Recursion Pharmaceuticals Aktie

Good morning, everyone, and thank you for joining Recursion's Third Quarter Earnings Call. We're hosting earnings live from the CGA-IGC Annual Meeting in Denver. Today, we'll share the new FAP data that was presented yesterday at CGA by Dr. Lisa Boardman, along with updates on our oncology program, REC-1245, and progress across our partnerships.

Before we get into the data, I want to start with the philosophy behind what we're building. To pioneer a new blueprint for making medicines with AI, we believe you have to tie together 3 core capabilities in one engine, uncover novel biology, design differentiated molecules against challenging targets and generate clinical evidence that matters for patients. And that is what we're building at Recursion. And the higher bar we set for ourselves is repeatability, not one target, not one molecule or one clinical result, but the ability to do this across programs again and again.

Let me show you the progress we're making. Let me start with what we're seeing in patients. We now have positive clinical data from 2 first-in-class programs. With REC-4881 in FAP, we saw deep reductions in polyp burden that persisted off treatment. Now we're also seeing activity across the GI and reduction of larger high-risk polyps. We'll walk you through some of that data today. With REC-1245, we've seen partial responses in heavily pretreated cancer patients, 2 confirmed and 1 confirmed PR. And in that program, our AI platform guided all 3 decisions from biological insight on which target to pursue, how to design the molecule and which patients to study.

If you take a step back to molecule design, Sanofi recently optioned a potential first-in-class oral I&I candidate. The chemistry challenge behind that was to design a single molecule that could inhibit 2 targets. And then when we talk about novel causal biology, we use the data that we have generated to build proprietary maps of biology and identify targets that may be driving disease. Genentech has now advanced the first neuroscience target from those maps, and we're working together on small molecules against it. That's the repeatability we're starting to see, the same engine of covering novel targets, producing new molecules and generating evidence in patients.

Behind the progress you just saw is one AI platform that connects biology, chemistry and the clinic. We draw on more than 50 petabytes of proprietary data to make decisions about which targets to pursue, which molecules to design and which patients to study. What makes that approach repeatable is how we learn. We generate our own proprietary data, which builds our models, which then in turn guide the experiments we run. And what we learn in the lab and in the clinic now then feeds back into those models. So an insight from one program can shape what we test in the next.

Let's start by looking a little bit more into the data from REC-4881 and REC-1245. Here, we're sharing a summary of the early clinical data from these programs. First, with REC-4881 and FAP, remember, there's no approved drug treatment. We previously have shown deep durable reductions in total polyp burden. What's new is, first, we have updated this data to expand the adult population to 18 and over, and we continue to see these substantial reductions. In addition to that, we're seeing significant reduction throughout the GI tract, including the upper GI, which is a huge area of unmet need.

In addition to that, we're also seeing reductions in the number of larger, higher-risk polyps. The safety profile remains favorable with mostly grade 1 or 2 adverse events and no additional discontinuations in this update. With REC-1245, as I mentioned, we are seeing early single-agent activity in heavily pretreated patients with solid tumors. Vicki will walk you through the partial responses and the exploratory expansion cohort we have opened.

And again, to say, look, the biological insight behind each program came from our platform. And with REC-1245, you can see how our use of the platform continues to evolve. We're not just using it for biological insight on the target, but we're also using it to design the molecule and now it's helping us decide which patient populations to study. You can see that same evolution in our partners as well.

Genentech optioned the first unexplored neuroscience target from our proprietary maps. Sanofi optioned the first development candidate from our collaboration, a potential first-in-class oral small molecule for I&I. And each of these programs have over $300 million in milestones with high single to double-digit royalties. In total, in our partnerships, we have now achieved more than $535 million in cash inflows from our partners. What gives me confidence in the repeatability of this work is that in roughly 2 years, we have earned more than $100 million across 9 milestones tied to the progress in these partnerships.

And I'd love to remind everybody that, look, underpinning all of this progress is how we use our data and the fact that we have one single engine. One of the biggest rate-limiting steps in science is the access to high-quality relevant data. And we have built a scientific data factory designed to generate this at scale. Our AI models draw on more than 50 petabytes of proprietary biological data to identify these novel potential causal targets.

We then test those predictions experimentally and feed the results back into the models to guide the next experiment. That closed loop is how proprietary data can become a map. A map can then service a new target, and that target can ultimately become a new program that either we or our partner chooses to advance.

We see also gains in productivity. Let's just take an example in designing molecules. In designing first-in-class molecules, we are now synthesizing 80% to 90% fewer compounds across our portfolio relative to industry benchmarks. In the clinic, ClinT agent can screen thousands of trial sites in minutes to help us find sites that actually see the patients we're trying to reach. We're seeing enrollment run about 30% to 60% faster than industry benchmarks. And we are advancing all of this work from a stronger financial position.

We just completed our oversubscribed $200 million financing in October, bringing in leading institutional investors and now expect our cash runway to extend into 2029. At the same time, as previously communicated, we have lowered our 2026 cash operating expense guidance from $390 million to $375 million. And then now looking ahead for 2027, we expect that figure to come down even further to under $360 million.

Before we go into a deep dive of the clinical programs, let me take you behind the scenes of a couple of our partnership programs and some of the updates to date. Let's start with the Sanofi announcement. We set out to design oral small molecule that could inhibit 2 targets involved in inflammatory disease. Now that's a difficult chemistry problem. Not only is it first-in-class, but a change that improves activity in one target could come at the expense of the other. And you still have to preserve the properties that actually allow this to be a chronically dosed oral compound.

For patients who may need this treatment for years, an oral option could make a meaningful difference in daily life. That's the kind of first-in-class design challenge we want to take on. Sanofi brought deep expertise in immunology and drug discovery. And our team used our AI platform to design the molecules, test them in the lab and use what we learned to refine that next design. Sanofi has now optioned the candidate and will lead and fund this further development with Phase I expected to begin in 2027.

This is the first development candidate Sanofi has optioned from our collaboration and our sixth milestone across 5 programs together. That track record is what gives us confidence in the repeatability of our work. The option triggered a $10 million payment, bringing total upfront and milestone payments from Sanofi to be $144 million. Under the collaboration, each program has up to $343 million in potential milestone payments plus double-digit royalties on net sales.

Now while with Sanofi, the challenge was to design small molecule across 2 targets, with Roche, we started with a separate problem or with a different problem. We knew what cancer pathway we wanted to reach, but it had been long considered undruggable. [indiscernible] reveal a way in. Roche highlighted this work at its September Investor Day this year. Together, we mapped how cells responded when we altered individual genes and then expose them to compounds.

It's this gene compound relationship that helped our biology engine to identify a new mechanism for acting on this pathway. We then advanced the most promising compounds and confirmed in vivo that they acted on that intended pathway. Much more work to come on here. And then in the neuroscience space, instead of knowing whether there was an undruggable pathway, the question was slightly different. The question was, can our Maps of Biology find novel causal targets that were previously unknown. This work is a great example of what deep collaboration looks like for us.

We generated high-dimensional proprietary Maps of Biology from human neurons to uncover genes with unexpected links to disease biology. Our teams have worked side-by-side to prioritize the most compelling targets and then tested them experimentally, building that body of evidence around which ones are worth advancing. Genentech has now option the first target from that work and small molecule discovery is underway that we're working on together.

Importantly, the value of the collaboration doesn't stop with a single target. We are continuing to work in advance using the same reusable maps to generate and test additional hypotheses, creating the potential for multiple programs to emerge from that one shared data set. Now what ties all these examples together is how we're using one platform across different programs and different disease areas.

With Roche-Genentech, we uncovered new biology that led to a novel target and more coming in neuroscience. With Roche-Genentech in oncology, we found a new way to act on a pathway long considered undruggable. And with Sanofi, we designed 1 oral molecule to act on 2 targets in immunology. And in our own pipeline with REC-4881 and REC-1245 in rare disease and in oncology, we're beginning to see those biological insights translate into activity in patients.

Now before I turn to the REC-4881 data, I also want to share an important Board update. Today, we announced that our Board unanimously appointed Fred Hassan as Chair. Fred has repeatedly built, scaled and transformed some of the most important companies in our industry, including as Chairman and CEO of Schering-Plough and Pharmacia. He also brings a wealth of Board perspective and experience from companies, including Amgen, BridgeBio and Prometheus.

What matters to me is the perspective Fred brings at this stage of Recursion. He understands firsthand what it takes to transform companies, to move companies from clinical proof through approval and launch while building enduring businesses around them. And he also shares our conviction that data and AI are creating a new blueprint from drug discovery and development.

Fred will serve as a Nonexecutive Chair, and I'm very excited to work with him as we enter the next chapter. I'm also grateful to Rob Hershberg for his commitment to Recursion and for the perspective he has brought on from many years as physician scientists and biotech CEO. Rob will remain on the Board, and we will continue to benefit from his insights.

Now let's turn to REC-4881, our investigational oral MEK1/2 inhibitor for FAP. Yesterday, in Denver, Dr. Lisa Boardman presented the updated TUPELO results in the presidential plenary at CGA-IGC. What makes this meeting particularly relevant to FAP is who's in the room. It brings together the genetic counselors who help families understand their risk, the specialized GI physicians and surgeons who take care of them and the researchers working on new treatments.

I'll share the highlights here. This update gives a broader and more complete picture of the program. It includes additional patients where we have now expanded eligibility to adults aged 18 and older. Importantly, we continue to see the depth and durability of response across and throughout the GI shot. After 12 weeks of treatment, 87% of evaluable patients had a reduction in total polyp burden.

What stands out to me is the depth of the response and how quickly we're seeing these responses. About 1 in 3 patients had greater than equal to 65% total polyp burden reduction in just 12 weeks. And when we follow these patients for about 12 weeks after stopping treatment, we actually see about 85% still had a lower polyp burden than when they first started.

The next thing that's really important is the activity we saw both upper and lower GI. Remember, I mentioned the high unmet need in upper GI polyp burden reduction. We see reduction throughout the entire GI with median reductions of about 55% and 50% in upper and lower GI, respectively. And more importantly, that activity extended to larger higher-risk polyps.

Of the 8 patients who had polyps larger than 5 millimeters at baseline, 6 had fewer of those polyps at week 13 and 4 saw those counts fall by at least 80%. And when we look at safety, we see a favorable for chronic dosing with no new safety signals and no additional discontinuations in this update. Majority of the AEs that we're seeing are Grade 1, 2.

To put these results in context, consider what it means to live with FAP. We estimate that more than 50,000 patients across the U.S. and EU5 have been diagnosed with this disease. It's a serious lifelong condition driven by loss of function in APC, a gene that normally helps keep cell growth in check. Patients develop precancerous polyps throughout the GI tract, hundreds to thousands, and there's no approved drug to treat FAP.

Even after surgery to remove the colon, patients still face a lifetime of surveillance and life-altering surgeries. There's a psychological burden with that, too. Every endoscopy brings another period of waiting for patients and their families. What will define this time? Will I need another operation?

I want to spend a moment on the upper GI because this is one of the greatest unmet needs that we have heard post-colectomy from KOLs time and again. Look at the polyp on the left. It's flat against the wall of the duodenum. That wall is thin and removing such a broad-based polyp carries a risk of bleeding and perforation.

When the disease can no longer be managed endoscopically, patients need surgery. These are life-altering surgeries to remove the duodenum, sometimes along with part of the pancreas. That's a major, major surgery with high morbidity and long recovery. And upper GI cancer remains a serious risk even after the colon has been removed. That's why we paid such close attention to whether REC-4881 could reduce polyp burden in this part of the GI tract.

It is that persistent and prolific disease burden that also helps the population we could ultimately serve. Our immediate focus is the estimated 37,000 patients across the U.S. and EU5 who are post-colectomy and require decades of lifelong surveillance and life-altering surgeries. We also have the potential to explore pre-colectomy patients following this immediate focus.

Together, these patients and these populations underpin our estimate of a roughly $15 billion total addressable market. But ultimately, the opportunity is defined by what we can change for patients. Could a medicine, an oral medicine meaningfully reduce this disease burden? Could it delay the need for another one of these multiple life-altering interventions or change the overall trajectory of lifelong surveillance. These are the outcomes that we're focusing on.

So let's jump directly into the program. How did we get from APC loss to a MEK1/2 inhibitor? That was not an obvious therapeutic connection. It was an underappreciated biology that our platform surfaced. Our platform compared the changes we saw in images of APC deficient cells with the effects of thousands of compounds to see could they reverse the disease-associated changes to healthy. And while thousands of these compounds were screened, MEK1/2 inhibition stood out as the top in terms of our AI model prediction and REC-4881, in particular, produced one of the strongest reversals of the disease-associated changes we measured. That's the unappreciated synthetic rescue we saw and what led us to license this particular molecule from Takeda to study in FAP.

Now while the platform service this novel insight, the biology also gave us good rationale, which is something that we're learning more about. Loss of APC is known for decades to activate beta-catenin signaling, which helps polyps form. Now as these polyps evolve, additional mutations can also activate the MAPK pathway. What REC-4881 is doing is interacting with both of those pathways. One, inhibiting the MAPK pathway, which is downstream of the additional mutations that these polyps accumulate.

In addition to that, given the crosstalk that is known between MAPK pathway and beta-catenin, MEK1/2 can also dampen some of the growth signals driven by APC loss. It is that platform insight that led us to test a MEK1/2 inhibitor in patients with FAP for the first time.

So what we see when we tested that hypothesis in patients, this tubular update includes additional patients, as I mentioned, and we've expanded that study eligibility to patients 18 and over. Each bar that you're seeing here represents 1 patient. And after 12 weeks of treatment, 13 of 15 evaluable patients is about 87% had a reduction, deep, deep reduction in total polyp burden.

So next, what we looked at is what happens if you follow the patients for 12 weeks on treatment and then we take the drug off and follow them off treatment. Essentially, what you see is a rapid reduction in polyps in the first 12 weeks. And what's interesting is even when they're off drug, you see a majority of the patients, about 85% or so, the polyp burden still remains below baseline. For patients living with a disease where they may need to manage it for decades, how much time they spend on treatment matters. The possibility of dosing flexibility could therefore be quite meaningful for patients and the persistence we're seeing off treatment gives us an important signal to explore as we think about long-term management.

We also wanted to know what happens to the larger higher-risk polyp. Here, we're counting polyp greater than 5 millimeters. As you can see, 8 patients had at least 1 at baseline. And by week 13, 6 of the 8 patients had fewer of these larger polyp and 4 so their counts go down by 80%. The individual counts really puts that into perspective. Two patients each went from more than 90 polyps, and when I say 90 polyps, 90 larger polyps to just 1. Another went from 5 to none. In these 3 patients, these counts also held at week 25 when they were off treatment for 12 weeks.

Now moving from the overall polyp burden. We will now break it into upper and lower GI. What we see is reduction in both the upper and lower GI. In this slide, specifically, you're looking at the upper GI polyp burden reduction. Remember, this is one of the greatest areas of unmet need post-colectomy. And just after 13 weeks, we see a 92% of the patients had a reduction in their upper GI polyp burden where 1 out of 2 saw about 65% -- equal to greater than 65% reduction in their poly burden and 2 patients had near clearance of 98% or greater.

Now let's again look at what happens to these patients when we take them off drug for 12 weeks. 11 of the 12 evaluable patients or about 92% or so, still had upper GI polyp burden reduction below baseline. That persistence matters. In the natural history that we've seen before, spontaneous regression was not something that was observed. Polyp burden generally increased over time. And the fact that most of these patients remain below where they started after 3 months of treatment is very encouraging.

These images also show what the change looked like in a patient with deep reduction. This patient had about 98% polyp burden reduction. On the left, you can see the polyps in the duodenum before treatment. On the right is the same part of the duodenum at week 13. Across the GI tract, this patient's total polyp burden measured by adding the polyp diameters fell from 1,344 millimeters to just 32 millimeters after 12 weeks of treatment. That's a 98% reduction. And of the 96 polyps that measured more than 5 millimeter at baseline, only remained above that threshold.

So how do clinicians put this in context? One measure they use is the Spigelman stage scoring. This grades the severity of duodenal polyps from stage 0 to stage 4, which is the most advanced. It takes into account the number of size of the polyps as well as what the pathologist sees under the microscope. Clinicians use that information to determine how often someone needs surveillance and also what type of intervention, including multiple types of surgery.

What you see here is at week 13, 6 of the 12 evaluable patients, about 50% actually down stage on their Spigelman score. Four went from stage 3 to stage 2, 1 went from stage 4 to stage 2 and another went from stage 2 to stage 0. And what's even more interesting is of the 6 patients, 4 had reached their week 25 assessment after 12 weeks of treatment, and all 4 remained at least 1 Spigelman stage below baseline. The remaining 2 patients had not yet reached their week 25 assessment at the time of cutoff.

And now let's finally also look at the lower GI tract. Again, these are post-colectomy patients, which means it's really focusing on the remaining rectum or the ileal pouch. 12 weeks post treatment, we see 82% had a reduction in their lower GI polyp burden. And again, you see a few patients that had deep, deep reductions of 88% and 100%. And again, when you look 12 weeks after stopping REC-4881, 7 of the 10 evaluable patients had lower polyp burden in the GI below the baseline.

What encourages me across the totality of the data we're seeing is the consistency. We saw reductions in both the upper and lower GI tract. We see reductions in the lower polyps that are difficult to manage and are higher risk. And we see that even when these patients come back for endoscopy 3 months after stopping treatment, most still had less polyp burden than they had at the start of baseline.

Next, let's look at safety. At 4 milligrams QD, 21 of 26 patients, safety evaluable patients, so that's about 81% had Grade 1 or 2 treatment-related side effects. The most common was an acne like rash. 4 patients had a Grade 3 event related treatment. 2 of these are skin reactions and 2 had elevations in CPK detected on blood test. We saw no grade 4 or 5 treatment-related AEs.

In terms of discontinuations, we saw 4 patients stop treatment because of Grade 1 or 2 events. Two of these cases involve pre-existing conditions, while the other 2 patients had stopped for Grade 1 diarrhea and Grade 2 rash. Most of the discontinuations occurred early in the study. 3 of the 4 have been reported in our first readout, and there are no additional discontinuations we see in the latest update. As the study progressed, we have refined how we work with sites to recognize and manage side effects early, including the increasing use of measures for rash.

We will carry that experience into our next study. So let me bring this back together for what it means to live with FAP. Even after having their colon removed, patients face decades of surveillance and repeated and invasive procedures. Ployps in the upper GI tract can be especially difficult to remove and patients have no approved drug to treat the disease.

In TUPELO, we have expanded eligibility to adults 18 and over. After 12 weeks in oral REC-4881, we saw deep, deep reductions of the polyps, which are then sustained when the patients are off drug. And we're also seeing fewer of the larger, higher-risk polys. For patients, the question is whether these reductions can change the course of their care? Could an oral therapy give someone more time between not only these endoscopies, but also surgeries? Over decades of living with FAP, having an oral drug could make a meaningful difference.

We're in ongoing discussions with the FDA and the design of the registrational study. And based on regulatory alignment, we're targeting initiation in the second half of 2027. With that, I'm going to hand it over to Vicki. She'll bring in KOL and patient perspectives, which are especially important as we think about these data, what they could mean in the context of lifelong disease management. She'll then walk you through the REC-1245 program and the early clinical data that we're seeing there.

Thank you, Najat, and welcome, everyone, to our KOL panel. I am pleased to moderate today's discussion to help contextualize the new Phase II data for REC-4881 presented at CGA-IGC to explore its relevance to the current unmet medical needs in FAP. Joining us today are 2 esteemed global FAP specialists. First, we have Dr. Evelien Dekker, Head of the Expert Clinic for Familial and Hereditary Polyposis and Colorectal Cancer, and Professor of Gastrointestinal Oncology at the Amsterdam University Medical Center. We are also joined by Dr. Bryson Katona, Director of the GI Cancer Genetics Program at the University of Pennsylvania, Perelman School of Medicine and TUPELO study investigator.

Dr. Dekker and Dr. Katona, thank you both very much for taking the time to join us. And Dr. Dekker, I'll start the first question with you. For those less familiar with FAP, could you walk us through the natural history of the disease, how patients are managed today and where the greatest unmet needs are given that there are no approved therapies?

Yes. So looking at the FAP patients and their disease course, most of the patients will be diagnosed at a young age because we know family members are having FAP and they will be genetically tested. We advise around age of 12. So that will be -- at that time, they will come to the pediatrician, and they will be scoped. They will be undergoing a full colonoscopy in our hospital, and we will look at the disease burden, which means we will see how many polyps are already present in the colorectal.

From then on, they will be checked upon regularly throughout childhood, usually 2 yearly, depending on the burden, it might be more frequent. And what we will by time do is resect polyps, mainly from the rectosigmoid because taking them out from the whole colon will be difficult because it's too many polyps because they will by time, develop hundreds to thousands colorectal polyps, which might, if we would not intervene, develop into cancer. To be sure we can prevent colorectal cancer, will undergo surgery so that part of the colon will be resected.

So at age 18, they are transferred to the adult gastroenterologists usually in a transitional phase. And we will follow the patients further. Again, if they're not operated yet, which might well be the case, we will check upon the full colon. If not, we will do the end of the colon. And approximately from age 25, we will also make sure we evaluate the stomach and the duodenum because not only in the colorectal, but also in the duodenum and then later also in the stomach, they might also develop precancerous lesions. So again, polyps. So lifelong checkups, necessary polypectomies or other interventions, maybe even surgery to make sure they have a long life.

And you did speak about the early management of the disease in the colon and the need for colectomies as well as how that shifts to more of an upper GI disease burden as well as in the retained rectum. Can you also tell us why does upper GI disease become such an important clinical challenge post-colectomy? And why is it more difficult to manage in some cases relative to the colonic disease?

So yes, patients have an increased risk for duodenal cancer and gastric cancer, and it is quite a challenge for us as gastroenterologists to battle that because preneoplastic lesions, so adenomas develop in the small bowel. But generally, they are more difficult to resect endoscopically and also endoscopic resection comes with more risk for complications.

So we try to, however, to do that endoscopically because otherwise, they might need surgery to prevent cancer. And duodenal surgery really is big surgery with major risks for both morbidity, but also in the end, mortality. The problem also is that after surgery, the risk in the small bowel remains. And it might be also more difficult to inspect the new area risk properly. So yes, there's definitely a challenge in the duodenum with -- specifically with morbidity of the operation is more difficult.

Thank you. Dr. Katona, turning to the REC-4881 to below data. We saw reductions in upper GI polyp burden in 92% of evaluable patients. And about half of those patients had deep reductions of about 65% or more and rapid reduction within that 12-week treatment window. From a clinician's perspective, how meaningful is the magnitude of polyp burden reduction like this? And how might that affect decisions about patient care?

Thanks, Vicki, for the great question and for the opportunity to discuss this data. When we think about reduction in polyp burden and what that means clinically, I think oftentimes what it means clinically is a little bit different from what people want to see from a regulatory standpoint. And unfortunately, polyp burden is not a regulatory endpoint that's typically used.

But from the clinician standpoint, it's incredibly meaningful. Most of what we -- most of the decisions that we make for patients with FAP are related to the polyp burden. And so like when we're determining when someone will go for a surgery, how soon they have to come back for another upper endoscopy or colonoscopy is typically dependent on what type of polyp burden we're seeing on the exam. So clinically, I think the polyp burden and seeing reductions in the polyp burden can have a major impact for patients.

So just to give you an example, so say somebody who is getting upper endoscopies because of duodeno-polyposis, if the number of polyps that you're seeing in the duodenum starts to reduce, they're starting to get smaller, I think a lot of clinicians may feel comfortable potentially extending out the interval at which an individual would undergo an upper endoscopy. So I think from the clinical side, the polyp burden and reductions in polyp burden, especially large reductions in polyp burden, do lead to meaningful clinical changes for our FAP patients.

Great. One particularly interesting finding was that 92% of evaluable patients with upper GI polyp burden reductions remain below baseline after 12 weeks of treatment. How do you interpret that durability or sustained effect of the drug? And how does durability affect how you might approach management of this lifelong disease with patients?

It's a great question. I think that the durability of a reduction in polyp burden is an important metric to understand for any type of chemo preventive agent in FAP. From my perspective, clinically, seeing a durable response is always very encouraging. I think in the kind of the dream chemo preventive agent for FAP is one that somebody doesn't have to take lifelong and potentially could be taken for a shorter amount of time or maybe potentially with intermittent dosing pattern so that individuals could see a response, have durability of that response. But I think it is a very encouraging first sign for this drug.

Thanks, Dr. Katona. I'd like to move on to the safety data, Dr. Katona. The safety data showed that most patients, about 80%, had treatment-related AEs of only Grade 1 or 2, and those included adverse events such as rash, including acneform dermatitis, diarrhea and left ventricular ejection fraction changes. 4 patients had Grade 3 events, and there were no events above grade 3 that were treatment related. Given your experience managing patients with FAP, how do you think about the benefit risk profile for REC-4881 in this setting?

It's a great question and an important question to think about the risk-benefit profile and how any side effects would be tolerated by patients. I think a patient's willingness to tolerate side effects is a very personalized decision based on the patient, but then also based on how advanced or what stage their FAP is at.

So just to give you an example scenario, if you have a patient who maybe has a very attenuated FAP phenotype, they're having a couple of polyps that form in their colon every year and they're getting the colonoscopy with complete polyp clearance, that individual may have a low threshold to tolerate any side effects from a medication since their polyps are being completely managed by colonoscopy. But if you go to the other side, if you go to somebody who has very severe FAP with advanced stage duodenal polyposis, say, for example, Spigelman Stage 4 duodenal polyposis, and they're getting upper endoscopies every 3 to 6 months and they're right on the cusp of potentially needing a, as Evelien mentioned earlier, a major duodenal surgery for resection, I think that individual is going to have a much higher tolerance for side effects, one to potentially avoid a potentially morbid and life-altering large surgery, but then also potentially to have to cut down on such a high burden of endoscopic monitoring.

One other important point to think about side effects and patient tolerance is, are we developing strategies where patients can essentially manage these side effects or take proactive steps to prevent them? So I think that's the other important thing to think about for side effects, especially when you know what the side effects are and especially if it's kind of a narrow window of side effects where you potentially could either prevent or kind of aggressively treat them to reduce their impact.

Thank you for that. I think it's a really important point and speaks to the need for educating both the patients themselves as well as the physicians who are treating them as you bring new therapies into the market. My next question is for both of you. I'll start with you, Dr. Katona, and then you, Dr. Dekker. As people learn about this space, they often ask whether there is a threshold of polyp burden reduction that would be clinically meaningful. For example, if you have a reduction, but you still have residual polyps, would that change how you treat the disease? What are your thoughts about that in clinical practice?

This I would say, that any reduction in polyps is a good thing in our mind. And I think that the fact that polyps don't go away is also okay as well. Those of us that take care of FAP patients, we know they're going to have polyps there. And essentially, we follow them to basically make sure that these polyps don't progress to cancer down the road. But I think when you're thinking about reduction in polyp, I don't think you need to think about it as all polyps have to be eliminated.

But then also to get to your point about threshold, I don't think that there's a specific threshold that we expect. Because again, I think that often is a personalized thing as well. So if you have, for example, somebody who has 2 polyps in their rectum and those polyps decreased by 80%, that's great. But I would be more excited if I had somebody with 100 polyps in their rectum that all decreased by 50% in size. And so I think it's -- I think any polyp reduction is a good thing to see, but I don't -- I think it's very difficult to put like a specific number because I think it's a very personalized scenario based on the patient.

Great. Thank you for that. And Dr. Dekker?

Yes. I would like to add that I agree about the personalized checkups we do. It's not a one size fits all. We scope everyone 1 yearly and see what happens. So if the polyp burden reduces, usually that means that we can prolong intervals at which we scope. And maybe even there could be not only a regression of the polyps present, but also less growth of new polyps. Because we know during the life of those FAP patients, new polyps keep coming, right? So if that would be a combination, reduction of the size of any present polyps and less new polyps, that would translate into, I think, a lot of benefits for the patients.

Great. Thank you, Dr. Dekker. This question is also for both of you, and this time I'll start with Dr. Dekker. To recap, we talked about the intensity of surveillance to decisions about polyp removal and major surgery, which is driven by the extent and severity of polyp disease. In closing, how significant is the unmet need for medical therapy in FAP for post-colectomy patients, especially patients with upper GI disease?

It's absolutely a huge unmet need for these patients, for these FAP patients, to have medication as a potential treatment because they suffer a lot from the big burden of surveillance as well as the risk for having cancer, both in the retained rectum or maybe in the pouch as well as upper, because the consequence when it's upper might be even larger if we get to the stage where presented polypectomies are not possible anymore and they need surgery. And again, more checkups after. It never stops during their life.

Thank you very much. Dr. Katona?

Yes, I can't agree more. I think Evelien's use of it's a huge need is spot on. People have been doing research in the chemo prevention space in FAP for decades and we still don't, at least in the United States, we still don't have an agent that is FDA approved.

And so we're -- FAP patients, we're diagnosing them. Our numbers of FAP patients are increasing. We're still having to send these patients for major surgeries. We're still -- we're doing a ton of scopes on these patients. And so the need, yes, the need is very large, and it's only going to increase over time as numbers of FAP patients continue to increase.

I'll close this out by asking if there's anything else that either of you wanted to mention that we haven't yet covered.

I'll just add on that I think this preliminary or early data is very exciting and promising. I'm definitely very intrigued to see how the rest of the results from this study turn out. And I know the FAP community has their fingers crossed and are very interested and hopeful as well.

Well, Dr. Dekker, Dr. Katona, thank you both for sharing your invaluable clinical perspectives on the management of FAP and helping us to contextualize the TUPELO data for REC-4881. We truly appreciate your time and expertise today.

Having heard the clinical and scientific perspective, it is now important to hear directly from individuals navigating the daily realities of FAP. We will now move on to our patient panel. I am looking forward to guiding today's conversation. This panel offers a special opportunity to step back from clinical data and hear directly from individuals living with FAP, allowing us to listen, learn and better understand their daily experiences.

Joining us today are Laura and Jesse Laster, who navigate FAP as a family. Jesse was diagnosed as a teenager and underwent a colectomy shortly afterward. To manage his ongoing cancer risk, he later required a complex while procedure. Today, Laura and Jesse are raising 2 children who also inherited the condition, and they are here to emphasize the need for continuous, lifelong care.

We are also joined by Lindsay Huff. Lindsay grew up knowing FAP would shape her life as she and her sister inherited the condition from their father. After undergoing a colectomy in her [indiscernible], Lindsay faced significant surgical complications, as did her sister who now requires a Whipple procedure. Lindsay is here to share her journey and show younger individuals that FAP does not have to hold them back.

Thank all of you for being here today. The first question is for all of you, but Jesse, let's start with you. FAP has been part of each of your lives and families for a long time. Can you walk us through your experience from when FAP first entered your life through your colectomy?

My journey with FAP started about the time I was 8 years old. I wasn't diagnosed at that time, but that's when I started really seeing my sister go through her journey -- or through her surgeries related to FAP. And then it was mainly thinking about -- thinking to the future when I would have to kind of go through similar situations.

So as you mentioned before, I was diagnosed at 15. I had my own colectomy and J-pouch at 16 years old. And it's really kind of the first part of my journey with FAP.

Thank you. Lindsay, maybe we'll move to you and then, Laura, I'll circle back to you.

Sure. So my experience with FAP has been lifelong. Started when I was born. My parents knew that it was likely that I would have FAP, but I know -- I believe it was either before or shortly after birth, I was genetically tested and found that I have the gene mutation for FAP. And my journey started around age 10 with getting my first colonoscopy and finding out that my polyp burden was significant enough that I would ultimately need my colectomy at age 10.

Thank you, Lindsay. And then Laura, as a family member who's had to navigate this, can you tell us what that experience has looked like for you?

Yes. So what the experience with FAP has looked like for me has really started when Jesse and I met in college and he explained that he had this condition and -- a little bit about it, but it was -- kind of seemed like I had this done and then you're good, now I am healthy. But actually, very quickly after we met, within a few months actually, he had a bowel obstruction. And I don't know, maybe that was my first clue that maybe the whole life with FAP is not just one-and-done thing after somebody has the colectomy procedure.

Thank you. And so maybe we'll talk a little bit more about that now and life after colectomy. Jesse, I'll start with you again, and then Lindsay. Many people assume that once someone has a colectomy, FAP is largely taken care of, I think, Laura, as you were just saying. What has been your experience, Jesse, in the years since colectomy? And how has your disease been managed over the years?

I'd like to start to kind of back up to before the colectomy, whenever I did -- or when I found out that I had FAP, it was very much the thought was that you just -- you get diagnosed with FAP, you have a colectomy and then that takes care of everything. I remember maybe in the following years after my colectomy, bowel obstructions became an issue due to scar tissue from the surgery. But even beyond that, I had such a high polyp burden in my stomach that doctors started being concerned with that and then some of the polyps in my small intestine and especially the duodenum that became more of an issue.

As mentioned before, I had a Whipple procedure because I had 1 polyp in particular that had what they call high-grade dysplasia or cancer in situ, they call it. They're just cancer located on to that one polyp. So that's -- I've since learned it's an ongoing maintenance for FAP.

Ongoing management throughout your life. Thanks, Jesse. And Lindsay?

My experience after my colectomy at age 10, our doctor had told us that every case of FAP is a little different, that sometimes FAP just sticks to the colon and sometimes other families have had concerns with it going to the upper GI. And so our management was doing upper and lower endoscopies every year. So after my colectomy, ultimately had another surgery at age 14 to have a J-pouch placed and then ultimately had some complications from those surgeries that resulted in a small bowel obstruction.

Thanks, Lindsay. I'll start this next question with you. How has monitoring or treatment of the upper GI tract factored into your experience with FAP? And how has that shaped the way that you think about your long-term health?

Sure. For me, I would say monitoring has been fairly stable with my upper GI. I do have a significant amount of polyps, but they're not staging anywhere super significant where there's concern at this time for any further surgical management. My sister has had a pretty significant upper GI burden. She's had multiple duodenum polyps. And at our last scope that we had in August of this year, they found 5 large duodenal polyps that they staged at Stage 4. So ultimately, she's getting a duodenectomy, which has been described to us as a pancreas-sparing Whipple procedure, and she'll be having that within the next week.

You've mentioned that your sister has an upcoming Whipple procedure. And over time, as she's been scoped, as the upper GI has been scoped, did you or she see this coming? I'm wondering, did she have time to sort of mentally prepare herself? Or was it kind of a more sudden?

I guess I'll start with she didn't really have a lot of time to prepare for this being a possibility. Last year, she had an upper endoscopy done, and it was found that she had a pretty large duodenal polyp. They didn't say whether it would come back or not, but they ended up doing a procedure to remove that polyp and place a stent.

Fast forward to this year, she's got 5 polyps that are long and kind of [ granchy ] and there's risk of obstruction. They ended up staging them and they were looking like Stage 4 polyps, which is about the worst you can get before they're officially cancerous. And basically, the only thing presented to her was this procedure.

Thank you. And then, Jesse, perhaps answering the same question, how has the experience of having FAP and the upper GI disease as well, how has that shaped how you think about your health over the long term?

I guess the way that I see my managing my health with FAP has changed over the years. Like I mentioned before, from going from it's taken care of and don't need to worry so much about it, having scopes every few years, was kind of thought to be the way to manage it. It went to having scopes every 6 months, so got a lot more complicated.

Certainly having the Whipple procedure and having the cancer diagnosis definitely changed the way I see that. I talked to my doctor about what are the next options if I get cancer again. And I was curious about can the Whipple be done a second time? And he said no. So there's a lot of concerns with ongoing management.

Thanks, Jesse. For this next question, Jesse, I'll start with you again, then Lindsay, and then back to you, Laura, as well. Jesse, when high-grade dysplasia was found and a while procedure was suggested, what were your considerations in terms of making the decision to move forward?

I think specifically for the Whipple procedure decision, there didn't appear to be much of a decision. The polyp that was found with high-grade dysplasia, it was presented to me as the only way to take it out was to take out the duodenum and the Whipple procedure was the procedure to do for that.

There were other decisions that I had to make in my journey. I've had doctors tell me that I needed to have a permanent ileostomy or even revisit my J-pouch. And then I even more recently had a doctor suggest that I may have to have my stomach removed. There's been a lot of decisions to make in my journey, and it's been a struggle, I guess, to manage quality of life and what is the best choice to make to manage my FAP.

And then, Lindsay, you've had some significant complications from some of your surgical procedures. Are you able to share some of those experiences? And then how does that influence how you think about any additional procedures when they're recommended?

Some of the serious complications, the major one that I have had is from my colectomy and my surgery to get a J-pouch, I had that temporary ileostomy and had it reversed. In 2016, I was a freshman in college, and I had adhesions from those previous surgeries. I was in my dorm room shower and I went over to grab soap and my bowel flipped. And so that caused a small bowel obstruction. I was emergently ambulanced to a local hospital. We didn't know whether they would know anything about FAP or anything going on with my health and ultimately had 70 centimeters of my small bowel removed. I had another temporary ileostomy placed from October of 2016 to May of 2017 when I finished my freshman year, and then ultimately had that reversed again.

Ongoing complications, I don't really have much bowel left at all. So there's a lot of issues with malnutrition, absorbing nutrients from the food that I'm eating, and continued concerns for adhesions and potentially looking to avoid another bowel obstruction that could happen. And ultimately, there's not enough bowel to reverse that in the future and I would have a permanent ileostomy for the rest of my life.

So thinking about knowing that complications are a possibility because they've already happened, if there was another procedure recommended, I believe it would be an emergent procedure and there really wouldn't be much decision. It would just be basically life-sparing for me.

Thank you for sharing that. This one is for all of you again. Thinking about your own experience and your family's future, what would a medicine for FAP need to do to make a meaningful difference in your life? And when you think about the future of FAP care, what would meaningfully change your daily life?

I think as far as my medical care goes, I would see a medication that reduced the amount of procedure that I had to do would definitely help with my care, reducing the polyp burden and especially the polyps that are concerning to my care team. That would be a great advantage, I think, to medical care.

An extension of that would be for my own 2 children. They're still early on in their journey. They haven't -- hadn't have any surgeries yet. I would see a great benefit to their medical care even to be able to reduce polyp burden and eliminate polyp burden potentially for them and then reduce or eliminate the need for surgery for them. So kind of an echo of my care needs would be on them too.

Thank you. Laura, anything you'd want to add to that?

Yes. I mean just in looking at a medicine that could be -- contribute to less polyp burden, as Jesse said, less procedures having to happen. And I mean where we live, we're driving 4 hours one way to deal with this. It's an expensive hospital that we go to. It doesn't accept all insurances. And so there has been just life considerations and having to do less ongoing procedure maintenance besides, of course, I want you around a long time, it's also just like an expense issues too. I mean it's something we need to do, but it's something we have to plan for as a family. So if there were less requirement of monitoring and avoiding so many procedures, all of which come with risks, that would be fantastic.

Thanks, Laura. Lindsay?

I have to echo a lot of what Jesse and Laura said. I think a medication that could reduce polyp burden or potentially eliminate polyp burden would be beneficial to prevent any future complications for myself or my family. Thinking to the future, my wife and I have started thinking about starting a family and what does that look like? I would love for my children to not have to go through the same experiences that I've had to go through.

And echoing again what Jesse and Laura talked about with the financial burden that comes with that, we -- my family flies across the country to go and receive specialized care. There's a significant amount, not only for travel expenses, but the procedures that we're having to have done. I think that would, with a lot of the financial burden that we have to think about when we're having medical care done as well. And I think a medication that could potentially lessen the amount of monitoring that we have to have could really help as well.

Along with the issues with insurance, we've talked about facilities take a lot of different insurances, but they don't always take all insurance. And so this year specifically, I am in South Dakota and I have my health care company's insurance plan, even though my insurance was going to be accepted as in-network because they didn't accept the insurance -- my insurance normally, they take 50% of what you're going to ultimately owe to the medical facility upfront, which for me, the total that they were estimating was around $15,000. And as a 28-year-old, I don't have $15,000, let alone much more under that, available to give upfront. There's a lot of financial hoops that you have to navigate and jump through and advocate for yourself for, especially when you need that specialized care.

Lindsay, Laura and Jesse, thank you so much for sharing your stories and giving us a window into the daily realities of living with FAP. Hearing your experiences firsthand brings such important perspective to our discussions and reminds us exactly why this work matters so much. We truly appreciate your time today.

I will now move on to talking about our REC-1245 program in solid tumors. REC-1245 is an investigational orally administered and potentially first-in-class molecular glue degrader of RBM39, a master splicing regulator. Degradation of RBM39 disrupts the splicing of stress response in DNA repair genes that cancer cells depend on for growth.

The molecule and the initial insight came from our engine. Our phenomics found that RBM39 degradation phenocopies CDK12 loss while sparing the CDK13 associated liabilities that have made that biology so difficult to drug. We screened 204 compounds to achieve a development candidate in 18 months, compared to an industry benchmark for small molecules of roughly 2,500 compounds over 4 years.

In dose escalation, we are seeing evidence of clinical activity, 3 partial responses, 2 confirmed responses in endometrial cancer, and 1 unconfirmed response in hepatocellular carcinoma. Disease control rate at 8 weeks was 91.7%, 11 of 12 patients, at doses of 45 milligrams and above. We've now selected 45 milligrams once daily for expansion, a dose with no serious treatment-related adverse events and no dose reductions reported.

Let's look at how Recursion's engine got us here. The heat map on the left is an actual platform output. RBM39 clusters phenotypically with CDK12 and not with CDK13. That's the insight. CDK12 loss has been understood as a vulnerability in genomically unstable cancers for a long time, but CDK12 inhibitors have struggled because they hit CDK13 too. RBM39 gave us a way into this pathway that others hadn't seen. In the middle, RBM39 ensures proper splicing of critical stress response in DNA repair genes. In nonclinical models, RBM39 degradation leads to disruptive mRNA splicing and impaired DNA damage repair and cancer cell death. REC-1245 is highly selective and orally bioavailable and drives rapid RBM39 degradation in human PBMCs within 24 hours, with significant tumor regression seen across multiple xenograft models driven by replication stress.

The potential opportunity is broad with more than 100,000 addressable patients a year across the U.S. and EU5, spanning hard-to-treat tumors like endometrial, HCC and ovarian, and transcriptionally driven cancers, including pediatric tumors like Ewing sarcoma and neuroblastoma.

We see 2 complementary development strategies: monotherapy in tumors that are intrinsically primed and vulnerable, which is where we are today in dose escalation and now expansion; and combination, where REC-1245 could create vulnerabilities alongside standard of care and potentially move into earlier lines of treatment.

Our Phase I dose escalation enrolled patients with unresectable or metastatic solid tumors or relapsed/refractory lymphoma who've progressed on available therapies. Six cohorts were enrolled on continuous daily dosing. The primary objectives are PK and safety. Antitumor activity is the secondary objective. Based on the results of the dose escalation, we have selected 45 milligrams daily for our expansion cohorts and are currently enrolling these, including 1 in endometrial cancer. I'll walk through safety data and evidence of antitumor activity on subsequent slides.

In terms of safety, REC-1245 was well tolerated at doses up to 45 milligrams daily. The 65-milligram daily dose exceeded the maximum tolerated dose with 3 of 4 patients enrolled on that cohort experiencing dose-limiting toxicities, including stomatitis, neutropenia, febrile neutropenia and anemia. The most frequently reported treatment-related adverse events across all doses were fatigue and anemia at 30% each, followed by nausea, thrombocytopenia and stomatitis. About 27% had a Grade 3 or higher treatment-related event, with the most frequent being anemia.

At the 45-milligram dose, which was selected for expansion cohorts, the most common treatment-related adverse events were Grade 1 and 2 fatigue, anemia and nausea. We saw no dose-limiting toxicities and no serious treatment-related adverse events. The 2 Grade 3 events at that dose were elevated creatinine and hypokalemia.

Turning to activity. This is a heavily pretreated patient population with patients having received between 2 and 8 prior lines of treatment. There was no evidence of antitumor activity at doses below 45 milligrams. Across the 12 response evaluable patients who started at 45 or 65 milligrams, 11 of 12 had evidence of disease control, defined as an initial response of SD or better at 8 weeks. Four of the 12 remain on treatment.

Among these 12 patients, there were 3 partial responses, seen here on the right, 2 confirmed responses in endometrial cancer at 38.2% and 48.9% reduction, and 1 unconfirmed response in hepatocellular carcinoma at 55.9% reduction. The 2 patients with endometrial cancer who have responses continue on treatment, while the HCC patient has discontinued.

I'll now walk through case studies of the 2 confirmed responses. First, a 70-year-old woman with recurrent progressive Stage 4 high-grade serous carcinoma of the endometrium, which is p53 mutant, HER2 non-amplified and microsatellite stable. She's had 4 prior lines of therapy, including platinum chemotherapy, a TKI, immunotherapy and an antibody drug conjugate. She started at 65 milligrams, interrupted for Grade 2 oral and abdominal pain and restarted at 45 milligrams. She has tolerated the 45-milligram dose well and abdominal pain resolved.

At the week 9 CT scan shown on the right, we saw a partial response with 32% reduction in target lesions in the lung, liver and pelvis. The response was confirmed at week 17 with a 49% reduction in target lesions. She's currently asymptomatic from her disease and remains on therapy.

Next is a 54-year-old woman with recurrent progressive Stage 4 endometrioid carcinoma, which is p53 mutant, HER2 amplified, ER positive and MMR proficient. She had received 7 prior lines of therapy, including platinum and taxane chemotherapy, a VEGF inhibitor, HER2-targeted therapy, immune checkpoint inhibitor, the tyrosine kinase inhibitor and antibody drug conjugate, and investigational CAR-T therapy.

She started REC-1245 at 45 milligrams. At her week 9 radiographic assessment shown on the right, she had a partial response with a 33.9% overall decrease in target lesions, including periportal and mediastinal lymph nodes. At week 13, that response was confirmed with a 38.2% reduction in target lesions. She has tolerated the drug well with predominantly Grade 1 and 2 GI adverse events and laboratory abnormalities, and she remains on treatment.

As I mentioned earlier, we are currently enrolling an expansion cohort of patients with endometrial cancer who have disease progression following available therapies. Endometrial cancer is the fourth most common cancer in women in the U.S. with roughly 68,000 new cases expected this year, and treatment options are limited in the metastatic setting after first-line therapy, especially for the roughly 3/4 of patients who do not have HER2 amplifications and are therefore ineligible for a HER2-targeted ADC. Five-year relative survival with distant-stage disease is around 20%.

Based on the mechanism of action and the activity seen to date with REC-1245, we see opportunities for 1245 as a monotherapy, including potentially in the post-ADC setting, and potentially in combination with existing treatments for endometrial cancer, including immunotherapies. These are hypotheses to be tested and REC-1245 is investigational.

We expect preliminary data from this cohort in the second half of 2027, and we're also exploring additional tumor types and dose expansion.

I'll now pass it to Ben Taylor for further business updates.

Thank you. Earlier this month, we completed an oversubscribed follow-on offering with $200 million in gross proceeds. Over 90% of the financing came from health care dedicated specialists and mutual funds that did not own Recursion prior to the financing. We were able to access this new pool of knowledgeable, long-term-focused investors based on the strength of our pipeline data, demonstrated repeatability of the AI engine, and continued operational discipline.

Including the net proceeds from the offering, we had a $749 million in pro forma cash and equivalents as of September 30, which does not include the recent $10 million milestone from Sanofi. We have extended our runway into 2029 through not only financing, but also continuing to concentrate capital on the highest-value operations. As I've said in the past, we want to continuously increase the impact of every dollar we spend. In this quarter, we are in a stronger position to do that than ever before.

Our 2026 full year cash operating expense guidance of less than $375 million, excluding partnership inflows and onetime expenses, represents a nearly 40% reduction from comparable 2024 pro forma expenses. We are going to continue our year-over-year efficiency gains. In 2027, we expect cash operating expenses to be less than $360 million. We fundamentally believe that $1 invested next year should achieve more than $1 today. Through adoption of technology and disciplined data-driven management, we expect to keep lowering OpEx while aggressively investing in the best of our pipeline, partnerships and AI engine.

We are concentrating our investment where we see the clearest path to value. In our pipeline, that means prioritizing highly differentiated wholly-owned programs and also knowing that we shouldn't try and do everything ourselves. Since 2024, our partnerships with Roche and Sanofi have produced more than $100 million from 9 progress-based milestones, bringing the total proceeds from partners to over $535 million.

Internally, we build and test our AI models on real pipeline programs so that we can efficiently identify which technology investments are likely to have real-world impact. Our in-house supercomputer allows us to move quickly while still controlling compute expense.

Finally, our investment over the last decade to create more than 50 petabytes of multimodal proprietary data gives us a large and reusable asset that creates unique insights beyond what is publicly available. With a stronger balance sheet and lower cost base, we believe we have the time and resources to deliver multiple important milestones across our pipeline and partnerships.

And with that, I'll turn it back over to Najat.

Thanks, Ben. Now let me bring it all together. The financial strength you just heard about from Ben gives us the ability to keep investing behind programs and capabilities where the evidence is the strongest. And this slide shows just what that means across the engine that we talked about at the beginning of the call: discover, design and deliver.

We now have delivered tangible proofs across all 3. In discovery, Roche/Genentech advanced the first unexplored target in neuroscience from our maps. In design, Sanofi auctioned our first-in-class development candidate, which is an oral molecule in I&I. And in the clinic, REC-4881 and REC-1245 are generating encouraging evidence in patients.

Now the focus is to do it again. In discovery, we are working towards additional targets with our partners and expanding our proprietary maps of biology across additional disease areas. In design, the candidate that Sanofi just optioned is expected to enter Phase I in 2027 with potential for additional partner milestones in 2027. And in the clinic, we expect to update our FDA conversations for REC-4881, generate additional Phase II data and, of course, depending on regulatory alignment, initiate a registrational study in the second half of 2027. We also expect preliminary REC-1245 expansion data and additional clinical readouts from the rest of our portfolio.

This is the bar that we've set for Recursion. Again, as I mentioned before, not one target, not one molecule or not one clinical result, but the ability to move repeatedly from novel biology to differentiated molecules to evidence in patients, all from one connected AI-native engine.

We have the capital to execute. We have concentrated in our focused set of priorities and a series of meaningful catalysts ahead. We're excited about what's to come next.

With that, we will now move on to Q&A.

With that, now let's go to some questions. We have quite a few questions on REC-4881. So I'll start with a question from Alec from Bank of America. On REC-4881, what does the activity in larger high-risk polyps tell us about the drug's ability to change standard of care? Vicki, I'll have you take that question first.

Thanks, Najat. So patients with FAP, as we've spoken about and as you've just heard, face a lifetime of surveillance and procedures, interventions such as surgeries and excisions. And the standard of care currently, there are no approved drugs in FAP that can alter the disease progression, right? So over time, polyps continue to grow, both in the lower as well as the upper GI tract. Patients have routine surveillance. The surveillance frequency is somewhat determined by the extent of polyposis, including the size of the polyps. And there's nothing out there that can, approved, that can alter the course of the disease and modify the need for these procedures.

So in terms of what we're seeing with the TUPELO data, we've seen substantial reductions in polyp burden overall as well as in both the upper and lower GI tracts. And we've seen an effect on larger polyps of 5 millimeters or greater. And as the polyps get larger, this drives the need for increased surveillance frequency. It drives the need for additional procedures, including excisions as well as potentially additional major surgeries down the line.

And so what we've heard is that KOLs are really impressed with the data. So we're here in Denver right now at the CGA-IGC conference. We've been meeting over the last 24 hours or so with many of these KOLs. And they're really impressed both by the extent of polyp burden reduction as well as the durability of that reduction.

And so how we're thinking about this is REC-4881 could potentially be an adjunct to routine surveillance. And the Phase III will determine whether that reduction -- that impressive polyp burden reduction that we're seeing can actually translate into benefit that's meaningful for patients in terms of reduction in the outcomes that are important, such as cancer development, high-grade dysplasia and also need for these surgical procedures.

Thank you, Vicki. I'll take the next question from Steve from KeyBanc. If this drug reaches commercialization, what percentage of the addressable patient population do you think will continue to go the surgery route?

I would say, Steve, I mean, similar to and building on what Vicki mentioned, these patients go through a series of interventions and a lot of them is driven by polyp burden, but definitely the size of the polyp, as well as also dysplasia and other factors. So we do think the fact that we are seeing a reduction in the large polyps, reduction in the polyp burden overall -- and the addressable population is pretty large, right? About 37,000 patients that are post-colectomy. We do think that there's potential for a pretty significant impact in terms of either avoiding or delaying these kind of life-altering interventions that you've heard on the patient panel. I mean a duodenectomy where you're either sparing the pancreas or even removing the pancreas, Whipple procedures, but also just [ epilectomy ]. There are so many procedures that these patients go through. So we're excited, and we'll see more -- we'll learn more in the coming months.

Next question. How are you thinking about chronic dosing for REC-4881? Commercially, do you expect patients would require scheduled drug holidays, 13 weeks on, 13 off pattern used in the clinical trial or regular treatment breaks?

I'm going to hand this over to Vicki, but I will say, we're both here in Denver at the CGA conference. And one thing that we've heard time and time again is that, at the end of the day, these patients are so young, they're in their 20s post-colectomy, or even younger, that having flexibility in terms of the dosing regimen -- and this is a question from Priyanka at JPM, so thank you, Priyanka, is a really, really big need and something that they haven't really had even an optionality before. But Vicki, I'll hand it over to you for more details.

Yes. Thanks, Najat. We are hearing a lot of excitement around the data, a lot of excitement around the durability in that treatment break. And that allows flexibility for patients too in terms of dosing, in terms of life events. So having the flexibility to be able to take a break from treatment is really meaningful for both patients and physicians. So I think that durability of effect is really important.

Ultimately, how this will play out, I think, ultimately will depend on the Phase III data in terms of the chronic dosing. But having that opportunity to have a drug that both shows significant polyp burden reduction, that may reduce -- ultimately be approved in the Phase III, may improve these events that are important for patients, and then the flexibility to be able to stop dosing and have an opportunity to take a treatment holiday or a treatment break.

Thank you so much, Vicki. Okay. Now we'll move on to the next set of questions. This is from Gil from Needham and also Alec from Bank of America. Given large AI labs are increasingly leaning into applications in health care, how are you positioning Recursion to compete? Does unique data win the day here or maybe something else?

I'll take that question. I think it's great to see the interest in terms of the convergence of AI and health care, which Recursion really pioneered for many, many years. Here's what I'll say. There's -- I always say there's 3 areas of moat for the company. One, yes, the data aspect, the data factory, the 50 petabytes of data, but not just the volume of data, the quality of data to build models and close the loop to find novel targets is critical. We pioneered in that a long time ago. We've built on that, some of the examples with Roche/Genentech, Sanofi, et cetera, that you see really build on not just data matters, especially in a space. As we've seen with LLMs, a lot of the human history of text and data exists, but you don't have that in science. So being able to create that data and then actually leverage it to find novel targets, design novel molecules and ultimately show data in the clinic, we think, is critical.

But the next step that's even beyond that is, look, I said that at the beginning of the earnings call today, we firmly believe in order to actually make differentiated medicines with AI, you need to tie together those 3 capabilities: novel targets, designing molecules for those novel targets and then actually showing that evidence in the clinic, which we're starting to do. That has taken a lot of time and a lot of learning and reps to build. That connected AI engine end-to-end, it's a huge amount of work where you actually have your data set build the models. The models then actually you see the predictions work out in the experiments. And those experiments and that data teaches the model that this sort of recursive learning, that technical aspect of it and actually having team members that are bilingual that can actually understand both science and AI, that is not an easy thing to build. So I think reps in terms of that closed-loop learning system is an incredible area of differentiation for us and something we continue to build on every single day.

And then the third is how does any company compete that has differentiated assets? At the end of the day, it's about competing based on really valuable data that's scientifically, medically and commercially critical to get better medicines for patients, and you see us having that as well, in that diversification and having that end game in mind is very critical.

Last thing I'll say, so we live in a space where 90% of biologies are known, 80% of diseases don't really have disease-modifying drugs, there's a huge amount of white space. So looking forward to the entire sector, whether it's the frontier labs, whether it's pharma, whether it's biotech companies like -- and then AI biotech companies that does all make progress here. You've heard about FAP. You've heard about oncology programs for patients that are so late-line treatment on the path to palliative care and trying to get data there and see activity and have them be asymptomatic disease. There's so much room for us to all play here. So we are excited about that. But our position is we want to be known as a company, the real deal in terms of making differentiated medicines using AI.

All right. Next question. This is a question from Dennis at Jefferies, on REC-1245. RBM39 chemotherapy has shown about 17% ORR, 7-month [ POR ], 3.9 months PFS, 12-month OS, et cetera. With the endometrial expansion cohort starting, what target profile are you expecting? And perhaps, can you comment on whether this MOA, if there's a benefit that can be seen via ORR or PFS, OS, et cetera? And lastly, how do you think RBM39 could fit in the late-line endometrial given Merck's Trop2 ADC data at ESMO and Gilead's Trop2 as well upcoming in Q4? I'll have Vicki start with that question.

Thanks, Najat. I think first, just to note, again, as you've heard in the webinar, RBM39 is novel biology. We're potentially first-in-class. This is again based on Recursion's engine, and we're seeing really promising data coming out of this.

So just to contextualize the data that we shared today is from a dose escalation cohort. These are patients, heterogeneous and heavily pretreated, right? So a very difficult patient population to treat, the early Phase I population. And in spite of this being a late-line population, we're seeing substantial evidence of activity.

So we have 2 confirmed responses in patients with endometrial cancer. That means they had 2 scans, one that showed an initial response based on [ RECIST ], and then a second that showed confirmation of that response. And then one unconfirmed response in hepatocellular carcinoma. And again, at a dose where we've been able to select a dose that's tolerable for patients to move forward into expansion cohorts, which we've now initiated recruiting.

So those expansion cohorts will give us a better sense of the activity and the safety in well-defined patient populations. We will take those data and compare them both to historical response rates for agents such as chemotherapy in this setting. We will be looking, of course, at response rate and a duration of response, as well as the safety profile to say, where do we go next with this particular drug in these particular tumor types? And of course, we're watching the competitive landscape very carefully, including the Trop2 ADCs. I will note that both of the endometrial cancer responders that we had in the dose escalation cohorts had received prior ADCs, including one of them who had received a Trop2 ADC. And we see that activity in patients specifically who had a prior exposure to ADCs.

So we see potential for a monotherapy approach. We see potential for combination approaches, including in patients who've received prior -- I'm sorry, including combination with ADCs. And these expansion cohorts, we're looking forward to sharing more data that will help guide what those next steps are for the molecule.

Thank you, Vicki. And yes, I mean just the other thing, to build on that, is in terms of RBM39, the MOA is distinct from ADCs. I think that gives us a potential for synergistic value. And then also the fact that these patients are such late-line patients, as Vicki mentioned, and have actually progressed on these ADCs, and we're still seeing [ single-agent ] activity, is something that's encouraging for us, but a lot more work to come on the expansion cohorts.

All right. So the last question we're going to take from Sean from Morgan Stanley. Following the recent financing, you now have runway into 2029, while guiding to lower operating expenses in 2027. How do you balance capital discipline with the opportunity to accelerate higher conviction or high-conviction programs? And under what circumstances would you choose to meaningfully increase investment behind REC-4881,REC-1245 or additional platform-generated assets?

It's a great question. I mean one of the things, and Ben mentioned this, that's really important for us, I mean it's not common to be lowering our guidance, and we are doing that in a couple of different ways. One is really concentrating capital on areas that we have high conviction and that is purely data-driven. So we see that with REC-4881, REC-1245, and we are fully funding those programs and accelerating them.

Some of the other additional programs in the pipeline, again, as we see data roll out next year, we'll do the same. I think the fact that we're incredibly disciplined based on data and our capital allocation has really helped us a lot as we have increased focus, ambition and discipline.

The other thing I'll also say in terms of the platform perspective and partnerships, we just crossed $535 million in cash inflows, I mean 9 milestones in the last 2 years, over $100 million in milestones, and I think that repeatability is very encouraging to see with our partners such as Roche/Genentech and also Sanofi. And that's something we'll keep progressing and we invest really effectively in them.

And then the last thing is the AI platform. I mean part of it is we're seeing productivity in our engine, that is really, really helping us stretch every dollar and be much more productive, whether it is in the design of molecules or even the generation of maps, which we're faster at, more efficient and more effective.

So all in all, it's all about being data-driven, making sure we stretch out every single dollar, making sure we return value, of course, for our patients and our shareholders.

So with that, I just want to say a big thank you to everyone for joining today. This was a combination of both our earnings and also a mini-FAP Day. A huge thank you to Dr. Dekker, Dr. Katona, and also patients, at the end of the day, this is why we do what we do, like Lindsay and Jesse and Laura, that are experiencing FAP and the dramatic and challenging impact of those diseases every single day. We do this for you. So thank you for joining us. Thank you for sharing your stories. And thank you all again, and we will talk very soon.

Good morning, everyone. I'm Sean Laman, head of U.S. Mid-Cap Biotech Equity Research at Morgan Stanley, and welcome to our Global Healthcare Conference. Before we commence, I'll make you aware of some important disclosures. For those disclosures, please go to the Morgan Stanley Research Disclosure website at www.morganstanley.com/researchdisclosures, and if you have any questions, please reach out to your Morgan Stanley sales representative. So, post that, we now welcome Recursion Pharmaceuticals with CEO and President Najat Khan and CFO Ben Taylor. Welcome to the both of you. Thank you.

Maybe do me a favor here and answer sort of three macro type questions and So just think about the rise of China-originated innovation. Does it change your competitive positioning and your R&D and BD playbook?

Yes. Yes, thank you so much, Sean, for the question, and it's great to be here. I would say it, in fact, reinforces what we do. I mean, Recursion started with the whole concept of better understanding and decoding biology. So a huge amount of focus for what we do is around novel biology. I was, in fact, just recently in Shanghai two weeks ago, and just the speed and the execution and the prowess in that, in optimizing known biology, was really impressive to see. And it's also reflected by all of the deals that have been done to date. But that next frontier has to be new and novel biology, really first-in-class targets. And that's evidenced by what we have in our clinical program. REC 4881, REC 1245, we'll talk a little bit more about that I'm sure, which are both first-in-class targets that came from our engine, novel biology, or novel biology linked to the right indication, which is the case for REC 4881.

The other thing I'll also say, we recently announced an unexplored new target that was optioned in neuroscience by Genentech. That comes from our proprietary data maps that we've generated. So taken together, that's the premise of what Recursion was really founded on around novel biology, and I think it becomes even more important, and some could say prescient, in terms of what's happening with the optimization and the speed with in China as well.

Sure, thank you. I'll skip the next question because it's AI related, but you are an AI-derived company. So, you know, which policy variable is it? Is it FDA, Medicare negotiation, MFN tariffs, or global pricing? Do you think about the most, and what matters most to your business on the regulatory front?

Yes, it's a great question, Sean. I think all of the above are, of course, important to take a drug and actually commercialize and ensure this access to patients. For the stage of the company that Recursion is, FDA regulatory guidance, et cetera, is critically important, and we have a very productive relationship with regulators. I would say, in addition to that, as we think about the commercial potential of our assets, of course, the other policies that you mentioned, MFN and others, are also critical.

Sure, thank you. Some questions here on strategy. So you've been elevated to CEO over the last 12 months. You've cut operating expenses materially. You know, a bit of level setting. How should investors view Recursion now?

Yes, I know, nine months, Sean, nine months. Can't take, I'm just kidding. Here's what I would say, Recursion is a fundamentally different company than 12 to 18 months ago. Recursion has transitioned from a company with a promise in the novel biology platform to a company where the engine, I like to call it, has been expanded to all parts that you need to make a drug: biology, chemistry, clinical. And to do that, what we have done is we are building the body of evidence in the clinic and beyond into in terms of differentiated outputs that are being generated by that engine. I think that is critical. And you're starting to see repeatability across the progress we're making both in our internal pipeline and also in our partnership.

And I'll go through that just in a second. And then the third piece that's materially changed, Sean, as you mentioned, is just our level of productivity. We have really leaned into AI agents, automation, as well as concentrating capital where we have the highest conviction. We do that every single day. And therefore, we've reduced our cost base about 40% pro forma post the integration with Exscientia. Just to give an example, you know, it's good to say those three pillars, but just to give an example in terms of where we're seeing that body of evidence translating from the engine. Number 1 is the first-in-class program in our clinic. So REC 4881, first in class for FAP, total addressable market, just to remind everyone, nothing approved. This is an oral drug. It's over $10 billion. We may also have multiple programs with 1, for instance, REC 1245, RBM39, first-in-class target, identified in our platform, the degrader made on our platform, and should have more data later this year.

That's just 1. Then we've also made a lot of progress with our partners. We've just crossed over $525 million in upfronts and milestones where we're actually advancing first-in-class or intractable biology programs with our partners. These are not isolated events. I think that's really important to understand. You're seeing this repeatability across from novel biology, novel programs, to the clinic. That's all produced by 1 recursive learning platform that we've been building over a decade. And the fact that we're doing that with better productivity, as I mentioned, 40% less of our cost base, I think it's a testament to if you get better at learning, you should be able to predict more going forward and be able to validate that in the clinic. So in a nutshell, I will say the company is really transformed to a company with differentiated assets, powered by an end-to-end AI engine with repeatability that's coming through both our internal pipeline and our partnership.

Thank you. And how should investors think about the key points of differentiation of your platform? And what do you think investors misunderstand or miss the most?

I think what investors haven't fully captured, I would say, is two things. And it's because of this transformation that's happened in the last 12 to 18 months, Sean, as I mentioned. First is actually the inherent value of the assets itself. So I mentioned we have multiple programs, but two that are first-in-class programs. The first one with really high-quality proof of concept data already delivered. I think that is underappreciated today because often people are associating Recursion with the pipeline from a couple of years ago, so I think I think that's 1. The second, when you look at the progress across the pipeline and the partnerships, like I said, these are not disconnected. There's that body of evidence collectively that's being generated. And I think people are underestimating how much the platform has changed. It's not a phenomics platform. You know, when I first joined Recursion, people would say, "Oh, yes, you do some phenomics." Absolutely not. It is expanded from that to multimodal data where we can find novel targets validated by what we just did with Genentech and what we have in the clinic. We can make novel molecules, and we have evidence in the clinic, and I think that's a very different engine than what was around 12 to 18 months ago. And I would always look for that repeatability because if you have an engine, I want to know that you can do that over and over again. So I think those two pieces are underappreciated, and, you know, we have a lot of exciting catalysts coming up to dig into that more.

Awesome. I'll move on to REC 4881 and FAP. Maybe to give some investors, tying the platform and the discovery piece together. So how was the platform used to discover 4881? What are some of the properties of the pipeline candidate that make you most excited?

Yes, so thank you, Sean. This is REC 4881. This is our first-in-class oral asset in FAP. So just to answer the first question, this came, as I like to call our platform, there's three components: the biology, design, and clinical AI. This came from the biology part of the platform. It has been known for 50 years that FAP is driven by a loss of function in the gene APC. But nobody had ever studied and nobody had ever made the unobvious insight that a MEK1-2 inhibitor could actually reverse the challenges caused by APC loss of function. That arc was not known. Nobody had tested it clinically. So what we did is we looked at human / healthy / isogenic cells and we knocked out APC.

That's the root cause of the disease. And then, using our platform, we screened thousands of compounds to see which compound can actually phenotypically reverse. This is a cell morphology. Reverse the cells from disease state to healthy state. This is where we use ML models. And the top of the list of predictions from our foundation models was this asset, which happened to be a MEK1-2 inhibitor. We did not go in with the hypothesis. So think about how you do discovery today. You have a hypothesis and a mechanism, and you go test it out. Here is an unbiased, going from disease to healthy, and that's the mechanism that came up at the top of the list. But this is where it's really important. The last question you asked me, we generated our own data. We have our own data factory in Salt Lake City. Super important as you think about LLMs, frontier labs, everyone's talking about models being commoditized. What's the next frontier? Having your own high-quality data. But the next step that's really important, actually testing that in the lab. So we did our preclinical model. It looked good with that asset. Then we did our healthy volunteer. It looked good. And as Sean just mentioned, we did our proof of concept study where we saw most rapid and durable polyp burden reduction. Rapid in three months, almost. Almost half the polyps are gone. And what was even more remarkable, especially for KOLs, is when they're off drugs for three months. The effect not only persists but deepens in some patients. And remember, FAP is a chronic disease. It starts when you're 10, 11, even younger. And you're living with this disease for the rest of your life. It is the most penetrant GI cancer compared to, you can compare to BRCA1/2 in breast cancer, 50,000 patients, U.S. and EU5. The standard of care today is surgery. Surveillance and surgery. Patients have more than 50% to 60% of their GI anatomy removed. Imagine losing that. Your colon's removed, parts of your upper GI is removed, your pancreas, your gallbladder. It's incredibly invasive. So if you can get a disease-modifying drug that can slow that progression of the disease, so that you can actually reduce the surgeries, reduce the polypectomies, all of the invasive procedures. That would be incredibly meaningful for patients. And last thing is there is 0 approved drugs. So ours, just to summarize, rapid reduction in polyps, which is, and by the way, every polyp is precancerous. That's the root cause of the disease. Durable even when the patients are off drugs for three months. Oral, which we all know when you get a drug that's oral for patients, it's chronic. And the last thing I didn't mention, which has been a really exciting piece that we're learning more about, we see polyp burden reduction in the upper GI and in the lower GI. Let's take a second to answer why that's important. Even after your colon is removed, the polyps keep growing, upper GI and lower GI. The upper GI is where you have your other organs, pancreas, gallbladder, and so forth. It's really challenging because that's where you end up getting the invasive surgeries. And another thing about the anatomy of the upper GI is thin, means when you remove polyps, the risk of perforation and bleeding is higher. We're the only asset, investigational asset, to date that's shown such rapid, durable polyp reduction and also in the upper GI. So that's really meaningful from an unmet need perspective.

Thank you. We are in active FDA engagement to define a registrational pathway with updated data at CGA, IGC, and regulatory clarity expected this half. Can you provide more color on what we should look for in the second half for those updates?

Yes, great question. As Sean mentioned, there's two important aspects. One is we're sharing additional information at the CGA meeting in early November. This is the Congress, the main Congress for Inherited GI Diseases. So this is essentially our target audience for FAP. A presidential plenary to share additional information on our Phase 2 data, just some of the data I mentioned, but more information. So that's numero uno. Number 2, we are also in active conversations with the FDA. Just to remind everyone, we have orphan drug designation. And then we have fast track designation as well. So it's been a very productive set of conversations with regulators. Our base case is essentially what you see, ERAPA-like progression-free survival. What does that entail? These are the meaningful clinical events that you would want to slow down: surgeries, you know, dysplasia and so forth. That's what we're focused in our conversations with the FDA, and more to come.

Thank you. There has been a history of toxicity with MEK1-2 inhibitors, so maybe talk us through what you're seeing on the safety profile. Also, can we touch on the durability of treatment raises the possibility of intermittent dosing, so can you talk us through how you think about dosing as that potentially part of the registrational package?

So two questions. Let's talk about the safety profile. It's a MEK1-2 inhibitor. Over 85% of the AEs, treatment-related AEs we've seen so far, are Grade 1, 2. The main ones that we see is dermatitis, which is very much in line with what you see with MEK1-2 inhibitors. We've seen some transient CPK changes and so forth. Just to the question Sean asked, there are actually MEK1-2 inhibitors already approved in oncology, but there's also precedence in rare diseases. It's approved in NF1, which is a rare disease for pediatrics and for adults. And from that perspective, I would say what we have seen is this is very much in line. And from that perspective, I would say what we have seen is, this is very much in line. The dermatitis is something that we are also managing proactively with prophylactic solutions, which has really helped. And I would say even in the real world, most of the CPK has been quite transient. Drops in LVEF.

So it's very well known in terms of the profile, given both in ONC and also in rare diseases. That's something we take very seriously. The other thing I just forgot to note is ours is a QD. And we will also be looking at intermittent dosing, just given what we have seen with the durability of this drug. Some of the other MEK1-2 inhibitors that exist are primarily BID. So again, if you think about it from a patient perspective, a QD, especially for chronic dosing, I do take something on a QD basis, and I can tell you even that my adherence isn't always the best. So that's a little bit around the safety profile. Question was great, yes, on the durability, you know, we don't know exactly why, but we have some hypotheses, you know, in terms of we're doing some preclinical work, you know, when you look at polyps on the durability of response. And is there potential that a MEK1-2 is actually evolving both the composition of the polyp as well as the stroma, so the microenvironment of the polyp as well? So there's some translational work we're doing in-house. But I will say what's really interesting from an MOA perspective for this MEK1-2 inhibitor, it acts via a dual mechanism. One is, of course, the MAP kinase pathway suppresses. We call this the evolution pathway for polyps, because as polyps mature, you don't just get the APC mutation. You get KRAS, BRAF, and additional ones. And the dual part is the second part is it also has crosstalk with beta-catenin. So we think that's the other reason that is also effective. It has it has a dual hit in terms of on the polyps. And then some of the work that we're doing in terms of is there potentially you're changing the composition of the cells in the polyp itself and the stromal area, more to come. But look, what we're very excited about and many KOLs, I think this is one of the aspects they found really compelling, is that durability when the patients, not just durability off drugs, nobody's tested it off drugs, and in some cases, even deepening.

Sure. Thank you. Moving on to REC 1245 and the broader pipeline. So REC 1245 targets RBM39, which came out of your maps. How was the molecule discovered?

Yeah, the discovery of the molecule came from the same map that I just mentioned for REC 4881. This is what I mean. Non-obvious insights from the same engine starts to get interesting. And just recently, we had a new target by Genentech from the same type of engine. So this is what I mean, that you start to see the repeatability. So let me just explain for REC 1245. So at Recursion in Salt Lake City, we have a large data factory where we can manufacture millions to billions to trillions of cells. And we knock out every single gene in that in those cells, so 20,000. And you can measure not just phenomics, transcriptomics, and then for validation, we add proteomics and other assays to the list. Very different from the Recursion 12 to 18 months ago. So we're at REC 1245, looking at CDK12. I'm sure many of you know that CDK12 has been a target of interest for some time for DDR modulation and so forth. But it's been hard to drug because of the similarity to CDK13. So we looked at what are other potential entry points by looking at all of the other genes that we knocked out that have potential shared biology to CDK12. That's when RBM39 came out. That was a non-obvious insight. And the next thing that we did was we rapidly, within 18 months, designed a novel degrader. Then we went into the clinic, and what we have so and by the way, there's two big paths, I like to call this asset, potentially multiple doors that you can open. One is the genomically unstable tumors, think about heavily pretreated patients with solid tumors that are genomically unstable. The other is pediatric, you know, transcription tumors, like Ewing, for instance. So the beauty of it is, I think there's a potential, and you know, we'll let the data drive it for us, monotherapy, if we see increasingly high dependency in RBM39, you can actually put those cells over the edge monotherapy, or potentially combo, you know, think about the PARP, PRMT5. So there is multiple options, that's it, that's why they address patient population, about 100,000 patients is quite high, but this is a novel target, same engine that I talked about, 4881, same engine that's creating the novel targets for Genentech, and the design platform that made this molecule, same engine as what we're doing with Sanofi, where we've gotten five milestones to date, another important one up in I&I. So that's how we had discovered the target. And at that point, there was not this connection between RBM39 and CDK12 and the pathway and mechanistic understanding. So it's not just a novel target. Is going is once you understand the target, you actually get a better understanding of the patient population you'd want to enrich. That's powerful before you dose a single patient. We've been talking about it for the long time, but we're making some good headway there. But we have more data coming on that later this year.

Sure, and you've got other programs, multiple programs advancing. How does the data set look or the news flow look over the next six, 12, and 24 months?

Yes, I mean, we have multiple other programs. For instance, we just initiated our PI3K1047 program, DASP. We have IND enabling programs in ENPP1. We have LSD. I think the way, if I step back, Sean, I would think about it as, we are constantly looking at our programs. I mean, because we have a repeatable engine, we have multiple shots on goal. And like we did last year, we continue to look at the programs and say, where do we have the highest conviction? And double down. And we already have some great examples to pick from the multiple programs that are coming up. All of these programs and some sort of catalyst coming up in the next year.

each of us. Sure, thank you. Moving on to the partnership, so Genentech has advanced the first neuroscience target from the collaboration, previously unexplored target into joint small molecule discovery. What does that option mean economically, and what triggers the next one?

Yes, I mean, I'll start, and, you know, Ben, if you want to add to this as well, I think for us, as Sean mentioned, our partnership with Genentech is focused in neuroscience and in GI oncology, where it starts by generating these novel maps of biology, which Genentech has optioned two already for a lot of people, about $30 million each. And then the next step is how can we take this rich, novel, proprietary data, predict which targets might be causal, and then validate them back in the lab. The first of those was actually just accepted, as Sean mentioned, by Genentech. One thing to note, for those of you who are watching, whether the scaling laws of biology and AI will make a difference, this is one of the first examples of that, that you can actually find non-obvious biology that is causal from data that nobody else has in the world. But, Ben, I don't know if you want to talk about totality of what we brought in and also to Sean's point with the selection.

Absolutely. Absolutely. So, as you mentioned earlier, we've brought in $525 million in upfronts and milestones from our partnerships. We've hit over a dozen milestones. This was another key one because it was also advancing the biology side of the platform. What we're doing now is moving beyond the maps and actually translating it into a pipeline of neuroscience programs that are all based on completely novel targets. So what we've looked at now is the first milestone coming in for that target validation. We'll move through small molecule validation. It's important to remember we run all of our partnership business to be at least break-even or profitable from the beginning. And so what we're trying to do is really use it not only to build our platform, but also build a lot of NPV value where we continue to expand the pipeline using the scale and the technologies that we have, and then have a really nice profitable back end to all of those programs.

So maybe just two things to note. One is for each program, the milestones are about $300 million with single digit royalties for Roche. And not only has the target been optioned, but we're actually working on developing the small molecule for the target. So the journey continues beyond optioning the target. The second thing is we expect potential for more targets from the same data set. Now remember, if you think about the internet corpus of data, that's really what's available. And that's what was used to build a lot of these LLM-based companies. Guess what? That doesn't exist in biology. That doesn't exist in science. This is why having a data factory is so differentiated for Recursion. Again, not just having a data factory, but having the wet and dry lab connected so you can validate and see if the stuff is causal, and then actually design the molecules and take it to the clinic. So the fact that we have connected it end to end, and then we have the data factory, means these maps are reusable. That's the point I'm trying to make. It's a durable asset, because you're mining it over and over again, find novel targets, and expect to see more of that.

Sure, sure. Maybe just to clarify in my mind, so the neuroscience maps used whole genome CRISPR knockouts

in iPSC-derived neurons. Is that data set reusable across other partners, or is it exclusive to Roche?

It's exclusive to Roche. We have a 10-year exclusive neuroscience partnership with Roche. I mean, kudos to Roche and their phenomenal partners, just like Sanofi. These are not, we don't do volume partnerships. We do very integrated partnerships because again, we're doing things no one has ever done before. It's not a rinse and repeat play. This is a complete frontier play. And so yes, it's exclusive with Roche.

I think, important note on that, so Roche has paid. They've provided us over $215 million so far for the program, but we actually own all of the data. So through that partnership, they only have access to the work product. And to Najat's point, we do have another five years on the partnership and hope to continue it. They're a fantastic neuroscience partner.

But we are actually keeping the in-house capabilities and data that come from it. I know the Roche Genentech partnership, which is focused on novel biology and now moving into programs. Awesome, thank you, Ben. You notched a fifth Sanofi milestone. How should we think about the Sanofi collaboration going forward and what are you focused on in that collaboration?

For Sanofi, it's focused on developing molecules, small molecules, for intractable I&I and or oncology target. So, Sean, as you mentioned, the five milestones we've achieved so far is for lead series, for making those compounds against those intractable and first-in-class I&I and oncology targets. The next one that we've given guidance, potential one that we've given guidance for, would be a milestone at development candidate for a first-in-class oral small molecule program. That would be the most mature one. I think that's, as a standalone from a value proposition, it's a highly valuable asset, but there's also a read-through in terms of the design part of our platform. Just to remind everyone, we don't design everything small molecule degraders. That's what we are focused on, because we think that the activation energy to figure that out. It's pretty high, so that's why the competitiveness is high. And if that is optioned a development candidate by Sanofi, which we've given guidance for, that would then enter their pipeline. And this is what Ben was saying. Then it becomes we don't have operating costs and we start to recoup the milestones, again, just to reiterate, the milestones The Sanofi are over $300 million, and they have team level royalties as well. So these are meaningful milestones, but I want to underscore for a company like us as we're building our engine, I keep talking about everything you've heard about. The Roche Target, the clinical drugs that we have in our pipeline, it's all coming from the same engine. And the recursive learning from the engine is incredibly important. So these partners have been phenomenally helpful for us to not just have cool data, but to actually learn how to turn the platform into an AI-native product or asset engine.

Sure, thank you. Maybe onto one of my favorite topics, you know, the platform and the AI question. Yeah. Yeah, so you cite roughly 330 compounds per development candidate against 2,500 industry-wide, and the target to candidate in 1 and a half years there, there seems to be, versus four, there seems to be consistency in the industry on that metric, but that's a productivity claim. Can you walk us through the evidence that the molecules are better rather than just cheaper?

Yes, I mean, look, quality matters the most, and then faster is a nice, it's definitely a nice couple. Here's what I would say, what we just talked about right now. I mean, we have brought in over of the $525 million in upfront and milestones, $125 million of that is in milestones. I think we have one of the highest milestones achieved for any of these AI-native drug discovery and development companies. That matters because upfront is promise, milestones is truth. So it's the quality of the molecules, for instance, with Sanofi, for some of the other work that we're doing, that's a signal. The other thing I'd also say is the work that we have in our programs that we already talked about. So this is what I mean. Like, look, we're learning as we go, but we're building this body of evidence across biology, chemistry, and the clinic. And just to reiterate the speed part of it, speed... This is important because if the thesis is that you can have high-quality data that generates better and more performant predictions. Then you have to validate only a few to get to the right answer. We believe that's true. You see that in every other industry. Our timelines are about 3x faster for small molecules, which are hard, than industry. And we are physically making about 90% less molecules, physically, because the ones that we're physically making are closer to the end point of the desired state. I think that's a velocity in the engine that I would, again, by itself, you don't claim victory. At the end of the day, the drugs have to work in the clinic. But I always say these are green shoots to watch.

Anything else? Yes. The only thing I'd say. One of the really important elements is, even though we use AI, we actually create physical things. Who can take all of these chemistries into the lab and test them against other chemistries that have been developed for the same targets? So there's a very quantitative analysis that you can do to say, is this exceeding what has traditionally been done or not? And so we've had more than a dozen different development candidates, as Najat mentioned, we've had external partners that have been validating it. We're seeing really powerful output from the design platform as well as the biology platform. And you know.

Just like everybody's using some sort of LLM, it gets better the more you use it. That recursive learning is actually really important. So do we get everything right? No. I mean, but look at our success rate in industry, 10%. You get to 20, you double it, right? So that recursive learning, it would be hard to do that if you just do novel target biology. You have to make the compound and then you have to go in the clinic and you have to learn from really fast. And I think that's what our focus is, do all of that to unlock a new TAM that's why the first-in-class programs matter, whether it's our own pipeline, or with Sanofi, or with Roche, and do it fast so you can learn fast. You can start things, double down on things that work, and what doesn't work, learn from it really fast. That's how we're trying to collapse that engine and the timeline. But the reason is to improve that recursive learning.

Sure, thank you. The model's described as the data factory, the lab in the loop, and the differentiated pipeline. Frontier models, as you just mentioned, keep improving. Available to everyone, which of those three is the hardest to replicate?

Look, I think the hardest to replicate is actually integrating it together and having an end-to-end engine. There are so many steps that go into, like, biology all the way to the patient and showing data. I think unless you learn fast, you're not going to bend that probability curve that we all want to bend. But in order to do that, Sean, I would say the models are great. Model architecture is important. But unless you have the right data to teach it, and then the predictions go back into the physical world to experiment and learn from it, what's working versus not. I call it almost the patient in the loop because it's not even just the lab because we're going all the way to the clinic. But the data factory is critical, especially in this space. Because unlike some of LLM and others, LLMs are not going to be sufficient for science. You do need other modalities of data. And we have proprietary protocols for the cells that we're making. We make like a trillion cells, iPSC-derived neuronal cells. I used to be in lab, didn't even want I want to make 100. I mean, it's super hard. I used to just work with HeLa and CHO cells. These were so much easier. The level of sophistication in our science data factory, amount of automation that it takes, we've been at it for 10 years. And there's something to be said about that, because you really become much more mature and seasoned on what works versus not. And then now we have the output to show for it, not just data sets, but actually novel targets being optioned by really high-quality research organizations.

Well, thank you, Najat. We've just run out of time. So thank you, Najat. Thank you, Ben. I appreciate you participating.

Thank you, Sean. Thank you so much. Thank you.

This live transcript is auto-generated without human intervention or review.

Hi, everyone. I'm Jill Hall, Head of U.S. Small and Mid-cap Strategy here at BofA Global Research. So, I just wanted to welcome everyone to our day 1 of our virtual SMidCap event, excited to hear from corporates across the small and mid-cap space across sectors, great breadth of coverage here by our analysts. They cover about 1,000 small and mid-caps in the U.S.

So, I'm happy to bring nearly 20 companies today and feel free to reach out to me if I can help with the schedule or getting you signed up for any additional sessions, or if you're interested in broader small- and mid-cap research, or some of our compilation e-mails we send out on the fundamental side.

So I'd like to pass it over to Alec for this session to introduce the company.

Perfect. Thanks, Jill. As Jill said, my name is Alec Stranahan. I'm senior analyst covering biotech here at BofA. I cover around 30 stocks ranging from $40 billion all the way down to $400 million in market cap. And one of the more interesting names, I would say, is Recursion, and it's my pleasure to be joined by Ben Taylor, who is Chief Financial Officer and President of Recursion U.K. And I would say speaks just about as well about the science and AI aspects of Recursion as he does about the finance side. So Ben, really happy to have you with us.

Thanks, Alec. I always appreciate the intro and the conversation.

Yes. Yes, great. So as Jill said, I'm going to run through some questions here with Ben in a fireside format, but hope to keep the conversation topical for those dialed in. And if you do have a question, please utilize the raise hand feature [indiscernible] or you can e-mail me separately, and we'll get your questions asked. Ben, maybe just to tee things up for investors maybe newer to the Recursion story, and maybe AI and drug discovery as a whole, what is that? Why is AI needed in the drug discovery process? And how is Recursion maybe blazing the trail here?

Yes. And I think it's really good to level set because AI, especially currently is waved around like it's a magic wand, which it absolutely isn't. The way that we really think about it is it's a better analytical system. So, it's more similar to the evolution of starting to use computers or starting to use spreadsheets, how those things have changed the way that we do business and look at data analysis. I think AI is another step up in that.

And so, where we have been able to apply it to drug discovery and what makes us different is it allows you to both produce and analyze data in a different way than you ever could before and also do it in a more multiparameter way than you've ever been able to do it before. And so by putting together the data with modeling systems, with the ability to compute, we can really get to a different outcome than was historically possible.

And so, the foundation of the company is actually around changing the probability of success and really trying to unlock new parts of biology and chemistry, rather than efficiency. But we have also been able to do it much more efficiently, and we published some of the statistics on that, showing pretty dramatic reductions in the time and cost to be able to get to those differentiated outcomes. So, there's a lot of different pieces at play.

I think the nice thing about being in our shoes is we're actually at the point where we have multiple clinical programs. We have partnerships that have been running for multiple years. And so, you don't actually have to understand all of the AI just like people didn't understand drug discovery and biotech for many, many years. You just have to understand the output of it. And so that's really what we're focused on.

Okay. And maybe along those lines, and this is a question I get asked a lot, which is like -- for those paying attention, ChatGPT was kind of -- you could see it coming. But if you were -- most of the population, it was just like one day, we didn't have LLMs and the next day, we did and the world feels like it's changed. Is there a ChatGPT moment in drug discovery? Like is there a moment where the entire industry you think will just one day be AI in terms of the back end on the drug discovery side? Or is it maybe a little bit different of a situation?

No, I think it's absolutely the same. But rather than it being the entire world sort of figuring it out at once, I think what you've seen is more layers. I mean, if I go back to the days when we were a private company, literally no one, large pharmas, the investor base, no one was really using AI to evaluate drug discovery or try and build some of the models that we're doing now. And now, you look at pharma, there have been multiple large pharma who have announced $1 billion-ish investments towards building out AI and investing in that.

And that's because we're actually getting better results over and over again. It's repeatable. It's not just one-off. We're doing things in a better way. We're more efficient and achieving things. I mean, all of our partner milestones, we've achieved well over a dozen partner milestones. All of those milestones where they paid us millions of dollars were things that they couldn't do internally. And so it was seeing us demonstrate you can actually get to that different outcome.

And so I think that there's still another level of having it be more generally accepted. So inside of the industry, people are absolutely using it. It's funny, some of the biotech companies that are coming up now, they don't talk as much about the AI because it's such a hype thing to talk about in biotech. But the outputs that they're doing. If you look at Prevail, I mean, there's a lot of AI that was involved in how they achieved that outcome, and it was a great outcome.

So, I think on the industry side, it's already occurred. I think on the investor side, there's a lot of speculation about when, and how, and who. And so we haven't quite got to that moment yet, but hopefully soon. I mean, really, on the investor side, I think it's a matter of demonstrating that the products really make a difference. And we're very cuspy on that, it feels like.

Yes. Yes, I agree. Are there maybe 1 or 2 examples that you think are kind of proof points for how the AI-driven approach or what Recursion is doing specifically, can produce better medicines? Or is it really in the clinic? Or is it maybe just all wrapped up and fit into that?

Well, it's funny. Hopefully, you'll respect this. Coming from a data-driven company, we don't like anecdotal examples. And so we really look for an accumulation of evidence that something is changing. And so, how do we know if our biology AI is working, Like being able to target new ideas that weren't in the literature, they weren't commonly known. And so, now we've seen multiple examples of that. Like 2 clinical examples for us with 4881, which we'll talk about later and 1245, so in FAP with MEK1/2 and then with RBM39 as another target for DDR and other areas. Those were novel biological insights.

And then what we just saw with the Roche collaboration milestone is not only had Roche opted-in on the biology maps that we created around neuroscience, so neuronal cells and microglial cells, but now we've started to take completely novel programs from those maps and transition them into the design phase. So, that's actually saying this is something that wasn't in existence as a neuroscience target or known biology, and now we're transitioning it into something that can be a drug because we've done the target validation work on it and experimentally validated it.

So, I think those are 3 points that all point in the same direction of finding new connections and biology that doesn't -- that didn't exist before. One of the things that always blows my mind, if you look at all of the drugs that the pharmaceutical and biotech industry have created over their entire lifespan with all of the good people and all of the money that we put in, the approved drugs only cover about 3.5% of the genome. If you add on all of the drugs that are currently in development, we think of this massive pipeline of drugs that are coming through and all the innovation that's going on, you're still only covering about 13% of the genome.

And so the reality is we keep digging in the same holes. And so what we want to do is create new data, look at it in new ways so that we can actually break outside of those holes that we've been digging over and over again and really find new paths. So that's why the biology side is really exciting to us. I think we can also talk about the chemistry side. I mean, we've had many milestones with Sanofi and other partners advancing through, as well as being able to demonstrate how our chemistry is actually achieving things that other chemistries have not. I think we'll hopefully be turning over cards on the clinical side on that soon, but it's been exciting to watch both of those come together.

Yes. No, that's a great way to sort of sum things up, I think, for the state of the industry. And we hear more and more the actual generative AI models are not necessarily the big differentiator. There's still maybe an edge to be gained on compute, and you guys have your supercomputer in-house that you've already built out. But it's really more and more, it feels like the data side that's.

And then kind of how you pump that through the funnel internally and then spin the flywheel based on that data and generate new data from quality foundational data sets, that seems how you build a good platform. Maybe you can just talk about that piece, sort of how Recursion's built that kind of from day 1 and sort of where you see that unique data set being leveraged either internally or through your partnerships?

Yes. Well, and it's a really great point because if you think about why haven't we gone into more of that 87%, where there's literally not even a drug candidate out there, much less something approved. I mean a big part of that is because you can only go after what you have some sort of diagnostic or assay system or some way of understanding what good looks like.

And so this is where coming in and being able to create and look at data in new ways makes such a difference because if you're just focused on the algorithm and you're just trying to look at the data that's already in existence, you're going to get a lot of the same answers. You're just -- you're certainly not going to have differentiation from other people who are doing the same work. And so you need to really be creating novel data to be able to look at it in different ways and analyze it.

Most of the data that's created across the biopharma industry is not good for machine learning because it's been created in a format that's usually very local. It's usually done for a single project. The annotations are quite different from project to project and you can't use it as fluidly. And so that's where starting from the beginning, I mean, more than a decade ago, we have been creating novel data in sort of a machine learning annotated format that we can then consolidate down.

And so that's where we get to over 50 petabytes of having that data in existence that's very differentiated from what is in existence in other places. And if Najat was on, so Najat, our CEO, who used to head up AI at J&J, looked at it and said, in all of the data inside of J&J, probably only about 30% of it was really something that could be used for Machine Learning, and even that had a lot of data wrangling.

And so this is where trying to build up unique ways of looking at the data can get you down into really new spaces. And I think that's where a lot of the excitement on our side and as well as across the industry is how can I analyze this biology signal in a different way, whether it's phenotypically, or applying -- overlaying cellular imaging with transcriptomics and proteomics and other sort of sets, you get this much richer set of data.

I think last point, this is also a place where we've really felt the differentiation value of just having a lot of the breadth that we have because we started in the phenotypic cellular imaging, like that was the original base, and it's sort of a new language for being able to look at biology, which is incredibly powerful. But then being able to overlay that with orthogonal data sets like the transcriptomics I was talking about, proteomics, or looking at real-world patient data and figuring out genomic signatures.

What that allows you to do is actually synthetically or virtually compare the data output and the analysis. And so you can actually dramatically improve your ability to find real signals because no matter what source you're using, there's always going to be a lot of noise in it. And so we published a paper in Nature Biotech just recently that showed we were able to actually outperform models that were even up to 100x the data set size that we were using.

But it was because our data was better annotated and we were able to use multimodal sources to basically enhance the signal that we were seeing and make far better predictions on it. And that's sort of how you get to more composite biology and systems biology and where we hope to go in the future.

Great. Yes. I mean the saying garbage in, garbage out, that still holds true, maybe even more so today. But maybe to ground then some of the conversation that we've been having so far. You mentioned your Roche/Genentech partnership. You recently selected the first neuroscience target here from your collaboration for further drug development. So, I guess, what did that target need to demonstrate for you and Roche to be convinced that this is something worthwhile taking forward? And why should investors maybe view this decision as important validation for the platform?

Yes, absolutely. So just to take the step back on the partnership as a whole. So Roche had originally given us $150 million to go out and basically build maps. Most of that was to build maps in 2 different neurology cell lines, so neuronal cells and microglial cells. And those maps were basically whole-genome knockouts and other perturbations of those cell lines, and then looking at how did it morphologically changed, and then we overlay on some of the transcriptomics signals to understand what might be happening within those cells. So then, they optioned in both of those maps. They don't get any of the data, but they are able to query the maps, basically.

And that was $2 million, $30 million payments to be able to do that. So that's already all happened. Now basically, what the milestone that we just got is from those queries, we did a list of potential targets. And from those targets, we selected a few that we then wanted to go through and do validation on. Some of that validation is still AI-based and really how we selected the targets that we wanted to do. But most of it, we went to experimental systems.

And we said, let's test this in a disease model we know. Let's look for how this is disease-modifying to different cells of interest or diseases of interest. And so in a very classical way, demonstrate that this novel virtual finding is having real experimental impact. And so we went through a process with Roche. They are obviously a world leader in neuroscience, and that was what triggered the validation. And so now we're taking that target and recursion is designing the molecule to be able to target.

Okay. Okay. Got it. And I think one point that we shouldn't sort of gloss over is that this was a target in neuroscience that is novel, right? There's been decades and decades of research in CNS diseases and your platform and your neuro maps were able to uncover something new. So talking about digging holes in fresh soil.

Absolutely. Well, and by definition, everything that we do out of that partnership, this isn't going to be something that you can find in the literature, or that someone is doing an alternative program for. This is wholly new work in neuroscience. So, really, really exciting to see that progress, and hopefully, a lot more to come.

Okay. And maybe you could just remind us of the structure of the partnership with Roche. How many candidates could you bring forward? What are sort of the economics around bringing those forward? And then maybe you can talk about how that structure is maybe similar or different from your Sanofi partnership as well.

Yes, absolutely. So the Roche parternship originally a 10-year term, which obviously we can extend and we were about halfway through it right now. I almost feel foolish stating the number, but it's up to 40 design programs that we can advance and actually part of the rationale for the Recursion Exscientia merger, which is now almost 2 years ago, was to bring the design capabilities in-house with Recursion target ID.

And so it's actually really exciting that Roche wants Recursion to do all of the chemistry work and direct design work because prior to the merger, I don't know if that would have been true because we really have built out the capabilities. And so it's great to see that coming together. Now, what we hope to do is basically create a pipeline of additional targets coming out in neuroscience. We also have work ongoing in some GI oncology as well, and to be advancing that as a long-term partnership with Roche.

The Sanofi is a little bit different in that there was not a target-ID portion to it. It was -- this was a legacy Exscientia deal. So it was really focused on, hey, there's a target that we mutually are interested in. No one's ever been able to drug it before can we advance a candidate in it. And so now we've already seen 5 programs hit their first discovery milestone there. And so the next milestone for all of those would be basically opting in for Sanofi to take it forward into clinical trials.

And that is really exciting because that not only marks that we'll be advancing, again, really exciting new potential blockbuster programs, but also that it ends our operational obligations. So all of the payments from development candidate onwards are basically profit for us. And so if you look at the Sanofi collaboration, each program has a potential for up to $343 million in milestones, $193 million of that is pre-commercial.

So this is not some massively back-end loaded deal. And then our royalties on it are average in the low-double digits. So we actually capture a pretty substantial part of the NPV. Both of those programs and the Roche design elements are actually similar to Sanofi, not quite as high on the economics, but close. And for both of those, they're really designed to be more of collaborations, but ones where we are always at breakeven or profit on a direct cost basis. So Sanofi and Roche and our other partners pay us ahead of time for our expenses. So it's a really capital-efficient way for us to grow value.

Yes. That was an important point that I think you also mentioned on your 1Q call about the cost to service these partnerships, and that's a question I've gotten. So it's good that they're designed to not be a burden on that. And we've seen in the space, a lot of different approaches to monetizing these AI drug discovery platforms. You've got like the Schrodinger in that world that are more like a SaaS type revenue model. And then you've got like in Insilico's, which are maybe kind of a mix.

And then you guys are more of like a hands-on, let's do interesting science together and leverage the platform to push those forward, but it's a little bit more hands-on, but you also get larger, chunkier deals out of that as well. So maybe you could just talk about kind of the philosophy around the partnership model and how you balance that with in-house development.

Yes. Well, and it's interesting. So if you go back to our original mandate as a company, there was really 2 parts to it. One was how do you change the probability of success using technology, right? Like we're in a 95% failure environment. No one is making data-informed decisions not because they don't want to, but because they can't. The data is not good enough, the models aren't good enough. You can't make good predictions. And so that's how you get to a 95% failure environment. So please improve on that.

The other part of it was how do you make this into an actual business model rather than just a binary risk bet. And so that's where our ability to do things at scale more efficiently really comes into play. And so you can almost think of the partnership business as an outgrowth of that. So from early days, we decided having our own therapeutics was really important because it allowed us to demonstrate that the platform is working and also we were creating a massive amount of value, so being able to capture it as we get into those points. And that's where all of the upcoming clinical data is so exciting because those are obviously massive potential transitional points for us.

The partnership business, though, is a beautiful part of being able to fund the company, build the platform and grow the long-term NPV really well. And it makes sense because we do these things at -- or we can do things at scale. And so we would actually be leaving some of our capability dormant if we didn't have the partnership build. So the fact that we can get paid early on, use that to actually do a lot of applied development. So one thing that most people don't know, about 65% plus of our budget is actually applied.

So even when I'm talking about platform and technology development, like we're doing it on real programs. And so that's part of our edge, like we know if our models work because if they don't, the drug doesn't get made, right, or the biology doesn't work out. And so like the partnerships is a great applied platform for us where we build out our platform, test our models and be able to add that in as a part of the overall product engine.

And so we've always loved it. We don't want to become a service company that's a different set of economics. It's a different business model. There's a lot of infrastructure. In fact, how we run our partnerships is basically exactly how we run our internal programs. We just have a partner that we're strategically working with and talking about what good looks like. And so that's where we differentiate in our model from the more service-oriented side.

Yes. And I'd say your in-house pipeline is also a differentiator for you guys. And you're doing quite a lot across oncology and I&I, other areas. So maybe we can talk maybe for the next 10 minutes or so on the internal pipeline. Maybe starting with FAP, Ben, just because that's -- I mean, lead asset, you can bounce around, but it's the furthest in development, and we've got maybe more here. So maybe talk about that program, sort of the origin story and then the disease if people are familiar.

Yes, absolutely. So FAP, if you're not familiar, just a quick background. more than 50,000 patients U.S. and EU5. That's probably underreported because that number is about 70% hereditary, but it is possible to get this through somatic mutations as well. And so a number of those patients may just get caught at the colorectal cancer stage rather than when they actually had the FAP. So, but 50,000 very large for an orphan indication starts with an APC mutation that basically leads to chronic, cancerous or malignant, however you want to say it, polyp growth.

So none of those polyps that are growing are benign. They will develop into cancer eventually. This typically starts in the colon, but it actually spreads throughout the GI system. So down into the rectum and up -- can even reach up into the stomach. And so as this patient progresses through decades, they basically -- they'll typically have a colonectomy in their -- colectomy in their mid-20s and then we'll be going through basically surgeries, excisions to be removing those polyps throughout entire life. And so they can come in and see their doctors several times a year, if it's a serious case to be investigating how the polyps are growing and have them removed.

And so there's, on average, about 10 major surgeries for these patients throughout their lifetime and about 70 treatments with excisions. So it's just a massive surgical burden and quality of life burden on these patients. So what we were able to demonstrate is that we were able to bring down within 3 months, a little over 40% reduction in the polyp burden. And so that's a combination of size and number of polyps. Some of these patients can have hundreds or even thousands of polyps spread throughout their GI tract.

And so that's a massive change in a short period of time. What was also really exciting, we're the first drug to ever show that you could take the patient off drug and maintain that response. So we actually had a slight deepening of response in the data over the 3 months that the patients were off drug, but that is very, very exciting. And so more to come on that. We are currently in discussions with the FDA on the pivotal trial design, and we'll give an update on that later on this year. We are also presenting at a conference most people probably have never heard of, but it's the one where all of the docs that care about FAP go to, and we're presenting at the presidential plenary with the data.

Can give us more detail on why they're so excited about it. I know you're going to ask about FDA trial design because everyone does. The short story is going to be we can't get in front of the FDA, and we're going to let those talk through. But we know from the existing trials that are out there, which is one, there is a baseline that would be acceptable that we can move forward with. We think even if we don't change the design endpoints at all from that, we'd still be able to drive better enrollment through our ClinTech platform, which we can talk about as well.

So just to summarize, it's a large indication, no approved medicines. It could result in cancer if it's not treated. And even if it is treated with repeat surgery, oftentimes, patients still get cancer. So -- and if you have thousands of polyps, like how are you going to surgically excise those, right? So it really makes sense for a drug to come in with a systemic mechanism of action to come in those.

And you have shown, I think, 43% after 12 weeks of treatment and then 53% speaking to that deepening. I guess thinking about the conference presentation on November 2, what do you think we'll learn there in terms of is it longer follow-up, maybe individual patient data? It sounds like a great platform for you guys to garner excitement for enrolling a potentially Registrational Study, right, and activating site with the investigators.

Yes. So a couple of different pieces. At that conference, we'll definitely be providing longer follow-up, which is exciting as well. So the first cut we did was 3 months of treatment. And so being able to look at these patients over 3 months treatment on and 3 months off. Most of the drugs that have been attempted in this area, and there's not a lot were over a 12-month period and showed much less of a response than we did. And so being able to talk about the durability of it, I think, is important.

Also with looking at different ways to be able to manage some of the known side effects of MEK1/2 inhibitors, which appear to be very manageable, but we'll give you additional data on that at the upcoming conference. And then a really important point, we have effect both in the upper and lower GI. So the one other drug that is currently -- it's called eRapa, it's basically encapsulated rapamycin showed a little under 20% response after 12 months. But importantly, it was all in the lower GI, and there was no effect on the upper GI.

Well, the upper GI is actually where you get a lot of the more serious polyps as the disease progresses because you can't do that with a normal colonoscopy. You actually have to go through a more invasive endoscopic procedure to be able to monitor and excise the polyps in there. And that's also what can end up leading to things like a Whipple procedure where they're going in and obviously taking out a really damaging amount of your internal organs. So we were really excited, and we'll have more detail on it to show a very similar response in both upper and lower GI. And this is based on some of the PK properties that we had selected the drug for, which gives it a real advantage.

Okay. And then when you think about -- obviously, you need to iron out the details with the FDA, what a Registrational Study could look like. But what are sort of the main takeaways you expect to get from those conversations? Is it sort of around whether polyp reduction could support an approval, kind of the length of follow-up? And do you have a sense of what a reasonable comparator would be? Is natural history sort of used in this indication, given there really are any approved medicines?

Yes. Well, it's -- what I can say is we've had productive discussions with the FDA. No one's ever gone in with either the depth of response or sort of the natural history data, to your point, that we've been able to show. So we actually -- this is another place where we can put our resources to use to be able to look at data in a new way.

So in a very short period of time, we were able to create an LLM model using over 250,000 patient records to look at clinical practice, what is standard clinical practice of patients with FAP, what are doctors actually doing? I mean this includes all the physician notes and being able to query that and saying, okay, in this situation or with how patients progress or when do they get surgeries, those sort of things.

And so this is new information that has never been seen before. We also created a natural history database with one of the universities in the Netherlands that tracked FAP patients over 20 years, and we are able to show that these patients do have spontaneous or they do have continuous annual polyp growth. And that averages in the north of 50%. And so, there's real polyp growth that continues to happen for these patients. So, we're going to go in there, but I do want to set the base even if it ended up the exact same trial design as the one that had been approved, we'd be fine.

And what we're able to do with our ClinTech platform is we've shown increases of 30% to 60% in baseline enrollments, and we've put out some of the literally site level statistics on what we were able to achieve. And we do this by starting with the data and finding out where the patients are and driving the CROs rather than having the CROs drive us. And we've just been able to achieve different things in our clinical trials. And so we would focus on driving enrollment and getting to the right patients and understanding the right patient populations to get that trial done in the best way possible. So we'll see. Maybe there are endpoint changes, maybe there aren't, but we feel good about it either way.

Okay. Yes. And I'm glad you brought up the ClinTech platform because, as we know, the bottleneck, even with all the AI tools we have continues to be clinical trials in terms of getting drugs to approval. So, if there are ways to speed enrollment, or even recalculate your powering assumptions to decrease the sample size that you need? I think these are all things that you guys are doing and applying to your studies, right?

Yes. I mean we use real-world data literally on every single one of our programs. And we build AI models around it. We run simulated trials. We're doing all of the data science around the clinical aspects as well. And it does really make a difference. I mean we understand our patients much better and which of our inclusion/exclusion criteria really matter and how it could change, what sort of drug interactions we have to be most focused on, all of those sort of things, we're able to just get a much higher level of understanding for our patient populations before even starting the trial.

Okay. Yes, that makes sense. Maybe turning then to a couple of questions around your oncology portfolio. You're bringing PI3K-alpha towards the clinic. I think Go/No-Go is targeted for second half of this year. You also have RBM39, which is another asset for solid tumors. And I think we'll get second half dose-escalation data as well from there. So, I guess maybe just talk about your efforts in oncology and sort of what's getting you most excited from the emerging data.

Yes. So let's start with the RBM39 program because that's -- it's a really exciting one. This is one where it's basically we looked at a transcriptional target, CDK12, that the industry has always wanted to drug, but it is actually a really poor pocket to be able to drug. And so, we just took a more phenotypic approach to understanding the biology, and what we found is RBM39 actually results in a very similar biological change to inhibiting CDK7, and that sort of led us down a path.

And we ended up creating a novel degrader to be able to go after it. And this is a first-in-class that we brought into the clinic. Really exciting early PK/PD profile coming through, very large potential patient population, I mean, this is one where you can think of it sort of like a next-gen DDR drug, right? So certainly, if you've got a genomically unstable cancer, this has a lot of potential from a mechanistic point alone or if it's combined with drugs that cause that genomic instability, it could be really exciting.

And so more data on that soon, second half of this year. We'll give an update. And that's a program we really like both for its novelty, but also for its potential and where it could go. As far as the PI3K inhibitor, it's funny. We get asked the question a lot like why in the world would you do another PI3K. There's a lot of them out there. And this is interesting. It's one of the -- so if you look at a number of programs that were in the pipeline coming over from the more design-oriented side with the Exscientia.

PI3K, we looked at it and saw things like hyperglycemia being a dose-limiting side effect for a lot of the PI3K inhibitors. And so we wanted to create something that had far more selectivity so that you could go far deeper because that hyperglycemia is caused by the wild-type inhibition. And it has 2 negative effects. One, obviously, hyperglycemia and the patients can be very limiting on its own. But second, there's a lot of theory that, that actually causes the tumor to grow more quickly.

And so what we wanted to be able to do is really knock out that signal so that you could drive maximum potency on the H1047R mutants and really get deep into it. So it's a good example sort of like LSD1, like MULE1, like CDK7, while, CDK7 is a little bit different, but like those other 2, where there's a single really clear side effect that if you can remove that side effect through chemistry, then you have a potential of really unlocking that target class in a new way. And so that's the goal. It's actually -- even though the second-gen PI3Ks have shown less hyperglycemia, you still do see the hyperglycemia.

And in fact, those hyperglycemia rates are significantly higher if you're prediabetic and they exclude the diabetic patients. So there's a diabetic and prediabetic orphan population. But even in the nondiabetic patients, I think that you're getting into some limitations in being able to push dose because of the side effects like hyperglycemia. That's not the only one, but that is obviously one that will cause a lot of focus.

Yes. Yes. I mean even the next-gen, even if they have less hyperglycemia, they have other AEs like I think stomatitis is a big one as well. Does your molecule avoid that?

Yes. So we haven't seen any of the signals so far. But obviously, we're -- so we're IND-cleared and should be starting the Phase I soon, and we'll get a signal, but it looked very clean in the preclinical that we ran, and we didn't see -- so we're getting more than an order of magnitude more selectivity over the other compounds in the space. And I think it's also a nice example. There's a lot of me-too chemistry in the space and you just don't see a lot of differentiation. It's like a little bit better.

But if you want to get to an order of magnitude plus differentiation, you really have to be able to look at novel chemistry and go down different routes and explore the space in a different way. And so that's another reason why it's a nice highlight to we can actually go to places the rest of the industry can't because of what we're doing with AI.

Yes. Right, right. I want to pause for a second and see if there's any questions from the line. I'm not seeing any raised hands, but operator, if you could maybe read the instructions for asking a question, and we can see if anyone has one here in the last 5 minutes or so. I think operator, is maybe on mute. But I'll just say that if you do have a question, feel free to e-mail me or I think there's a raise hand feature as well, and I'll be tracking that. Ben, I do have one question, and I think I would be remiss if I didn't ask the CFO a cash runway and capital allocation question.

No, look, this has been a core focus for Najat and I and the whole organization. I mean if you look at our pro forma premerger expense to today, we're looking at about a 40% reduction. And honestly, we're not really doing less -- so we have eliminated a few of the programs that just looked like they weren't going to have impact. And that's really a sign of how we've gone through the entire budget. It's what can we see a clear line of sight that this is going to make a difference. And we do that for everything. We even do that for G&A, right? Like this is across the entire company. And if we can't see that clear line of sight, then we don't do it.

And if we can't, then we say, how can we optimize this work process? How can we ask the hard questions first? How can we get the data that we need and do this in the best way possible. And so we're just doing a lot more with a lot less. And so that's been really exciting to see. So we ended the quarter with $556 million in cash. And we expect that gives us a runway at least to early 2028 without any additional financing.

And so, we're in a good place to turn over some of these data cards that are coming up, 4881, RBM39, we've got good partnership milestones that hopefully add cash as well. And hopefully, we get across the line with our first Development Candidate with Sanofi, which I think would be really, really exciting, both in the program, but also in that partnership. Yes. I think we've got a lot going on, but trying to do it as efficiently as possible.

Yes. Yes, I think that's the name of the game and certainly something that AI is and Recursion specifically has been built from day 1 around. Maybe just a final point to end on here, Ben. Just to kind of focus investor attention over the next 12 to 18 months, what do you think would represent the strongest proof that the Recursion platform is working? Is it kind of the clinical efficacy on the in-house molecules? Is it progress with your current partners or new partnerships? Where do you think is the greatest opportunity for sort of validation over the next year or so?

Yes. And I'm going to put it into the investor context because I mean, look, we're publishing papers that benchmark top of industry across multiple different areas in AI drug discovery and development. So I feel like that's well validated. We've hit a whole bunch of milestones. So I feel like our partners are saying our technology is doing things that they couldn't do and that add value.

So you really come back to people looking for clinical data. I think a majority of people who follow the therapeutics world want to see more of that data come through. I think 4881 was a wake-up call for a lot of people because it was pretty compelling and biologically unexpected. And so hoping to surprise more people in those sort of ways. And I think people are paying a lot more attention now than they were before, which is great.

The other piece I'd say is just looking at the accumulation of data. So we don't expect all of our clinical programs to work. That would be crazy, but if you start to see more programs reading out positive data beyond 4881, we've got RBM39. We have CDK7 and, LSD1, PI3K coming into the clinic, some other programs that are coming up. I mean, they all start to point towards the same thing. Is the platform creating differentiated medicines because that's what makes a difference. And I think that will be a meaningful turning point in perception.

Yes. Very good. Well, I think with that, we're over time. So we'll have to end it there. But then really, I want to thank you for the great discussion for participating in the conference and for Jill, for hosting it. And yes, looking forward to all the updates over the coming months.

Sounds great. Really appreciate you having me.

Thank you. Thanks all.

Good morning, everyone, and thank you for joining us. Before we begin, I'd like to remind everyone that today's discussion will include forward-looking statements. Next slide. Please refer to today's press release and our SEC filings for additional details. At Recursion, our mission is to decode biology to radically improve patient lives. And we do this by building transformational medicines with an AI-native product engine.

Over the past year, we have reached an important inflection point. We are no longer just discussing the potential of our platform. We are demonstrating the ability of our AI-native product engine to generate differentiated programs and medicines. Just as a reminder, the engine you see on the left-hand side is built as a continuous learning system. Proprietary multimodal data created in our data factory powers Frontier AI models.

And these models can generate new hypotheses where every single prediction is tested experimentally. Each cycle strengthens both the engine and the products it creates. Ultimately, though, the measure of any engine is its output. So let's talk about that. First, our internal pipeline continues to mature. We now have 5 clinical stage programs, including REC-4881 in FAP, where we have generated some of the most promising clinical data in the company's history.

Remember, in a disease with no approved therapy and a TAM of almost $10 billion. Second, we continue to make significant progress in our partnerships while learning from the best in the industry and also while validating our engine externally. Together with leading biopharma partners, we have generated more than $500 million in realized inflows while advancing differentiated programs with Sanofi and Roche, Genentech.

So today, I'll share how we continue to strengthen our product engine and how we take these advances and are translating it into differentiated medicines, differentiated partnerships and ultimately better outcomes for patients. So the question that naturally comes up, what makes our product engine different? There are many companies applying AI to drug discovery. We believe our advantage isn't AI alone.

It's a combination of 3 capabilities that reinforce one another. First, we generate our own proprietary multimodal biological and molecular data at scale. This matters because AI can only learn well from high-quality data and much of the most valuable biology has never been measured systematically. Our 50 petabytes of data is designed specifically to train models, discover new biological relationships and improve over time as new algorithms emerge.

Second, we connect these models directly to experimentation through a Lab-in-the-Loop system, spanning biology, design and increasingly the clinic. Every prediction, as I mentioned before, is validated experimentally. Every result feeds back into those models. It is that recursive loop that helps us to move faster, improve our decision quality and systematically build confidence in our programs. And third, and most importantly, we convert these capabilities into differentiated assets.

That includes both our internal clinical programs such as REC-4881 in FAP, REC-1245, RBM39 in solid tumors as well as our partnered programs with Sanofi, Roche-Genentech. So how are we doing? Let's look at the progress we've made over the year to-date. As we look back over the first half or so of the year, I'm very pleased with the progress we're making across all 3 dimensions of our business: our internal pipeline, our partnerships and the continued advancement of our AI-native product engine.

On the internal pipeline, we advanced REC-4881 with our initial FDA engagement following encouraging Phase II data and additional Phase II data coming later this year that Vicki will talk about shortly. We have continued to build confidence in REC-1245 with early clinical safety and pharmacokinetic data. And we just received IND clearance for REC-7735, positioning it to enter the clinic later this year.

At the same time, our partnerships are also making progress. As you'll remember from earlier this year, we achieved another milestone with Sanofi, our fifth to-date on developing a novel lead series for a very challenging first-in-class oncology target. But I'd like to pause on a new milestone in particular that we're announcing today.

Together with Roche-Genentech, we are thrilled to announce that Genentech advanced the collaboration's first neuroscience target, a new unexplored target in neuroscience into a joint early discovery program, providing early evidence that Recursion's platform can generate novel biologically-validated targets for drug discovery. To me, this represents much more than another partnership milestone.

In an area where progress has been slow for decades, it provides early evidence that a fundamentally different approach, combining proprietary disease-relevant Atlases, purpose-built foundation models and that rigorous computational and experimental assays that we use to build confidence that these targets are actually causal. And of course, last but definitely not the least, the deep collaboration, scientific and technical with the partner can uncover previously unexplored therapeutic targets.

While it's still early, I believe this is an important proof point for both Recursion and the broader field. It suggests that an AI-native engine can move beyond optimizing known biology to discovering new biology, compelling enough to advance into drug discovery with one of the world's leading neuroscience organizations. So that's just the left-hand side, but we have a lot more coming ahead.

For REC-4881, we will present additional Phase II data at the CGA-IGC Conference, a premier medical congress for inherited GI disorders, our specific target audience for FAP. And we will also provide an update on our FDA interactions as well as continue advancing what we believe could become a transformational therapy for patients with FAP. Remember, nothing approved to-date, no approved therapies. For REC-1245, we are continuing our dose escalation and generating additional Phase I data, and we'll have a more wholesome update later this year.

With Sanofi, we expect the potential nomination of an oral I&I development candidate, a very important milestone that would further validate our ability to design differentiated small molecules against challenging targets with the potential to impact multiple immune-mediated diseases. And finally, we expect to initiate the Phase I study for REC-7735, further expanding our clinical oncology pipeline with another precision design program from our engine.

Taking together, these milestones reflect a company that is delivering ambitious proof points that matter while executing with focus and discipline. But equally important, we continue to strengthen the engine itself. Let me show you a few examples of how that innovation across biology, chemistry and clinical development is making our engine faster and smarter.

Let's start with biology. One of the biggest challenges in the industry is that much of human biology remains unexplored. We believe the answer isn't simply building larger AI models. It's generating proprietary disease-relevant data that these models can actually learn from. To do that, we have generated and aggregated more than 50 petabytes of multimodal biological data, creating what we believe is one of the largest proprietary data sets in the industry.

And as that data set grows, our models become better at discovering novel biology and every new discovery further strengthens the engine. That learning then carries into design. Because our biology models generate higher confidence hypotheses, our chemistry platform focuses on designing better molecules more efficiently. There's much to share here, but one thing I'll mention is we are advancing candidates using roughly 330 compounds over approximately 1.5 years.

So going from target to candidate in 1.5 years compared with industry benchmarks for small molecules of roughly 2,500 compounds over 4 years. That's a meaningful improvement in both speed and capital efficiency. And finally, we extend that same philosophy into the clinic. But clinical development is where a lot of value is ultimately created and where also a lot of programs fail.

By bringing AI into trial design, picking the right patients, I can't enforce that enough and site selection, we're already seeing improvements in enrollment, speed and patient matching, helping us to run smarter and more efficient studies. But one more important point, this isn't 3 different capabilities. It's one continuous learning system. Every experiment improves our data, better data improves our models, better models make better molecules and then clinical data is fed back into the system to make the next generation of products even stronger.

Perhaps the best example of the flywheel in action is what we have demonstrated with Roche-Genentech, and we're announcing today where our biology engine discovered a previously unexplored and new neuroscience target. I'd like to spend a few minutes just to take you behind the scenes as to how we got there and why we believe this represents an important new approach to discovery medicines.

Together with Roche-Genentech, as we worked in this area to discover a new unexplored target from our AI-driven map of biology, we focused on a few specific elements. Why does that matter? First, this wasn't about finding another target within a well-studied biology. It was about uncovering previously unexplored biology and building enough evidence experimentally to advance it into drug discovery with one of the leading neuroscience organizations in the world.

Second, we believe this validates something bigger than a single target. It provides early evidence that when you combine the right data, build the right models, do very rigorous computational and experimental validation and pair that with the right complementary collaboration, you can actually systematically uncover novel biology. And we believe this is just the beginning. The underlying biological maps are reusable.

This is a really important point with the potential to generate many more therapeutic opportunities over time. Finally, across our collaboration with Roche-Genentech, we've now achieved more than $260 million in upfront and milestone payments with the opportunity for more than $300 million in additional development, commercialization and sales milestones for each future small molecule program.

All right. So let me show you how we built this engine. To understand why this milestone matters, the question is why neuroscience. It's worth stepping back and asking that question. Neuroscience remains one of the greatest unmet needs in medicine. More than 3 billion people worldwide are affected by neurological diseases. And yet CNS drugs, as we know, continue to have amongst the lowest approval rates in industry.

Neuroscience is particularly challenging because the biology is extraordinarily complex, difficult to model, and we have repeatedly returned to the same small set of well-understood targets with only incremental success. We believe meaningful progress will require discovering new biology, not just simply optimizing what is already known. And that's exactly what this collaboration was designed to do.

So the next question comes, what does it actually take to discover a target that people will have confidence in? And before I go into the details, just a huge, huge thank you to Roche-Genentech for this deep shoulder-to-shoulder collaboration. It's one of the few rare ones that I've seen where the teams are looking at the same data, the same models, going through what validation needs to be done. So that joint collaboration is critical here. So everything starts with disease-relevant biology.

We asked ourselves a simple question. What -- are we setting neurons in a context that actually reflects human disease? In our case, that meant creating iPSC-derived neuronal cells, both neuronal and microglial cells at an unprecedented scale, more than 1 trillion neurons and hundreds of billions of microglia. What this does is it creates a rich disease-relevant Atlas that can be reused again and again to discover multiple future targets.

We view this Atlas as one of the most important long-term competitive advantages. But generating proprietary data, while important, isn't enough. The next challenge is making sense of it. Before asking the models to find something new, we grounded every analysis in causal biology that we understand today. So really grounding it in genetics. We introduced hundreds of disease-causing perturbations and anchored our searches around well-established drivers of neurological disease.

That matters because it gives every subsequent prediction of biology from a causal target from the very beginning. Rather than switching blindly across the genome, we are searching from a foundation grounded in causal genetics and disease biology. Now as that's established, AI can help us on our foundation models ask a much more interesting question. What is not seen? What can be unexplored biology that we don't know of today.

This is where our foundation models come in. Instead of evaluating one hypothesis at a time, the models compare the biological signatures of more than 17,000 genes across tens of millions of data points. They build relationships across the entire genome and identify genes that consistently behave like known disease drivers even if they have never been implicated in that disease before.

That allows data and foundation models, not pre-conceived hypotheses to compile a prioritized list of new novel potential targets. Now AI can generate hypotheses, but medicines and programs require evidence. Together with Roche and Genentech, we predicted every target -- we looked at every predicted target and then put that through a rigorous experimental validation cascade.

We built confidence in layers. First, we established that the target actually sits in the right biological pathway. Second, we show that changing the target actually can improve cellular function, for instance, neurons or microglia. And finally, very critical. We demonstrate that this target and modulating it can meaningfully affect disease-relevant biology using multiple orthogonal assays.

These assays are very robust, but they also include other multi-omic data layers such as proteomics, transcriptomics, et cetera. While no single experiment tells the story, what we do here is build a body of causal evidence before advancing the target. So putting it all together, our collaboration combines 4 capabilities, generating disease-relevant biology at unprecedented scale and it's challenging to do to actually have a trillion iPSC-derived neuronal cells that are high quality, standardized, viable. It takes a lot of specialized protocols and know-how to do that.

Second, we use foundation models to systematically explore that biology. Third, we navigate from well-understood disease mechanisms towards previously unexplored new biology. And finally, a very important step is validating all of these predictions experimentally before we advance it.

So our first neuroscience target, as I mentioned before, have now advanced into a jointly developed small molecule discovery program supported by our design platform. And again, what excites us most is, of course, this target, but the fact that this kind of data is highly reusable, the potential to mine it over and over again for unexplored targets and also that this wasn't the result of one algorithm or one experiment.

It's the result of a new operating model for discovering medicines. Before I hand it over to Vicki, I would like to highlight as we move on to our internal programs, the pipeline. As you can see here, we have multiple programs in the clinic.

We're constantly looking at the data to make data-driven decisions for REC-4881 and FAP, where there's no approved therapies today and REC-1245 targeting RBM, a novel first-in-class target first-in-class degrader with limited clinical competition to-date. Combined with additional internal and partner assets, we believe this creates a diversified portfolio with multiple opportunities to create value in the coming years.

With that, I'm going to turn it to Vicki to walk you through the internal pipeline in more detail.

Thank you, Najat. I'll start off this morning by talking about our REC-4881 program in FAP. FAP is a rare disease that requires lifelong management. Patients with FAP develop hundreds to thousands of adenomatous polyps in their GI tract and require colectomy to reduce the risk of colorectal cancer.

Following colectomy, polyps may continue to develop and grow, both in the residual lower GI tract as well as in the duodenum in the upper GI tract. Patients require ongoing endoscopic surveillance, may require additional surgeries, and they continue to be at risk for GI cancers. With over 50,000 post-colectomy patients in the U.S. and EU5, there are no approved systemic therapies to alter the course of disease.

This represents an over $10 billion potential addressable market. REC-4881 is an oral MEK1/2 inhibitor with a differentiated dual mechanism of action in FAP with the potential to inhibit both new polyp formation via cross-talk inhibition of the beta-catenin pathway as well as to directly interrupt signaling of the MAP-kinase pathway, which is a key signaling pathway in advanced disease. So again, blocking potentially both new polyp formation as well as the existing polyps within the GI tract.

So with that, I'd like to take a minute to discuss the impact of this disease on patients through a story of a woman named Jenny, who lives with FAP. Like approximately 70% of FAP patients, Jenny inherited the genetic mutation responsible for FAP from a parent, in her case, her mother. Seeing what her mother experienced had profound psychological impacts on Jenny, who knew from the young age of 8 that she also carried this mutation.

She has since had to endure multiple surgeries, which have led to chronic and life-altering complications, including frequent bowel movements, malabsorption and dehydration, chronic abdominal pain and anxiety with medical PTSD from all of the surgeries and procedures.

We have heard from both patients like Jenny as well as their treating physicians an interest in a pharmaceutical intervention that can prevent polyp growth and disease progression and ultimately lead to a reduction in the need for repeat surgical procedures. REC-4881 has shown promising clinical data in the ongoing Phase II TUPELO study.

Patients who had undergone colectomy for FAP receiving REC-4881 showed a median polyp burden reduction of 43% after 3 months of treatment. That treatment effect was durable with sustained reductions after 3 months off treatment. Additionally, reductions in polyp burden were seen in both duodenal disease in the upper GI tract as well as the lower GI tract.

The upper GI tract in particular, is an area of high unmet need as approximately 90% of FAP patients will develop upper GI polyps. When removal of these upper GI polyps becomes necessary, the thin mucosal wall of the upper GI tract increases the likelihood of complications, including bleeding and perforation.

REC-4881 has a manageable safety profile with predominantly mild to moderate adverse events, consistent with the safety profile of other MEK inhibitors. We continue to enroll patients on the Phase II TUPELO trial, including patients 18 years of age and older as well as a dose optimization cohort. We are pleased to share that additional REC-4881 data will be presented during the Presidential Plenary session at the CGA-IGC Conference in November.

As Najat mentioned earlier, this conference is focused specifically on inherited GI cancer syndromes with a target audience, which includes physicians who treat FAP patients. We also look forward to providing an update on FDA discussions later this year. Now I'll move on to REC-7735. PI3-kinase is frequently mutated in several cancers and is a clinically validated therapeutic target.

Lack of selectivity for the mutated form over the wild type is a key challenge for existing agents as inhibition of wild-type PI3-kinase drives hyperglycemia. Increases in blood glucose are both a safety issue, which often limits dosing and an efficacy issue as the resulting hyperinsulinemia can reactivate signaling through the PI3-kinase pathway, undercutting the efficacy of less selective drugs.

REC-7735 is precision designed to be 100-fold selective -- greater than 100-fold selective for the H1047R mutation, which is the most frequent activating mutation in PI3-kinase. Recursion's AI-native platform identified a previously unpublished binding site and delivered a development candidate in 10 months with no identified off-target liabilities.

As hyperglycemia and the resultant hyperinsulinemia are driven by inhibition of wild-type PI3K, the selectivity of REC-7735 is expected to result in an improved safety profile with respect to hyperglycemia and may allow expansion into patients such as diabetic and prediabetic patients who are unable to tolerate current PI3-kinase targeting options.

An improved therapeutic index, as I have described, may allow us to expand treatable patient populations, both within existing PI3-kinase alpha inhibitor indications as well as in additional solid tumors in which PIK3CA mutations are prevalent including potentially triple-negative breast cancer, ovarian cancer and endometrial cancer, just to name a few.

Additionally, the improved therapeutic index may allow expansions into earlier stages of disease within oncology as well as non-oncology populations such as PI3-kinase-driven vascular anomalies. With the IND now cleared by FDA, we intend to initiate the Phase I ZINNIA trial later this year.

Dose escalation will begin in patients with PIK3CA-H1047R mutant solid tumors. Once tolerability is confirmed at an active dose, we intend to expand into the hyperglycemia vulnerable patient cohort to confirm the improved tolerability in this patient population.

Dose optimization of 2 active and tolerated doses will then be performed in ER-positive HER2-negative breast cancer patients. We may also expand into additional tumor types based on emerging data. We expect to share the first data from this dose escalation part of the trial in the first half of 2028.

And with that, I'll turn it back over to Najat.

Thanks, Vicki. And shifting gears a bit, we often get asked about whether advances in Frontier AI can reduce or increase Recursion's competitive advantage. We believe we have a truly competitive -- unique competitive edge. As reasoning models and agents continue to improve. Next slide, they become dramatically more powerful when paired with proprietary data, automated labs and real experimental feedback.

That's exactly the system we've been building for years. Now we are deploying agents across biology, chemistry and clinical development across the engine and also alongside our scientists. In biology, here are some very quick examples. Our target discovery connector is helping patients -- is helping scientists interrogate our proprietary biological maps in hours rather than weeks.

These are the large maps that we just talked about earlier in our partnership with Roche-Genentech, but also the internal maps that Recursion has built over years, accelerating the discovery of novel targets. In chemistry, our design agent reasons across structure, SAR and experimental data to prioritize the next design hypothesis, critical inflection points in programs.

This helps our scientists decide what to make next and compress design cycles from roughly 4 hours of structural analysis to about 30 minutes. And in clinical development, the Agentic workflows are already improving patient enrollment, contributing to about 1.3 to 1.6-fold improvement over historical benchmarks.

That's significant. These are still early examples, but I will have Chris Radoux, our Director of Structure-based Technology, who is in this day in and day out, walk you through a real example in practice. Chris?

[Presentation]

What you just saw wasn't a chatbot answering a question. It was an AI agent reasoning across our proprietary experimental data, our in silico data, our historical project knowledge and structural biology to surface insights that would otherwise require scientists long time, but then also nonobvious insights. That's because in drug discovery, the bottleneck is really just generating ideas.

It's finding the right idea quickly enough to keep the make, test, learn cycle moving. As these agents continue to improve alongside Frontier models, we believe they will become an incredibly powerful multiplier of what we have already built. And finally, I'd like to highlight another aspect of our AI strategy.

AI is advancing incredibly quickly, and no single model will remain state of art forever. Our strategy isn't to depend on any one model. It's to build an AI-native product engine that can rapidly develop and adopt the best advances, whether they're developed at Recursion or by the broader Open Source community. Nesso-1 is a great example.

We developed an open source this model. This is a binding affinity model that delivers Boltz-2 level accuracy with 10 to 20x faster inference, helping advance the field while enabling dramatically faster design cycles. But look, the real advantage is in the model itself. It's our operating system. It's our operating model. It's our ability to rapidly integrate these models into our proprietary data. That increases prediction performance, accelerates the make, test, learn loop and allows us to evaluate many more compounds at a lower cost.

And finally, great technology only creates value if you have the right people to translate it into medicine. We firmly believe that. And that's why we have strengthened our leadership team in 2 critical areas. First, Dr. Hoifung Poon joins us as Chief AI Officer. Hoifung is one of the world's leading AI researchers with more than 15 years at Microsoft Research, where he led pioneering work in biomedical foundation models and AI for health care.

Importantly, though, he's not just a researcher. He has repeatedly translated Frontier AI into real-world applications and deployed that at scale. At Recursion, he will unify our end-to-end AI strategy, bringing together Frontier research and Applied AI across biology, chemistry and the clinic.

Second, Dr. Donovan Chin joins us to head up drug design. Donovan has spent more than 2 decades solving some of the hardest problems in drug discovery from small molecules and RNA-targeted therapeutics to proximity-based medicines and peptide modalities. Across Parabilis, Arrakis, and Novartis, he repeatedly helps unlock targets that were previously considered difficult or even impossible to drug.

That breadth across modalities and that depth and experience of translating computational design into medicines is exactly the kind of capability we need to continue building at Recursion. Together, Hoifung and Donovan strengthened the 2 engines that will continue to define our future, world-class AI and world-class scientific design.

Now I'm going to turn it over to Ben to give us a financial update.

Thank you, Najat. As I've said in the past, we want to continuously increase the impact of every dollar we spend. We are demonstrating this today by lowering our 2026 full year cash operating expense guidance to $375 million. In total, our revised 2026 guidance represents a nearly 40% reduction from comparable 2024 pro forma expenses.

Through disciplined data-driven management, we have been able to continue lowering OpEx while still advancing our differentiated internal pipeline, achieving a series of partnership milestones and maintaining a leadership position in AI-powered drug discovery. We have been able to increase our return on investment through multiple levers across the company.

In our clinical pipeline, we use our Cleantech platform to drive more efficient enrollment and planning of our clinical trials, reducing the time and cost to reach important data. Najat and Chris described some of the systems that we use to make our internal discovery both more efficient and more effective. We also focus our technologies on predicting and answering the hard questions first so that we can prioritize those programs with clear potential clinical and commercial differentiation as early as possible.

Because we deliver outcomes that are truly novel and differentiated, like our Roche-Genentech milestone today, our partnerships have achieved over $500 million in cash inflows, including more than a dozen successful discovery milestones. All of our partnerships are designed to be breakeven or profitable on a direct cost basis from the start with substantial value growth as we achieve milestones.

In our product engine, we are able to build, test and integrate AI models on real projects using the scale of our internal pipeline and partnerships. We know not only if the model benchmarks well, but if it matters when it's applied to a drug program. This direct application allows us to determine early which technology investments are likely to have real-world impact.

We apply the same disciplined management style to our corporate operations. We have been able to maintain G&A at a relatively low percentage of total cost, which helps us maximize the scientific ROI of every dollar we spend. We ended the quarter with approximately $557 million in cash and equivalents, which we believe provides us with an operating runway through early 2028.

And with that, I'll turn it back over to Najat.

Thanks, Ben. I'll close by looking ahead. We have built an AI-native product engine. Now the focus is expanding its impact while continuing to translate its capabilities into the right programs and repeatable proof points.

So on our wholly-owned portfolio, you should expect to see continued progress across multiple programs, additional Phase II data for REC-4881and a regulatory update before year-end, continued advancement of REC-1245 with a more wholesome update later this year, the initiation of REC-7735, that Vicki just mentioned, and progress across the broader pipeline.

We are on track across those multiple fronts. With our partners, we expect to build on this year's momentum. Following the advancement of the first previously unexplored neuroscience target with Genentech, we see the potential for additional programs to emerge from our [ maps ].

And with Sanofi, we expect the potential to continue the progression of AI designed molecules towards development candidates and later-stage milestones. We're entering an exciting period with multiple opportunities to demonstrate the power of our engine.

With that, thank you again for the time today, and I'd be happy to take your questions.

Great. So I'm just going to go through some of the questions. The first question coming from Alec from BofA and Sean from Morgan Stanley. How does the collaboration with Roche-Genentech form a template for how you can leverage your platform with other partners? Maybe 2 to 3 aspects that you think are transferable and provide proof points.

Yes. I mean it's a great question. Thank you, both. Big picture, the way we develop our novel data sets for creating novel maps, and then we take those novel targets and design compounds all the way into the clinic, that sort of Lab-in-the-Loop is something we use for both our internal programs and for our partner programs.

So that template is something that will only get better and faster as we go on and we can -- in terms of new partners or current partners, we will continue to scale that. As I mentioned before, our differentiation really lies in 3 areas.

One is that data factory. I mean, especially in biology, given so much of it is not known well, having access to great biology and data is incredibly important. And that takes years to build.

I want to emphasize that, understanding how to generate that data, validate that data, develop the models and also have a supercomputer, which we have in a hidden location in Salt Lake City, having that entire stack to make sense of that data back into the lab and validated. I think that is something we are one of the very few companies that can do that, and we continue to drive momentum there.

Next question. Can you provide -- and this is a question from Sean from Morgan Stanley, Gil from Needham and Brendan from Cowen. Can you provide an update on FDA engagement on REC-4881 in FAP, the registrational pathway and the data coming up at CGA-IGC? Vicki, do you want to start it?

Sure. I'd be happy to. Maybe I'll start with the upcoming data at CGA-IGC. So we presented data from the Phase II TUPELO trial for the first time back in December of last year via a webinar. We do think it's really important to put these data in front of the physicians who treat patients with FAP.

And so this will be an updated data set, again, presented in an oral presentation at the Presidential Plenary session at that meeting, which occurs in November where you may see additional analyses that help contextualize the clinical relevance of the data as well as potentially additional patients in that analysis as well.

So we look forward to sharing those details with the FAP treating community later this year. With respect to the FDA engagement, as we've said, these are ongoing. I think it's important to remember, there's very limited regulatory precedent in FAP. And so our engagement here really is around making sure that we derisk the study design from a regulatory standpoint, including things like what is the appropriate primary endpoint to demonstrate clinical benefit.

And I would say, as somebody who worked at FDA many years ago, those discussions, those conversations have been productive and I think are helping us get to a better point in terms of the study design. So nothing out of the ordinary there.

Again, this is just a rare disease with limited precedent, and we continue to have a productive dialogue with FDA and look forward to sharing -- once we have sort of something more concrete to share, look forward to sharing more details on that later this year.

Thank you, Vicki. All right. I'll move on to the next question. Ben, this is for you from Priyanka, JPM and Gil from Needham. Can you provide more color on what operating efficiencies were done to reduce the OpEx guidance? Is there potential for further belt tightening on OpEx in second half of 2026?

Yes. Great question. And I think as you saw in the presentation Najat covered, we haven't changed any of our full year guidance on what outcomes we're trying to achieve over the course of the year. And I think that's really important to remember because this reduction in guidance is actually from doing the same amount or more with less.

And so what we've really tried to focus on is how can we get to the most important answer first. You heard some of the description of the technologies that Chris took us through that Najat took us through. And that really makes a difference on how we can operate and how we can deliver those outcomes. So I think we started the year and we had some ideas of where we could go.

What we've seen is they actually have impact. We are actually getting to the answers faster and more cheaply. I think the numbers that everyone should use are the numbers that we give in guidance, which is the $375 million. That is our expectation of where we will be operating. But at our core, we are always looking for a better and faster way to do everything that we do. We are a technology company, and we should be getting more and more efficient over time. So we will keep looking and update you as we know more.

Thanks, Ben. Yes. And just to maybe reiterate that, we have -- we always have a commitment in order to ensure that every dollar goes further with some of the improvements we're seeing in our engine. You saw some of the examples around the fact that we design 90%.

We make -- physically make 90% less compounds for the one that goes into the clinic. We took about 1.5 years versus 4 years versus industry. Those are meaningful improvements in the velocity that we see in our engine, and we ensure that, that actually parlays into our spend.

We mentioned earlier this year that we changed our budget to an outcomes-based budget so that every single aspect like Alec and Sean going back to your question, when we do a partnership, we know exactly the fully loaded cost of building a map of a program and so forth. And that really helps us to ensure that those efficiencies are realized.

The other thing I'll also say, we continue to focus on our G&A and ensure that every single dollar is actually going to our programs and our partnerships. So we will continue to put pressure. That's our commitment, just like our commitment is to deliver on proof points from what can be really a value inflection point for the broader community in terms of programs and the use of AI to create value.

Okay. With that, I'll go to the next question, a platform question from Alec from BofA and many others, okay. With multiple tech companies entering drug development and as Generative AI becomes increasingly available, how does Recursion differentiate itself today and in the future? And what do you believe remains Recursion's durable competitive advantage competitors will find hardest to replicate over the next 5 years?

Great question, Alec, and everyone else who asked that. I think that's why you saw the second slide in the presentation was really around our durable moat and our differentiation, and that evolves over time. I think number one is the data factory. Look, you just said Generative AI is becoming increasingly available, maybe some would say even commoditized.

Where does the differentiation come from? If 80%, 90% of biology, as I've known, it has to come from high-quality data generation. Models depend on good quality data to be trained on. And you saw with the example with Roche-Genentech that we shared today, but also across the board, starting with disease-relevant data sets also matters. That just doesn't exist. So in order to build that 1 trillion iPSC-derived neuronal cells, that's a cell manufacturing capacity that we have in our Salt Lake City Labs.

Over years, we have gone through the pain and suffering of what works and what doesn't work. So think about it as a really mature and increasingly validated capability. So that's one on the data factory. And that's not just for biology.

You heard from Chris Radoux, 10 years of actually doing small molecule design, millions to billions of virtual cells -- virtual molecules that have been generated also gives us a lot of rich data, not just in areas that are known to the world like kinases, but actually other targets that are less known and not as available in the protein database, PDB, for instance, and others.

So that's one big pillar. Second, I can't emphasize enough is that Lab-in-the-Loop, that operating model because it's one thing to have great data. It's another thing to have great models. But really important, we need to validate these predictions. The only way we get this to be useful, utility at the end of the day to make a drug is if you're validating it back into the lab and that feedback, good or bad goes back into the models to make them better and smarter.

We do the same thing with AI agents. The more you engage with them, the more you give them feedback, they get better. I think that integrated Lab-in-the-Loop is [indiscernible]. It's hard to build for 2 reasons. It takes a lot of technical expertise, yes. It takes a lot of years of knowing what works, what doesn't works, yes. It takes tons of reps and with partners that are some of the best in the industry, we learn faster.

But so much of it is also culture. It's culture. I've always mentioned the piece that we have bilingual scientists that understand -- better understand, I would say, both science and tech that have appreciation of the challenges and opportunities of both, that open-mindedness where an agent gives you a different hypothesis from what you started, when you're in medicinal chemistry that's worked in that space for decades, that takes a different mindset, and I cannot emphasize that enough.

And then the third piece is what are we actually making from the agent? FAP, first-in-class oral for a disease where nothing has been approved. It's a stand-alone high-value asset. RBM39, first-in-class target, first-in-class degrader built from this platform with limited competition. So what you'll see in our pipeline is an incremental improvement but any 1 or 2 drugs that can actually be a stand-alone differentiated asset in its own right.

And we all know that takes time. So I think those are the 3 big areas that are not just an advantage for today, but continues because with every week, we're doing 2 more -- 2 million more experiments in our labs, the data moat grows. With every week, we actually have people turning through that lab and learning, that grows. And as you can see, with every week, month, we're making progress in our pipeline. And that takes time, resilience, focus and discipline, and that's what we're doing.

Okay. One more question for Vicki. PI3K questions from Brendan at Cowen and Dennis at Jefferies. Looks like REC-7735 passed your internal criteria for go/no-go decision with the Phase I to start for second half of 2026. Can you tell us a bit more about the go/no-go process?

What it is about the preclinical profile that gives you confidence that this is the right candidate? And also, what the Recursion AI platform has told you about the best development path forward in terms of study design, patient selection, et cetera? And then there's another sub-question, but I'll start with that.

Sure. So first, maybe to start off by saying we believe that there's room for improvement in the PI3-kinase space. So again, this is a very common mutation in certain malignancies, including hormone receptor positive breast cancer, but also extending beyond breast cancer into other GYN malignancies as well as head and neck cancer and colon cancer and others.

So important target still remaining unmet need in terms of maximizing the therapeutic index and ultimately, the efficacy that patients see. So the go/no-go process really involved a rigorous evaluation and confidence building in our preclinical data set. So the selectivity that allows us to hit the target hard without seeing additional toxicity.

So again, both in terms of the efficacy that we're seeing in preclinical models that look similar to -- at least similar, if not improved upon competitor profiles, the safety profile, including the lack of hyperglycemia, but also, of course, our GLP tox studies. These all helped us build confidence that this was the right molecule to move forward with into clinical trials.

Of course, ultimately, after evaluating these data, we made the decision to go forward. We've submitted the IND and that IND is now cleared, and we look forward again to initiating that study this year. In terms of the AI platform, I think one of the key pieces from a clinical perspective is these patients are going to be selected based on the H1047R mutation, so a biomarker, which will require a diagnostic.

And one of the key areas where I think the platform is helping us is in terms of our ability to find these patients, look for the right geographies and sites in which to conduct our clinical trial and help us accelerate the development -- sorry, the enrollment of this patient population.

Thank you, Vicki. Maybe just a couple of things to add. We talked about this early on, which is for this compound specifically, it is over 100x in activity of wild-type. So it's wild-type sparing. Why is that important? Important from a perspective of can we actually have the patients stay on increased dose intensity, dose duration, as Vicki mentioned, to really try to improve the outcomes from patients in the [ TI ].

But also even with grade 1/2 increase in -- for instance hyperglycemia, et cetera, we have seen elements that it can lead to reactivating the exact pathway, PI3K pathway that you're trying to suppress. And that has the potential for also compromising some of the efficacy that can be seen.

So there are multiple elements to why as we look at this compound, what we want to test in the clinic is, is it actually giving us a better safety profile and in turn, can it give us a better efficacy profile that would improve the therapeutic index. The other thing I would just say from the AI platform as well as Vicki mentioned, one is recruitment.

We know this is a competitive area. We're starting in with solid tumors, as Vicki mentioned, but it gives us optionality given based on what we will see in the profile to either go in onc or non-onc indications as well. And that's also another area where the platform can help. We're really thinking about what are the right patient groups and indications that we might select that others haven't maybe explored to-date.

So a lot more work to come, but step one is to go into the clinic and ensure that we are seeing the elements, the compound was really designed for. And recall, the compound was designed in 10 months, 242 compounds synthesized, 13 cycles and the pocket was a previously unpublished pocket. So we're not going after the same areas, which is why you see almost 130x selectivity over wild-type, super precise, super precision based.

H1047R is the most -- one of the most frequent mutations you see in this space, one of the ones that's tied to disease causality and progression the most. So we're excited. But again, it's part of multiple different programs that we're looking at. And based on data, we'll make the right go/no-go decisions as well.

One maybe just sub-question. When should we expect initial monotherapy data? I think Vicki had mentioned first half of 2028. So stay tuned. And with that, I'm not seeing any more questions on the screen. Thank you again so much for joining us today. Looking forward to the progress over the next set of weeks and months. And as always, we'll talk to you soon.

Ben Taylor, CFO of Recursion, here with us. Thanks so much for joining us.

Maybe to start, if you want to may provide an overview of your platform, how exactly it integrates AI in your process? And what are the key points of differentiation that you have versus other AI-enabled biotech companies?

Sure. Great. And first of all, thanks for having us here, Tommie. This is always a terrific conference and great to be here. So taking a step back, if you think about what Recursion does as a whole is we combine different ways of predicting models, whether it's biology, chemistry or ClinTech. And then we'll look at a program and say, why might that fail? Could it fail because we don't understand the biology? Could it fail because we don't understand the chemistry or maybe it's the patient population or how to design a clinical trial.

And then we try and take those 3 different pillars of the platform and say, can we create a better data-driven model that's going to be able to get us either to solve it in drug discovery or in clinical development and predict that failure point. And so what we've built now is really going beyond our own data generation, going beyond our modeling systems and really focusing on the proof points of that because we're not in a industry that creates models, we're in an industry that creates drugs.

So right now, what we have are 5 programs in the clinic, all of which have data over the next 12 months. A couple more coming behind that. We've got a partnership business that is primarily focused on Roche and Sanofi, but we've brought in over $500 million from those partnerships, and those are really high-value, high NPV programs. And underlying all of that is an AI-based platform. And so this is where we're AI native in the sense of we don't have infrastructure that we're replacing to put AI in place. We actually look at every single problem, and we say, is there a better way that we can do it using data, using AI, using computational power to be able to solve it differently than it's been done before.

And so the goal is to improve the probability of success. And because we're using technology, we do it generally faster and cheaper. So we've published a number of the statistics, but we've been able to really bring down development and discovery time lines and the cost of being able to do that as well.

Yes. And at this point, it's been over a year since your merger with Exscientia. Maybe at this time point, how you're thinking of the key priority for the company, both on the pipeline, BD and platform side?

Yes, yes. So Exscientia came in with the chemistry background to Recursion's core capabilities in biology. And what we've done since then is be able to combine them so we can identify novel targets and then we can design the drugs that are optimized for those targets. We green shielded a ClinTech business about 2 years ago. So Najat Khan, our CEO, when she came over from J&J, she was heading up the AI business as well as the portfolio review of all of J&J's programs.

And she had created inside of J&J a ClinTech business that used a lot of real-world data. And so we brought that in and built it up from scratch. And it's actually had a massive impact across all of our programs. We've been able to -- even though it's a relatively new part of the platform, we've been able to speed up our clinical trials by 30% to 60%. That's across multiple different pieces as well as identify lots of new sites in different patient populations.

So the platform has actually changed a great deal. Our capabilities have changed a great deal. I think underlying it, there's also a cultural piece. So maybe it's because of the chemistry and biology focus, but Exscientia had always been deeply analytical, really focused on getting to that exact solution. And we ran a business model that was very associated with that as well.

Recursion was always more big picture, driving towards the vision and trying to be really transformational across the industry. And so those 2 have actually come together really nicely. So we've done a lot of work on bringing in the cost structure. It's funny because we took 35% out of the budgets since the merger. And that wasn't actually what I would call cost cutting. What we did is we implemented a program that measured impact of literally every dollar that we spend. It doesn't matter if it's pipeline or tech or G&A and said, what's having the high impact and anything that didn't meet the bar was eliminated.

And so it's really more of impact modeling. And we continue to do that, and we continue to bring down efficiencies in that way while still trying to drive towards a pretty broad clinical pipeline and partnership business.

And we think you'd be quite active on the BD front. You brought in new technologies for the platform. You have pharma partnerships, you had an NVIDIA partnership. So maybe what are your thoughts on the overall strategy here? And where are the future directions for activity as we think external?

Yes. Well, and this is actually a really important point that I think a lot of people miss. Our original mandate was not just to change the probability of success. Basically, the investors that funded a lot of the original work said, I want to invest in biotech, but I don't want to invest in binary risk. And so I don't invest in biotech. So how can you create an actual ongoing business model. And so through good times and bad over the last many years, we've held to that and said, we need to be able to bring that risk diversification in. And that's why you see in our pipeline, we have best-in-class, we have first-in-class. We've got a real nice mix.

But in our partnerships in our BD as well, where we look at it and say, how can we get a balance of different partners that all contribute. We run that partnership business actually at breakeven or to a profit at all the times. So they pay us in advance for our direct costs as we get to in-licensing or profit that flows down to us. So that's been a really nice backbone. The other part of BD, we haven't done yet, but we are set up to do is basically out-licensing of our own programs. And so if you look at our pipeline, the 5 clinical programs, the 2 in preclinical or in IND-enabling, it is not a part of our business plan to take all of those programs ahead ourselves.

And so we want to look at the data. We want to see how they advance. I'm sure we will always have 1 or 2 at least that are our own wholly owned programs. But we're also very open to out-licensing and other aspects like that. We just want to get to a really valuable transition point. And so the way that we look at it is it's all option value. We can do portfolio management. We can do capital management because we have this diversified business.

Yes, that makes sense. And so maybe let's turn to the pipeline. If you'd like to give an overview of the data that we'll see within the next 12 to 18 months before we get into some more detail.

Absolutely. So we've got a number of data points, both towards the end of this year as well as the beginning of next. So this year, our most advanced program is 4881, and that's in an orphan disease called FAP. So we're currently in FDA discussions on the registrational trial design for that, and we'll give an update on the registrational trial design. We are also continuing to enroll a broader patient population than the original data cut had been on.

And so we'll put out that data probably first half of next year. But those are 2 major points, especially on the FDA registrational trial design. Everybody is waiting for that on FAP. And so we'll update that as -- the pipeline a program called RBM39, which is a really interesting example from our biology platform where we looked at the biological mimic of CDK12, which is a known good oncology target, but almost impossible to drug cleanly.

And we saw that RBM39 had a very similar biological fingerprint for it. And so we created a degrader molecule to be able to go after that. We've done the initial dose escalation, which we just talked about on the last earnings call. We saw a nice clean safety profile, dose-proportional PK/PD. We didn't put out a lot of data. I kind of think of it almost like a -- we did a Phase I futility analysis because novel target and a degrader system, just take a look early and make sure everything is going the way that you wanted to. It was. That has a lot of potential, and we'll probably talk about it later because it really targets genomically instable cancer mutations.

And so that's a pretty broad number of indications. More data on that later on this year as we continue up into that escalation. And then looking across the rest of our pipeline, early next year, we've got MALT1 and CDK7. Both of those will have data readouts. MALT1 in monotherapy, CDK7 in combination therapy. And then second half of next year, we have an LSD1 program that we just started up as well.

Okay. Great. So on your lead program in FAP, you're discussing population and trial design with regulators. What are the potential scenarios here for what you could announce in the second half? And as you think of ERAP's path in their trial, what could be the potential similarities and differences?

Yes. So ERAP set a template of what we could do in this area. So this is something there's no approved therapies. And there really wasn't a lot known about the disease as well. And we've published some of the natural history data that we've been able to do through our ClinTech platform that's really educated not only ourselves, but also the industry on this is what the management cycle, the health cycle for a FAP patient looks like based on all of the data.

And so what we want to do is take that to the regulators, take that to ourselves, be able to do analysis and say, this is what's important. This is what patient populations look like in different settings. These could be the higher risk aspects. Talk to them also about different potential endpoints. Now it would be fine if we went with the same endpoints that ERAP had used. That -- it's not a bad trial design. It's just what we want to do is look at other areas where there wasn't as much known coming into it.

So excisions is a really important part. This is -- it's not as invasive as obviously a colectomy. But when you've got literally hundreds of polyps throughout your gastrointestinal system and they're doing excisions on it every 3 to 12 months to remove those polyps and then they keep growing back, there's major bleeding episodes, there's infections. There's all sorts of problems in being able to digest and absorb. And so the comorbidities that go along with just that, which is only out of a colonoscopy or other endoscopic treatment are actually very substantial. And that wasn't included, for example, in the ERAPA trial. And so we want to ask questions like that. The FDA has been -- we haven't noticed any differences in any of our programs from the changes that have been going on at the FDA, but we're also not going to get in front of it, and we want to have the discussions.

Yes, makes sense. And you spoke a bit on the ClinTech platform that you have. Maybe if you could just lay out how it was used in FAP and how it's different than if you had used a more traditional method.

Yes. So there were 3 ways that we used it. 2 of them on the natural history side. So one, we took -- there's a university in Amsterdam that had been doing for 20 years, they've been tracking FAP patients. So it's the longest running database of FAP patients. And so we were able to take that and convert that into something that looked at, okay, what's your average polyp burden growth? What are the patient demographics for this look like? How does progression look and be able to compute that into something that showed on average, you've got about a 60% increase in polyp burden per year per patient, which is just a massive amount of increase.

No one had ever actually shown that polyps continue to grow, right? Like before that, you couldn't say they will spontaneously reduce or grow, right? And so we've pretty definitively shown which way that goes. In addition, -- we have a database of over 300 patient lives that include not only claims data, but also all of the physician notes and those other aspects.

And so what we did is we looked at all of the physician notes and found all of those that were related to FAP patients, it was about 250,000. And we -- in the course of the week, because we have our own coders, we have our own supercomputer, we were able to create an LLM that allows us to query in the same way that you would with Claude or Gemini, how many surgeries are these patients having here? What are their likely comorbidities? What are they getting treated with and really understand the entire patient journey in a much more detailed way.

And there is no standard of care manual for these patients. And so we were helping to define that based on what is actual clinical practice. So those 2 were completely novel and just play off our strength in data analytics and AI. What we also can do is basically run the simulated trials. So this is something that there is a classical history of being able to do clinical trial simulations. People do too little of it. They should do a lot more.

But what we do with all of our trials is we'll take the patient population. We'll change the inclusion/exclusion criteria. We'll look at different combinations or comorbidities and see how it affects the outcomes of the trial by basically being able to run different statistical analyses around it. And so we can better target what does a good trial look like, what does our statistical package look like? How can we do a more informed data-driven trial design.

Interesting. And then if we think about the FAP market, how big could this opportunity be as you think of prevalence, what your label could potentially include and the proportion of patients who are seeking systemic therapy?

Yes. So U.S. and EU5, there's about 50,000 to 70,000 depending on source, patients that are identified with it. That is mostly the hereditary population. So most FAP patients, at least most known FAP patients have germline mutations that are leading to this. And that's why they are diagnosed usually in adolescents or their teenagers. And so what we look at is, one, how can we treat those patients. There's no commercial therapy out there right now. And so their only option is surgical ones. So they either get resections. They almost always have a colectomy by the time they're in their 20s.

And then on average, have 10 different resections over the course of their lives as they go through. That doesn't include all of the colonoscopies where they're having the polyps removed on a 3- to 12-month basis. And again, these patients literally could have hundreds or thousands of polyps throughout their colon. So that's very severe. What we look at from the market is almost all of those patients could benefit from this because the polyps never stop. They're all precancersous. So literally, every single one of them is a precancerous. There's no benign polyps in FAP.

And so they have to be removed at some point. And so being able to control and manage the polyp growth and burden, there's something called Spigelman staging, which basically dictates to doctors when they need to take it out, and it's a combination of polyp burden and dysplasia. We're happy to see 40% of patients even in just the 3 months that we treated them for, 40% had a downstaging of their Spigelman scores, which is really exciting.

So what we want to do is get in there, look at this as really a chronic care market. Most, if not all, of those patients could benefit from a ongoing chronic therapy. There's probably also a pretty heavy underdiagnosis of the somatic mutations as well. So you can get this somatically. And currently, about 20% to 30% of identified patients have somatic mutations that have led to it. It's just the APC gene going off.

The question mark becomes how many more patients of those are just being caught when they get colorectal cancer? Because if it goes undiagnosed, they will get colorectal cancer. And so there should be some portion of that, that could also be treated earlier.

So maybe this is where we segue to the oncology portfolio. As you said you reported data from RBM39. What is giving you confidence in the therapeutic index that you're showing? And what -- when we think of what you're going to report next, what will be the key learnings from that next data set? And over time, how are you thinking of potentially making this a combination opportunity?

Yes. So RBM39, really interesting program. So CDK12 is a very common transcriptionally driven issue with oncology and could be a great target. RBM39 as a transcriptional component is able to also have a similar effect. And so what we've looked at and we've shown preclinically across dozens of different models is that whether it's DDR, some of the cell cycle issues, basically any of the different types of mutations that have that genomic instability could be affected by this.

You get that synthetically lethal mechanism of action or you at least weaken it down. If you look at how you could potentially use that, what you would want to do is be able to have that synthetically lethal component ongoing for a long period of time. What we got excited about in the data is the preclinical model said we'd need about a 70% or 80% IC70 or IC80 on a pretty sustained basis. So this is a protein that actually resynthesizes quite quickly. And within 8 to 12 hours, you can actually regain back to normal function. So what we needed to see was an extended period of being able to have that -- and so what we -- we haven't released all of the data because we wanted to gather more before we put it out, but the profile absolutely looks like QD dosing would be appropriate.

We had no DLTs in the run-up so far. That's encouraging from 2 perspectives. One, obviously, you don't want to see the DLTs, but also this is a degrader. And degraders have historically had a lot of problems with off-target toxicity. And so not seeing that is very encouraging. And then we also saw a dose-dependent PK/PD profile. So we're seeing the systemic distribution, and we're seeing initial biomarkers on the PD that look like we're all heading in the right direction. So in the next couple of doses, we'll be getting into the predicted therapeutic window for -- from the preclinical models. And we'll report more on that second half of the year.

Okay. Great. And do you think this is a mono or a combo opportunity overall?

I mean anything in cancer is a combo opportunity, just how it's used clinically. From a mechanistic perspective, if you had a genomically instable cancer, absolutely, you should see some activity. So what we tried to do is it wasn't a biomarker targeted Phase I/II. It was a biomarker enriched. So what we went is we went after indications that have a higher prevalence of genomically instable cancers. And so we do want to get some people who have genomically stable cancers because you always need the negative in data as well, but we'll also be looking at a number of different biomarkers that could indicate where this is the best patient population.

Okay. And you have CDK7 data in first half of '27. What's the success for this data set look like to you?

Yes. CDK7, it's one of those big -- it's a high-risk, high-reward target. So it's a master regulator around a lot of cell cycle and transcription. So you're really talking about a fundamental biological mechanism -- we've seen the CDK 4/6s, which basically target one stage of the cell cycle. That's a -- I know it's north of $10 billion drug class now, but it doesn't have the durability that you'd like to see. And part of that is because you're targeting that single point. And so the cancers are very good at getting around a single point aspect.

And so CDK7 could be a better mechanism because you're targeting basically 2 stages of the cell cycle as well as some of the transcriptional elements. And the real question there is, can you manage tox? So what we did is we designed something that had a very short half-life that was reversible that basically could get in quickly hit to a high level, anything that was a rapidly dividing cell and then get out of the body. So trying to avoid the heme tox is a big area that you'd see or the GI tox. Also, previous efforts had made things that were very much transporter substrate.

And so that's going to be very counterproductive because you're going to have a lot of GI tox and you're also going to have trouble staying in the tumor microenvironment. So these are examples of how we could use the design platform to say, I only want to be in the system for 6 to 8 hours, and I want to go into a novel chemical space that's not going to have that substrate issue. A number of different profile properties that we were able to design around. In the Phase I monotherapy, we saw that the design -- we saw all of those things that we had designed for happened. Now the question is really, okay, in a combination environment because this is more cytostatic than cytotoxic, in a combination environment, do I see efficacy, too? And can I do that in a manageable toxicity window. If that data is good, massive potential for it, but we'll have to wait and see.

Okay. So for PI3K, a very active market. You have Astra, Lilly, Relay Therapeutic, others. What is the niche that you hope to carve out with your program?

Yes. Well, and this is a funny one because it is a niche, but it's a very big niche. So about half of the patients that are suitable for PI3K therapy are prediabetic or diabetic. And so that means that anything that's causing hyperglycemia is going to be counterindicated against it.

So far, all of the PI3Ks that are more advanced or on the market are causing some level of hyperglycemia. And so that was the core thesis going into it. Can we create something that is ultra-potent and selective because the hyperglycemia is coming from a selectivity issue. So if you look at like the Relay or Scorpion compounds, they were about a 10x selective over wild type for the PI3K. And so you get a reduced hyperglycemia compared to some of the -- like a Piqray, for example.

But what we wanted to do is say that hyperglycemia is a sign that you're not just focusing on the mutant. And so we took the most common mutant, which is 1047, and we created something that's about 120, 150x selective over wild type. And it's been incredibly clean. We've had great preclinical efficacy with no hyperglycemia signals. So in one hand, it's a niche to go after that. On the other hand, it's actually a really large patient population. We put out some of the numbers. It's about 11,000 patients in breast cancer and about 20,000 patients outside of breast cancer.

And you have a MALT1 program. There has been some historical challenges with tox here. How are you thinking about addressing that?

Which program...

Your MALT1.

MALT1, yes, yes. So MALT1, another -- it's another one of the design stories. If you look at the compounds that were with J&J and Schrodinger, what we saw is basically, there's a UGT1A1 inhibition, which is not actually a part of the MALT1 mechanism, but causes hyperbilirubinemia. The issue with that is you would almost always use a MALT1 in combination with a BTK or a BCL-2. Both of those have liver tox. And so if you combine liver tox with hyperbilirubinemia, then you get Hy's Law.

And so you can't use those together, right? Or you're going to be triggering a lot of really serious toxicity. And so that was the design. Can we take that out? Because our thesis was that was not an on-target tox, that was an off-target tox. And so what we showed in our preclinical is we didn't have any UGT1A1 inhibition. We got into a very novel chemical space. We had great potency and selectivity. And so now that's in dose escalation for B-cell malignancies and could be some really interesting data. That's another one where you would also see it very early, right? Like you don't have to do a big trial to see are you hitting UGT1A1 or not? You'll get it in a handful of dose levels.

Okay. Is there anything else from the pipeline that we missed that you would highlight?

No. I mean we've always got different things coming along. I think getting back to that business model point, one of the real cores here is we wanted to -- and this was also post the merger. We wanted to take the best of the best from both companies. And so we actually killed a number of programs, both clinically and preclinically because we looked at it and everything had to pass a multiparameter analysis.

We had to say, okay, good science. Hopefully, it has that. But is there a good clinical pathway? Does the data support it? Is there a good commercial opportunity? And then can we do it ourselves or not. And so everything went through that, and we as we look at those programs, current clinical or IND-enabling programs has the potential to be a blockbuster, has the potential to really make a broader impact. The ones that are more design stories, we've also already seen some biological validation like we're just talking about with MALT1, for example, or LSD1 is similar, where there has been clinical evidence of those mechanisms modifying disease, but they come with a lot of tox.

And so can we take the tox out or trying to diversify that risk and say, let's go for some really novel biology like around FAP or RBM39. And so we're trying to just balance that all together, take the best of the best. And we definitely don't need them all to work, but hopefully, a lot of them will.

And then as you think of your cash runway, maybe you can walk us through some of the assumptions baked into that and how there are also considerations between how you recognize revenue from your partnerships versus look the OpEx from those?

Sure. Yes, always a point of confusion. So cash runway out to early 2028. And we've actually been able to extend that out a couple of times just by better cost management and efficiencies. And we're going to keep focusing on that and keep trying to drive it down, really focused on we're a technology company. We should be doing more with the dollar tomorrow than we are today.

So early 2028, the only assumptions besides our operating costs that we have in there. So our operating costs include advancement of all 5 of those clinical programs plus one of our IND-enabling programs and we assume that those go out. So obviously, that will be directly correlated. On the partnership side, all that we've included is probability weighting of inflows from our existing partnerships. So there's no new business development. And we haven't gotten into how we do probability weighting, but what I'd say is we try -- even though we always hope to exceed industry benchmarks, we use industry benchmarks for our own internal modeling to try and take a more conservative view.

And that's how we've modeled out the potential milestone payments from the partnership. And that's all. There's nothing else in there, trying to keep it clean.

And then just maybe one last one on the platform. As we look at the AI-enabled drug development space, we see a lot more activity on the preclinical discovery side versus the clinical side, and you've been doing a lot of work on the clinical side. Maybe if you could just speak to how this shift has happened and what the challenges are and kind of your unique position here.

Yes. Well, and that's -- it's actually really amazing how the story has transformed because we were absolutely -- both companies started up as basically a point solution. But what we found is you don't get to a 95% failure rate from a single point. You just have failures all across the spectrum. And so what we do now is we actually do a lot of applied development. How we ended up with our 3 pillars, the biology, the chemistry, the ClinTech is we looked at how programs are going to fail -- and what do we need to do to try and prevent that.

And so that just naturally led us to this broader platform. Now the interesting part is each one of those individual technologies that comes together into those pillars has to be independently validated. The only way to do that is actually working on programs. And so one of our big advantages is for 13, 14 years now, we've been doing applied development. So most people don't realize this, but about 2/3 -- a little over 2/3 of our budget is applied. So it's directly going into pipeline programs or our partnerships.

And so that's how we like to develop technology. We -- only about 10% to 15% of our budget is actually on what a lot of people would consider sort of the research, the basic research part. And the rest of it is really, I want to understand that this model works because I'm using it in a program and it's making a difference. And so you do that year after year with a lot of programs with big partners who say, yes, this makes sense or it doesn't. And you end up getting this very large base of validated platform capabilities.

Now we can look back and say, what's the problem? Do I have a tool in my chest that works on that? And what's more so is we have loved all of the development going on in agents. So we've got partnerships with a number of the large tech companies looking at we can deploy our own agents or agents that we bring in from the outside very quickly because all of the fundamentals that an agent would want to use have already been built and validated inside of our system. Like an agent will never fix your broken system. You have to have a working system that an agent can then run on. And so that has been a force multiplier for us of bringing even more efficiency in than we had before.

Yes. Well, thank you. I'm looking forward to all the progress.

Thanks. I appreciate you having us again.

Thanks for attending the 2026 Bank of America Healthcare Conference. My name is Alec Stranahan. I cover Recursion and SMid Biotech at Bank of America. And I'm pleased to be joined today by Ben Taylor, Chief Financial Officer of Recursion. Thanks for being here, Ben.

Happy to be here. Thanks for the invite. It's always good to come out every year.

Yes. Yes. Great. Well, maybe just to start, there's a lot going on at Recursion. This is AI-enabled drug discovery for those who are uninitiated, and Recursion is really a trailblazer here. They've been doing it for many, many years. I guess, Ben, to start, you've got sort of 3 value drivers at the company. You've got your internal pipeline. You've got your pharma partnerships, and you've also got the Recursion OS platform itself. I guess, where do you see the biggest re-rating opportunity for the company. Is it from the pipeline validation? Is it from advancing and expanding your partnerships? Or is it really from the platform and sort of what you're driving from that?

Well, at the heart of it, we're a therapeutics company. And so we really focus on making sure that we're advancing the pipeline and getting the value drivers from that. I mean if you look at one of the fundamental tenets of the business is to create a more risk-diversified biotech model. And so that has, over the last decade-plus involved a lot of platform and a lot of the partnerships as well. But in the end, it's all for the goal of actually developing therapeutics that are going to reach patients. And so right now, what we have is 5 different clinical programs that all have clinical data in the next 12 to 18 months that's really impactful for showing proof of concept and/or we've got 4881 that's starting its registrational trial. And so those are really the core value drivers, not only from an investor perspective, but also just for the core mission of advancing medicines towards the patients.

The other components, it's funny because for us, it's a continuous stream. We don't really view the platform separately from the pipeline and the partnerships. We're actually just doing the same work that we would do on our internal programs, but we're getting paid in advance for it and a nice NPV on it. So all of it really flows back to does this drug have differentiation? And is it likely to be a good product for patients.

Great. And maybe for those less familiar with the company, maybe you can just talk about sort of the full stack capabilities that you have and how that's evolved over time, where you're sort of at today and how you sort of see yourself settling into the rapidly evolving AI landscape?

Yes. If you take a step back and you think about, all right, how can you make a differentiated model in biotech, there's 2 components that you have to do. One is improve the probability of success. If you've got a 90-plus percent failure rate in the industry, you're always guessing. And so that's the part that really focuses on building a better data-driven model to it. If you think about how we continue to advance along and bring up the other pipeline programs, what we've been able to do is show that we can do that across a number of different facets on where that probability of success can come down.

So if you think about a program as a whole, you need to think about the failure of a clinical trial being based on statistics. It could be chemistry, it could be biology, it could be patient selection, it could be trial design. And so all of our systems are actually saying, can I create a better predictive model to be able to solve that recent [Audio Gap].

I want to be able to explore the biology better. I want to go to a completely novel chemical space. And so there's not a single point of technology that really defines us. We just have the largest integrated toolset to be able to solve those problems that cause failure for clinical trials. So I think that is a massive differentiator for us, the fact that it's been created within a unified system so that it works together, where most of the companies in the space right now are really developing point solutions for a single area.

The other aspect is the data. I mean, because we've been doing applied work for so long, we've created a massive dataset of over 50 petabytes of internal proprietary data that allows us to build better models. So if you're just using the public data or even if you're just using data that has been generated outside of the ML context, the fidelity of it is pretty low. And so it doesn't actually drive a lot of predictability. What you really need to do is create something that supports a better data analysis. We just had a paper published that showed a much smaller data set of highly annotated data actually creates far more predictive models than a massive data set of poorly annotated data.

Interesting. And I want to talk about something that you got asked at our breakfast this morning, and I thought your answer was pretty interesting. And just thinking in the near term, the industry is evolving in terms of where innovation is being rewarded, right? Typically, that was through pharma partnerships like what you have. A lot of those pharmaceutical companies are going to China now as well. I guess on the AI drug development landscape, with like AlphaFold kind of established now, where is the next sort of white space for development in the next couple of years?

Yes. Well, it's just -- it's absolutely massive. So if you think about all of the pharma industry, all of the biotech industry for as long as they have both been running and all of the approved medicines and all of the drugs that are currently in clinical trials, you only cover a little over 10% of the genome. And so that means almost 90% of biology is sort of in the dark to us from a therapeutic perspective. And the chemical space is actually in a similar state, if not even worse.

And so first of all, there's a lot of white space that we need to explore. And that's part of why we need new techniques to be able to evaluate it. But I really think where we're going to continue to see evolution is, one, over the next couple of years, well, where we are right now, people understand that patient selection is important, right? If you get the right patients into the trial, you can increase probability of success. And obviously, we've got a lot on that in ClinTech. What I think people are just starting to understand is the chemistry space, which is really potency and even selectivity are only a small piece of the puzzle because realistically, how is it absorbed? How is it metabolized? What are the other interactions that you're dealing with.

And so we, in legacy Exscientia had predicted multiple clinical trials that failed before there was any data on the clinical trials, just looking at the chemistry. And so I think people are just starting to understand, okay, a potent compound that works in mice, which is, by the way, a preclinical model that's basically designed just to focus on potency does not actually take into picture the entire patient environment and all of the biology. And so understanding that side from chemistry, I think, is just starting to come up.

And then the biology side of it, people still think of very linearly. ABC1 connects to XYZ2, and that's how biology works, which, of course, is not true. It's an incredibly complex multivariate problem. And so I think we are just starting to get the tools to be able to explore that better and be able to get into areas that are far off that what I was talking about earlier was really exciting. Virtual cell is obviously a big word and often used. But the publication that we just had in Nature Biotech, basically, what it showed is for a number of different cell lines, we were able to experimentally predict what would happen in that cell without having to be at all part of the training data.

And so we're actually getting to the point where you can use multimodal biology. So cellular phenomics, which is what Recursion was originally founded on, combined with transcriptomics, combined with other different aspects that you can bring in. And all of a sudden, that signal that's so cloudy in any single medium becomes much more clear and you can make predictions on actual underlying biology even without having original knowledge. Still a long way to go there, but we're just tip of the iceberg on it.

Yes. And I mean the way I think about AlphaFold is you had a good data set, and there's not that many. I mean, there's a lot of different ways that proteins can fold and interact, but it's not as complicated as like neutrons and electrons in like small molecules, right? So like the multimodality of that made it an easier first step actually, which was a little bit counterintuitive. Is the right way to be thinking about it that you're just going up in orders of complexity going from proteins to small molecules and then to a virtual cell?

Well, it's really interesting. I think there is going to be fits and starts across all the categories, but you're absolutely right in a way, like the AlphaFold, which, by the way, completely amazing and fantastic discovery, but it was based on 30 years of annotated protein data that existed in the public space so that there was a much larger base to be able to build from. And then being able to put that together, you're basically looking at physics principles and how that combines together and it becomes a computational statistical problem.

With chemistry, you're talking about -- so just to put it into context, like if you take all of the antibodies, for example, that are known and out there, you're talking about 10 to the 15th about. If you take all the potential medicinal chemistries, you're talking about something around 10 to the 60th, so effectively from what we're thinking about. The biology, obviously, you're interacting between all of those different pieces. But I think what's really important is AlphaFold itself didn't solve all proteins. It may be far better at estimating and predicting what then you can prove out experimentally later on in proteins that are not well known, but the prediction capability is far higher in areas where there's a lot known about proteins.

And the same is true in chemistry, like generative systems work better around areas where there has been a lot of data created. And so what you'll see is there's going to be aspects of biology, let's call it, the most complex because you're now combining all of those different tenets of whatever. And you're going to have amazing breakthroughs in that, but they're going to be around areas that you can build off of, just like we're continuing to advance in proteins and continuing to advance in chemistry.

Great. That's really interesting. But yes, maybe we can move from the theoretical now to the practical because this is...

You got me going.

Yes, I think that's super interesting. But in terms of what moves stocks, including Recursion, tends to be the tangible aspect of what's coming out of the platform. And for you guys, that would be your clinical and preclinical programs. So maybe we can just go down the list because you've got a bunch of interesting programs in your pipeline. Maybe starting with 4881 in FAP. For those less familiar, maybe just talk a little bit about sort of what this disease is. There's really no approved therapies here, although there's maybe some context that you can put 4881 into that makes sense from a biologic perspective. And sort of what's the clinical course for these patients? And what are you trying to solve for with treatment?

Yes. Really, really tough disease state. So this is usually identified hereditary, though there probably are a lot of somatic mutations that cause this as well. But 50,000 plus, and that's primarily the hereditary population who, throughout their entire life as long as they live, they're going to continue to get hundreds or even thousands of malignant polyps throughout their GI tract. And so this usually starts up in adolescents or teens. Patients will usually have a colectomy before they're 30, and then they'll continue to get further and further resections as time goes on. They're likely going in and having excisions of all the polyps can be on a several month basis. So really, really difficult disease state, very, very high comorbidities and completely chronic.

So what we've been able to show with the data is within 3 months of treatment, the polyp burden was reduced by about 50%. And then we were able to actually also take patients off of drug and maintain that polyp benefit, which is really, really important for a chronic disease drug. Those aspects basically gave us confidence that we would be able to move forward with a registrational pathway. And so we're currently in FDA discussions on starting the pivotal trial, and we'll give another update on that later on this year. But huge unmet need. There's no approved therapies for it. And we're really excited because that was a discovery that came out of our Bio AI platform. It was a novel connection that we were able to look and say, okay, how can we potentially reverse this APC-driven mutational effect with a drug and then convert that drug into now patient proof of concept?

Yes. And I guess as we're thinking about the translation of polyp burden to functional outcomes for patients, I know you've got a pretty substantial kind of natural history set that really helps you understand sort of the patients with this disease. I guess, how should we kind of bridge that gap between the 50% polyp burden to functional endpoints like I don't know, CRC prevention and others. And what's most important for the FDA, I guess?

Absolutely. Well, so every single one of those polyps is precancer. And so basically, if left untreated, all of these patients will progress into colorectal cancer. And so that's obviously one of the endpoints. But there's a number of different endpoints that you can look at with the FDA. There are things like there's a Spigelman score, which is sort of a composite of polyp burden and dysplasia that leads clinical assessment and treatment.

There are excisions because when someone without FAP goes in for a colonoscopy and gets a polyp or 2 removed, it's painful, it's uncomfortable. When you're going in every 3 or 6 or 9 months and having 100 polyps removed, you're talking about massive excisional burden, bleeding, infections, quality of life changes. And so you can look at those pieces. You can look at resections and serious surgeries. These polyps actually can go up into the duodenum and you can't do like a colonoscopy to be able to remove those polyps. You actually need a different endoscopic procedure, which is a far more serious procedure. So there's a lot of different true clinical endpoints.

If you talk to the clinicians, they view polyp burden as being incredibly important because it basically determines how they stage patients. The other thing that we were able to do, and this is sort of an example of how we have the infrastructure because we have a supercomputer, because we have AI scientists, because we have access to 300 million patients of real-world data, literally, in the course of a week, we created an LLM that looked across those 300,000 patients, found 250,000 patient records associated with FAP created something that we could query and say, okay, what is the standard of care in this setting? How many surgeries are these patients getting? The answer is 10, by the way, across their lifetime of serious surgeries, not counting all of the colonoscopies. And that gives us a lot of insight that never existed before that we can also take to the FDA and talk to them about the pathway.

Yes. I guess how essential is the real-world data package to your FDA discussions? And does this reduce the likelihood of a randomized controlled trial?

Yes. Well, I won't get in front of the FDA discussions, certainly. But what I'd say is it's definitely informative, definitely appreciated. I think even if it doesn't -- so the FDA loves data and loves transparency. Anyone who thinks differently hasn't worked with them, they love it. But even if it doesn't make a difference with our FDA pathway, it makes a big difference with how we design that trial, which patients we go after. We can do things like take a patient population and change our inclusion/exclusion criteria and say, okay, first of all, how does this change my patient population and commercial opportunity across the U.S. and Europe. And then we can say where are the sites with patients that have this sort of demographic and then we can build in there. That's why we've seen a 30% to 60% improvement in our enrollment rates across our clinical trials where we're using our ClinTech. So it makes a big difference.

Yes. And that should pay dividends across the pipeline where you're designing new studies. I guess on RBM39, this is your molecule that can, I guess, take a side door to shutting down CDK12. We're expecting clinical data in the first half, I believe. What's sort of the first-in-human data that you need to show to give you conviction to advance? And I guess how quickly could you make this go on?

Yes, absolutely. So we RBM39, just to take a step back, was a program that we founded with our biologics platform. It was basically CDK12 has been a very interesting oncology target for a long time. It's transcriptionally related. And so there's obviously a lot of opportunity with high mutational burden cancers, but incredibly difficult to target because the pockets of CDK12 and 13 look the same. You get a ton of side effects with CDK13. What we were able to do is say, how can we identify the same biology in another target? And that was a novel connection. And then what sort of drug could then reverse that disease impact. And both of those were discovered phenotypically on our platform, and then we optimize the molecule.

So what we have been enrolling in is looking at patient populations that generally are going to be higher in high -- like MSI-high populations or different areas that are going to be potentially more susceptible to this type of drug. And we wanted to do like an early look, almost like a futility analysis and saying, okay, this is a completely novel target that no one's discovered. This is a degrader, which has its own benefits and potential concerns. And we're doing early dose escalation and trying to be really efficient with all of our money. I'm sure we'll talk about that at some point soon, too. And so what we wanted to do is take a look, and this was -- the first look we did was last week on our earnings call.

So basically, we had -- we showed that we had PK/PD that was dose proportional. We had the profile that we wanted to see, which was basically this is a protein target that you'd need to knock down probably 70% plus on a pretty constant rate throughout the day. And we were seeing that exactly how we had designed it to happen, which is great. And we're just getting close to what would be the predicted therapeutic doses. This is monotherapy, but there's potential to see signal coming out of monotherapy because if it was a cancer type that exposed to synthetically lethal mechanisms or other things like that, we could potentially see a signal. So that's where we're going next. The next readout will be a more thorough data set second half of this year and could be really, really exciting. That's actually a very large addressable patient population has a different sort of splicing and stability disorder.

Right. And I guess on that point, we'll be looking to see which tumors you select. I guess, is there a decision tree that you make either around which indications to pursue like unmet need vis-a-vis market opportunity? And then how do you make the decision to mono or combo...

As you will respect, we are incredibly data-driven. So we do a lot of real-world profiling and statistical analysis around that. We also do that for our commercial analysis, and we can do both of those aspects in-house. And then we're going to be looking scientifically at the data. We've got theories based on what we see coming out of the bio platform and what we believe to be true. And we're going to be looking in the data and say what holds up and what doesn't.

Okay. And maybe shifting over to 7735, which you announced, I think, earlier this year. This is your mutant selective PI3K inhibitor. Obviously, a pretty hot space right now with Incyte and others like Relay, et cetera. I guess -- here, I guess you have the potential to address some of the safety shortcomings with the hyperglycemia and stomatitis that's been seen. Is that really the leg you stand on from a competitive perspective? Or is there something else you expect to gain from the molecule? And I guess what is sort of the IND-enabling data you need to show to push this into the clinic?

Yes. Well, I mean, there are other things in the hyperglycemia, but that's a really big one because this is always sort of the hidden secret in oncology drugs is you see the results come up from a clinical trial, you should always go back and look at the inclusion/exclusion criteria. So most patients right now, and we did this on a real-world analysis as well. I mean, if you've got diabetes or prediabetes, which is actually in a lot of the cancer indications, around half of the patients you're not going to get this drug or you're not going to get a PI3K that are currently out there or you are likely to come off it very quickly. So the real-world time of being on, for example, Piqray is only a couple of months, and that's because of side effect profile. Even if you don't have diabetes or prediabetes, grade 3 hyperglycemia is a very serious issue. And so that is a really big important part.

I think what is also interesting is by having -- so we're seeing 100, 150x selectivity over wild type. Now what that means is an ability, we believe, to drive deeper into the dosing of it for the mutants. And so I think there's 2 levels of the question. One is that orphan population that is not eligible for these therapies right now because of some sort of glycemia issue. But the other aspect is, can we go deeper on dose to drive higher response rates in this population before you start to get to MTDs. And so we've seen a lot of drugs come up in the space. Our original thesis on this continues to hold. As far as we can tell, we are about an order of magnitude more selective than the drugs that are currently advancing through development. And so we like this drug.

Yes. So we'll stay tuned for the data from that program in the second half. I guess I want to ask about another arm of your business, which is partnerships. The 2 main ones that we hear talked about are Sanofi and Roche. And maybe we can roll this into sort of the capital allocation burn question, which is how could milestones potentially bolster your cash position over the next 12 to 18 months? How are these partnerships going? And how do you sort of see the allocation between your internal development, servicing your partnerships and being good stewards of capital?

Yes. And I'm going to back it out into the capital allocation question as the core part to answer there because I think as I said earlier, we're a therapeutics company, and we think the long-term value is going to come out of the drugs that we developed. And so if you think about how we invest in platform and what we do with our partnerships, it's got to be towards driving that bigger goal. And so first of all, on sort of the expense management, what we did post the merger was take a look at everything across the company and actually put an analytical framework to measure impact and basically said, anything that we can't clearly see high impact from or high probability from, cut it. That literally took out 35% of the budget over the last year.

And so we continue on with that framework now. Our expense guidance for '26, even with the 5 clinical, 2 preclinical, the big partnerships with Sanofi and Roche, it's still -- so it's less than $390 million. We're focused on getting to that less number because we are, at our heart, a tech and efficiency company. We should always be getting more impact for less cost. And we're going to continue doing that. We're also going to continue to be really disciplined in how we make decisions. So right now, most people don't realize this, about 2/3 of our cost, almost 70% is going directly into pipeline programs and our partnerships. So there's not some big 30% of the budget going into platform development or something like that. That's not how we do it. It's applied platform and pipeline development.

And so what we like to be able to do is say, okay, is the pipeline advancing? Is the partnership advancing? And if it's not, I'm going to cut all of the associated spend with that. If it is, great because it's going to be generating value on its own. And so from the partnership side, we've brought in over $500 million from those partnerships. Just in the last about 2 years, we've hit 7 partner milestones, 5 with Sanofi, 2 with Roche that continue to bring in more money. And over time, that business -- right now, we run it basically to be breakeven to a mild profit because our partners prepay us for our expenses. As we get to opt-ins on Sanofi, as we get into advanced states on Roche, actually, that becomes all profit that rolls in on a go-forward basis because we have no more operational obligations.

Okay. Great. Well, there's obviously more we could talk about. Always a lot going on at Recursion, but I think for time, we'll leave it there. So thank you, Ben, for the great discussion, and thanks, everyone, for your interest in Recursion. Thank you.

Thank you.

Good morning, everyone, and thank you for joining us. Since stepping into this role, I've been focused on a singular question: how do we harness the full power of AI to consistently and with urgency create better medicines for patients? That requires bold ambition and a lot of focus and discipline to create value for patients and shareholders. And therefore, our approach has been deliberate.

First, we're focusing on signal over noise, generating proof and proof points across our both wholly owned programs and our partner programs with the goal to showcase where AI can truly make a difference in creating value. Second, we are continuing to evolve our platform into a repeatable, AI-driven product engine, not tech for the sake of tech, but tech that creates products of value. And third, underpinning it all is a strong commitment to financial discipline and thoughtful capital allocation, ensuring we're constantly being data-driven to prioritize and invest in our highest conviction opportunities to deliver durable value.

Today, I'm excited to share some of our updates. We're making meaningful progress across all these fronts, which I, along with Vicki and Ben, will share more with you today. With that, let's dive in. But before we do, please note that today we'll be making forward-looking statements on this call, and therefore, please refer to our SEC filings for more information.

To put the progress in context, I think it's worth briefly stepping back to how we built the foundation to enable it. Look, Recursion has been on an intentional and, in many ways, a pioneering journey. Early on, Recursion recognized both the immense potential of AI in drug discovery as well as the reality that, unlike many other domains, the underlying foundation, whether it's data, compute, in biology, and broadly in science, is still being built. The map is still not complete, and that fundamentally changed how you think about applying AI.

So we made a deliberate choice to invest ahead of the curve, to generate and curate proprietary data, which we'll talk about more today; to build a scaled compute infrastructure; to integrate automation; and very importantly, to develop models that are purposeful and in a true closed-loop, lab-in-a-loop system, a phrase that has become much more in vogue now, designed not just to predict, but to test, validate, and continuously learn. That has led to a differentiated foundation, which we continue to expand and refine today. But in parallel, what's critically important is our focus now on translating that foundation into tangible proof, advancing programs, high-quality candidates, and overall demonstrating repeatability as we evolve toward a truly product-focused AI engine.

But let's make this all much more concrete. As a result, where are we today? What are some of the facts? So first, we have established our clinical proof of concept, our first clinical proof of concept, with our REC-4881 allosteric MEK1/2 inhibitor focused on FAP. We showed significant reduction in the precancerous polyps that are a huge driver of the progressive nature of the disease, as well as showing durability, something that's quite unique in the data we've shared to date. And why is this important? These are patients that have no therapeutic solutions to date and require life-altering surgeries and have near-inevitable CRC risk. This is a great example of how we can translate AI-directed insights from our platform into true outcomes. More on the latest there shortly.

But look, this is not just a single asset story at Recursion. We now have 5 wholly-owned programs, each with clear inflection points over the next 12 to 18 months, creating not just a consistent cadence of catalysts, but then also a way for us to test, learn, and also be disciplined in our areas of programs that we invest in. And we'll share more data from one of these programs, REC-1245, our RBM39 degrader. But this momentum is not just in our wholly-owned pipeline. It also extends in our partnered portfolio. With over $500 million in inflows, and more importantly, I would say, 10 milestones delivered to date, I underscore that, it's one thing to announce partnerships, but we are really focused on delivering value from these partnerships, including many of which are first in industry. This underscores a track record of delivering tangible, differentiated outcomes, and we are deeply grateful to our partners for their close collaboration in everything that we do.

Underpinning all of this is, of course, our platform, an end-to-end AI-native product engine across biology, chemistry, and clinical development, powered by proprietary data and a lab-in-the-loop system, and designed for repeatability. And I'll share some of the latest stats from our platform later in the presentation. And then importantly, look, we have to do this with focus and discipline, extending our runway into early 2028, while reducing our operating expenses by 30% year-over-year. This is how we are moving from promise to proof.

So let me walk you through how it all comes together. How do we pull this together for the ultimate goal of delivering better medicines for patients? At the foundation is an AI-native product engine that combines proprietary multimodal data, integrated wet and dry labs, purpose-built models, and scale compute. Now we hear those words a lot, but what differentiates us is not one model, it's not one data set, it's not one program. It's the integration of our tools, technologies, and our teams. Look, our proprietary multimodal data has both proprietary data that we have generated in our labs, which we also integrate with public data. We sit in the sweet spot of leveraging both. Our automated wet labs in Salt Lake City and Milton Park, Oxford, for those that are not as familiar with Milton Park, are interconnected with purpose-built AI models, and we have in-house supercompute resources to rapidly build those algorithms and learn from them.

And I have to say, and it's not just words, we truly mean it. Spanning all of this is our greatest resource, you've heard me say this over and over again, bilingual talent. AI researchers who appreciate the humility in making medicines and who bring a completely different take to how we can make medicines, and drug developers and drug hunters with reps under their belts that have seen what it really takes to make a drug from start to finish, and who are open-minded about unlocking the potential of AI. Make no mistake, the culture and the talent and the integration it takes is one of the hardest things to do in this space, and I'm excited that we've made so much great progress there.

Now all these ingredients come together in a vertically integrated AI native platform, starting first, biology. So we can simulate and understand biology much more effectively. And we really want to move away from the stats that we only -- the industry only understands about 10% of biology. This is what allows us to identify novel targets. This is where we're pushing the boundaries to really understand the root cause of disease. The next click, this really came from the integration with Exscientia, applying generative chemistry and active learning and many other approaches to design precisionly created, differentiated molecules. This is what helps us create both first-in-class programs for those novel targets, as well as really high-value, best-in-class programs. For instance, optimizing therapeutic index for programs that have been around but haven't fully maximized their potential to date. And third, this is something we've built over the last year or so, applying our data and insights to also inform a smarter, more effective, and patient-centric path to all of clinical development.

Look, all of this is great to have, but we take it together to build a broad and diversified portfolio, both internally and with our close partners, with the ultimate goal of developing differentiated medicines for patients with significant unmet need. We do it faster, and we want to do it better. But how do we do this? Our strategy remains unchanged from what you've heard the last time. We want to be clear, focused, disciplined, while being ambitious. First, translating proof to products. We are advancing our deep pipeline learnings, and the goal is to have revenue-generating medicines for patients. And we do it by applying a rigorous data-driven prioritization approach. so that we only invest behind the most highest confidence opportunities.

Second, as you heard me say, scaling a differentiated AI-native product engine. Look, the platform is the heartbeat of so much that we do, where each prediction and experimentation allows us to compound our learning and advantage to drive repeatability in creating better products.

And third is pairing that bold ambition with disciplined execution. Rigorous capital allocation is something we think about constantly, ensuring that there's operational focus and that our milestones are measurable to sustain that long-term value creation. You'll hear more about that shortly. So let's just dive in into one of our first pillars, which is our wholly-owned pipeline. Look, I'm proud to share how this strategy is beginning to translate into early signals of pipeline progress. What you see here is a broad and increasingly diverse set of programs built on 2 key areas: number one, clear rationale for differentiation, that's coming from a platform; and second, a defined path -- a rapid and defined path, I should say, to upcoming milestones and decision points.

And the differentiation across these programs takes 2 forms: one, in some cases, it starts with a novel biological target or novel mechanistic insights. You'll see more of these coming from our discovery part of our platform; and then the other is driven by differentiated molecular design. And then the third, more recently, as we built up the clinical development AI platform, how we design, which patients do we pick, how do we design our protocols, and how do we execute in the clinic.

Let me double-click some of the latest highlights from the last quarter on these slides, a period that is marked by strong and accelerating clinical momentum. So let's start with REC-4881. This is our allosteric MEK1/2 inhibitor. As you recall, this program is rooted in a novel mechanistic insight with the potential to become the first precision therapy for FAP. As I mentioned earlier, and I can't mention it enough, a serious and [ underserved ] condition where patients often face very limited treatment options and no medical or therapeutic options to date. We have generated compelling proof of concept, and we're continuing to advance the program with urgency and vigor, including we've already initiated FDA engagement to define a potential registrational path forward. We're very excited to share more update on this in the second half of this year.

Next, turning to REC-1245. This is our platform-derived first-in-class target and degrader with the potential to address multiple solid tumors and lymphoma. We're excited to share today early clinical data around the safety, tolerability, and PK profile, as promised. To date, we have observed a well-tolerated profile with no dose-limiting toxicities to date. And we're continuing to advance the program with additional data expected later this year. In a few minutes, Dr. Vicki Goodman, our Chief Medical Officer, will walk through the details -- more in detail.

And finally, REC-4539. This is our LSD1 inhibitor for the potential treatment in solid tumors, including small cell lung cancer and also in AML. What differentiates this program is the underlying molecule designed with our generative platform to overcome some of the treatment-limiting on-target toxicity seen to date in earlier LSD1 inhibitors. We've now initiated our Phase I clinical trial and dosed our first patient with additional updates coming second half of 2027. I'll talk more about the program, the biology, unmet need, as well as the platform insight shortly.

All of the other programs remain on track. Now we're also continuing to see strong, consistent execution across our partner pipeline, where our platform is being applied in close partnership with our esteemed partners whose deep expertise, collaborations, and capabilities, we are deeply grateful for. And what's emerging, I want to highlight, is 2 potential unlocks. As an example with Sanofi, the unlock is use of AI on the chemistry design side, taking difficult and diverse protein targets in immunology and oncology, using our platform and AI in partnership with Sanofi to drug these challenging -- historically challenging targets. These programs are progressing towards key inflection points over the next 12 months, including a potential for development candidate, which is a big unlock in terms of potentially onboarding that asset to partners' portfolio.

And with Roche and Genentech, the unlock is on the biology side. Roche and Genentech have been pioneers in really thinking about leveraging biology perturbation at scale to really take large-scale multimodal maps and translate them into actionable and validated programs. So the unlock here is you hear a lot around large-scale data sets being generated across the industry. Well, the unlock is how do we translate that using foundational models that we're building and robust experimental target validation into not only validated targets, but potential first-in-class programs, something the field has long aspired to do. So we have a potential first on track in the next 12 months or so.

Look we've talked a lot about our wholly owned programs, our partnered programs, excited about the momentum we're building here, and about our platform that underscores it. But the secret sauce of any organization is talent. Talent is critical to everything we do, and continuing to build a strong, experienced, ambitious, and humble team is a key part of how we drive value. And with that, I'm really pleased to introduce newest member of our executive leadership team, Dr. Vicki Goodman, our new CMO. Vicki comes to us with an incredibly strong track record of delivering transformational medicines for patients across many parts of the industry, starting at the FDA, large pharma, and biotech. You can read about all her credentials on the slide, which I won't go through in detail. But simply put, Vicki is the right person with the right skill set to lead Recursion's clinical development in this next chapter of the journey, but more importantly, with the right heart and perseverance to go through the trials and tribulations that's drug discovery and development.

With that, I'm going to turn it over to Vicki. Vicki, why don't you kick us off with a few words about joining Recursion and then more details about REC-1245.

Thank you, Najat, for the kind introduction and for the opportunity to work with you and the rest of the Recursion team. One of the reasons I joined Recursion is because the breadth and differentiation of our pipeline represents one of the most exciting opportunities to translate AI advances into meaningful therapies for patients. Today marks exactly 1 month since I joined. And even in that short time, I've found Recursion to be a place where scientific rigor, intellectual curiosity, and a deep spirit of innovation are brought to bear on the creation of new medicines that matter. It's wonderful to be part of the team, and I look forward to continuing this important work.

Today, I have the privilege of presenting an exciting clinical update for REC-1245 from the ongoing DAHLIA Phase I study, including preliminary safety and pharmacokinetics. REC-1245 is an RBM39 degrader currently in Phase I for the treatment of patients with solid tumors and lymphomas. RBM39 is a novel target, which plays a central role in splicing fidelity. When RBM39 is degraded, it induces widespread splicing defects to which tumors that are already under stress, such as those with DNA damage repair deficiencies, global genomic instability, or replication stress, may be particularly sensitive.

Additionally, RBM39 is highly expressed in certain tumors and is associated with disease progression and poor survival. The relevant patient population is estimated to be over 100,000 patients in the U.S. and EU5. So that's the why RBM39 is an interesting target. The how we came to be working on RBM39 is a story we've touched on before. It's an example of how the biology element of our AI-driven platform enables the identification of novel therapeutic targets. Using genome-scale phenomic mapping, our maps of biology, RBM39 emerged as a functional analog of CDK12. This novel relationship, which came from an unbiased platform insight, was not obvious from sequence homology or traditional pathway analysis.

CDK12 is a well-known oncology target for its role in DNA damage response modulation, but it has generally suffered from challenges in selectivity because of how homologous to CDK13 it is. Following our insight, we developed molecular glues and degraders for RBM39, and we showed that these phenotypically mimic CDK12 loss but not CDK13. This provides a druggable potential analog for CDK12 without the CDK13-driven toxicity. We progressed from target ID to IND-enabling studies with roughly 200 compounds synthesized in 18 months, which is significantly faster than traditional approaches. We then needed to correlate our insights with the mechanism of action for RBM39 to translate them into clinically actionable hypotheses. We confirmed through in vitro studies that there is a greater sensitivity to REC-1245 in cell lines that have higher replication stress and DNA repair vulnerability versus cell lines that don't have higher replication stress. And in the panel on the right, you can also see that in vivo tumor regression in an MSI-high ovarian CDX model was also demonstrated.

We've carried these insights forward into the design of our DAHLIA Phase I/II clinical trial. Our early clinical strategy focuses on tumor types with those same characteristics that suggested sensitivity in our preclinical experiments. The safety and PK data we are sharing today is from 16 patients enrolled across the first 4 dose levels. All patients have advanced solid tumors, and 7 of the 16 have MSI-high or mismatch repair deficient tumors. Importantly, REC-1245 is well-tolerated. Across the dose levels evaluated to date, there have been no dose-limiting toxicities reported. The most common treatment-related adverse events that have been observed are GI-related, constipation, nausea, and vomiting. As you can see, these are generally low grade with one grade 3 event of nausea and vomiting reported. There have been no treatment-related serious adverse events. Dose escalation is ongoing and recruitment is on track.

We have an early PK/PD summary from the evaluated patients to date, and we'll have more dose escalation data and a fuller PK/PD update in the second half of the year. So far, we are seeing predictable dose-dependent exposure with exposures continuing to increase as we move through the dose levels and PK data that are supportive of daily dosing. Our initial PD data also confirm target engagement. We expect, as we move through the next 2 dose levels, to see exposures that are correlated with tumor regressions in mice. Overall, RBM39 represents an end-to-end example of how we're using AI to translate a novel insight into a potential medicine, not just identifying a target, but building a coherent biological hypothesis that informs clinical strategy. I look forward to sharing more data with you later this year.

And with that, I'll turn it back to Najat.

Thank you, Vicki. Now moving on to LSD1, REC-4539. First, I'm pleased to share and announce that we have dosed the first patient in our Phase I clinical trial. But taking a step back, let's discuss a little bit as to why we think that LSD1 is an interesting target for Recursion. As many of you know, LSD1 is an epigenetic regulator with a range of cellular functions and a promising oncology target across multiple cancer types. However, so far, clinically, the potential of LSD1 inhibitors have not been fully met. Previous clinical attempts to drug LSD1 have shown some efficacy, but have been limited by on-target and dose-limiting thrombocytopenia. So while the biology is understood, the challenge has been at the level of the molecule.

And therefore, we believe this has the potential to unlock a meaningful therapeutic opportunity, particularly in settings like extensive stage small cell lung cancer, where there are approximately 45,000 patients in the U.S. and EU5 with emerging but still limited treatment options after progression on first-line therapy. So for this program, the starting point model. And this is where our chemistry AI part of the platform really shines. We intentionally moved away from traditional bias chemistry, chemical space, and instead used a blank white sheet, active learning to explore a broader or information-rich space, which allowed us to identify novel starting points that wouldn't typically be pursued. What that led us to is identifying a new scaffold. And we iteratively refined it, ultimately arriving at the compound REC-4539, in approximately 20 months in just over 400 synthesized compounds, much faster and fewer than what is industry standard. Compare that also to what Vicki had mentioned with RBM39, 18 months and about 209 compounds synthesized to date.

So you're starting to see these, as I like to call them, green shoots in terms of the number of compounds synthesized, the speed, and also the efficiency, and how we're generating compounds. The focus here, though, was designing a molecule with properties that directly address the limitations we've seen in this class to date, specifically reversibility and a shorter predicted human half-life to potentially reduce the risk of cytopenias that have been one of the dose-limiting factors for prior LSD1 inhibitors. And we're sharing here also some preclinical data in small cell lung cancer that demonstrate that this compound had more minimal impact on platelets while maintaining efficacy, which is not shown here, but we've seen that data as well, while being comparable in efficacy to other agents in the class, but having more minimal impact on platelets versus other agents in this class.

In addition, there's also another feature of this compound. It's brain penetrant, which may be particularly relevant for patients with small cell lung cancer where up to 50% develop brain mets. This differentiation and the potential to improve tolerability has encouraged us to advance the compound into Phase I. We had our first patient dosed in April. This is a dose escalation study in select solid tumors, including small cell lung cancer, with expansion cohorts planned following the initial escalation phase. And I want to make one thing really clear. Vicki and team are structuring these programs to enable rapid data-driven decision-making. This is how we really manage our capital allocation, specifically to address rapidly whether the emerging clinical profile really supports the hypothesis of mitigating or reducing the risk of thrombocytopenia. We expect to share initial PK and safety data in the second half of 2027.

So look, I won't -- I think I've covered most of this, but similar to what Vicki shared, here's another example where we use our AI platform to solve for design challenges around a biology that's more validated, optimizing molecules where we believe there's been limitations to date. And recall, there are no FDA-approved LSD1 inhibitors to date, despite a well-defined patient population and significant unmet need that remains for patients. We look forward to sharing more clinical data for this program next year.

Now let's go to our second pillar, which is incredibly important. Beyond our portfolio, underpinning a lot of what we're doing is continuing to advance our end-to-end AI product engine, pushing the boundaries. So we continue to remain at the forefront of AI-driven innovation. Let me walk you through the platform and also some of the facts and the stats around how we're doing. We built the Recursion platform specifically to address the most persistent bottlenecks in drug discovery and development. And so we're always looking at how we're doing versus industry. So let's start with biology. Look, we have generated more than 10 high-dimensional maps of disease biology. And what's interesting here is more than half of them have been in partnership with Roche and Genentech. And these are already driving multiple novel programs that will target programs in our internal pipeline and then we're also working actively with our partners at Roche and Genentech to translate these maps into novel targets and first-in-class programs. This is an important unlock.

Why are these maps important? Biology is a systems level approach. We need to understand the interconnected circuits. And therefore, enhancing our ability to identify and prioritize targets with better confidence, with better understanding of the underlying biology is critical to really determining the root causes. If we go to the next slide, we are also synthesizing, as you heard me share briefly, more and more compounds where we are designing 90% or synthesizing 90% fewer compounds than the industry benchmark. So about 330 compounds on average versus 2,500 to 5,000 compounds, which is the industry standard today, while also advancing these programs to advance in development candidates roughly twice as fast. This is a meaningful step change in both efficiency and cycle time, and something that we watch very carefully on an ongoing basis.

The next point are ClinTech capabilities. Where we have deployed it, we're already seeing about 30% to 60% faster trial enrollment. This is very important for us, both for rare diseases and competitive areas such as in oncology, while increasing the eligible patient population for some of these programs from 10% to 40%. This directly impacts our time lines and speed at which we can generate high-quality clinical data. And the underpinning it all is an integrated platform with more than 50 petabytes of proprietary multimodal data. This is incredibly critical for not just building purposeful models, but ensuring that we have our data moat that we not just invest in, but continue to expand.

So suffice to say, these are not theoretical or isolated improvements. These are real, tangible gains that we keep measuring and focusing on to reinforce how our platform is changing the way we are discovering and developing our medicines. Again, I always like to say there are green shoots, but this is how we are pushing the frontier of what's possible with our platform.

Now let me double-click on a couple of recent examples in the biology layer where we're really pushing the next generation of our models that were recently published. Big picture, one area of focus for us in our biology platform is learning the language of biology. Sounds simple, not easy, incredibly important. And we do it across many data layers by generating perturbations at scale, whether it be genetic, chemical, and so forth. Why is that important? What is our goal? Our goal is to, a, understand biology more comprehensively; b, and then be able to predict and simulate perturbations before we are even running a single experiment. And third, that we can actually generalize beyond the data that we've already seen. That's really important with these foundation models. You want to predict responses in new, out-of-distribution contexts, such as novel targets, combinations, and cell types.

And why does it all matter? Given the vastness of the biological space and how little best in industry know 10%, this has the potential to unlock areas that remain intractable today. And that becomes even more powerful when we connect different data layers, high-content cell imaging, transcriptomics, proteomics, patient data, and more, to really have a more unified view of biology. Now it's not just in theory. We're making progress here already. Step 1 is to actually develop a new generation of models. So let me share with you 2 recent advances in transcriptomics foundation models that we just published in the last month.

The first is TxPert, which we recently published in Nature Biotech. TxPert is a model designed to predict how gene expression changes in response to different perturbations, essentially helping us to understand biology and how it will respond before we run the experiments. So similar to what you saw in chemistry, the reason why we can reduce the number of compounds we synthesize is because we predict and simulate more, and then, of course, we make less. And how can you do that more in biology? It's important for us to understand the systems approach in biology across different data layers and be able to predict well so that we can do less experimentation and only do the experimentation that really matters. What's particularly exciting about this model, and listen, we're still in early days, but it's great to see the progress from our teams, is learning patterns in biology. It's not just memorizing, it's actually learning the underlying patterns.

The second is generalizing beyond the data it was trained on. This is an important start. It's a start to predicting responses to new perturbations, new combinations, and even new cell types. Watch this area more. I think this is going to be really, really important in biology and foundation models, just given the vastness of what we're working with. And it's an important first step towards how we think about building a virtual cell, a term that is overused. But the importance of it is, again, can we simulate and explore biology more comprehensively, more computationally before we move into the lab? This is incredibly important so that we can be more effective and efficient and ultimately improve the probability of the targets that we could put into our programs.

Next is another model, which is complementary. TxFM, our transcriptomics foundation model, which represents a significant step of actually connecting lab biology to patient biology. I won't go into the details here, but just want to highlight a couple of things. First, it is built on a highly curated combination of both proprietary and public data, bringing together diverse data sets into a shared representation space. Why is this important? There's a lot of conversations, is quantity important or quality. Both matters. So you'll see from some of the early insights here that the quality and the model architecture was really important to ensure great model performance here.

So what's exciting is the following: number one, the result is a model that surfaces much richer understanding of biology and reducing experimental noise. Batch effects and so forth, very, very important in this space; number two is it outperforms a lot of leading foundation models. But more importantly, it outperforms models that are trained on 100x larger datasets, demonstrating that advantage I was mentioning on just the data, data curation approach, and also model architecture.

And what I like is the interpretability. That's where we're starting to go. And again, early days. It doesn't just rank genes, but it reveals the gene networks, the circuits, the patient subtypes from RNA data, so that transcriptomics can in time become a more systematic engine for understanding the mechanistic and target hypotheses that underpinning just one-off analysis. There is so much richness data, and we're driving to understand that even better. And practically, again, it's how do we get more efficient in the experiments we run? How can we do less reruns? How can we do better cross-study comparison and efficiently use our resources?

So today, both these models are starting to be deployed in our platform. For TxFM, we're starting to leverage it for target identification, better mechanistic understanding, and patient stratification. So sharing some latest and greatest here, and we will, as always, in the months to come and years to come, share how this is truly impacting our platform. That's what we care about. How do we take data models to really show the translation of proof into our programs, into our partner programs, and progress better medicines for patients.

With that, I'm going to turn over for our third pillar, which is how do we drive all of this important work with good discipline and good ambition. Ben?

Thank you, Najat. Our core focus from a financial perspective is ensuring we have adequate runway to achieve multiple upcoming milestones. We continued our trend of operating discipline with a 30% year-over-year reduction in cash operating expenses. We were able to achieve these savings while also growing our pipeline, partnerships, and platform by focusing only on those operations that had clear and measurable impact. In addition to our operational discipline and infrastructure simplification, we also expect ongoing efficiency gains from our technology advancements and the adoption of agents. During the quarter, we received our fifth milestone from Sanofi, advancing a potential first-in-class program for a novel biological target. We closed the quarter with $665 million in cash and equivalents, which we believe provides operating runway through early 2028 without additional financing. For 2026, we are maintaining our cash operating expense guidance of less than $390 million, which fully funds our expected milestones and partnerships during the period.

And to take you through those milestones, I'll turn it back over to Najat.

Thank you so much, Ben. And look, I'm going to close by saying we have a lot of important work ahead of us and very exciting work ahead of us. As we look ahead, we have a clear and consistent cadence of milestones, both across our wholly owned pipeline and our partner portfolio. In our wholly owned pipeline, we expect multiple clinical readouts over the next 12 to 18 months; in fact, for every single one of our clinical stage programs. Continuing to build on clinical evidence and test the hypothesis underlying our platform. In parallel, we're seeing continued progress across our partnered portfolio.

I'll recap the 2 potential unlocks I mentioned: one, looking at the use of AI to develop novel compounds for difficult-to-drug targets. We're excited about our work with Sanofi here and some of the development candidate decisions coming up in the next 12 to 18 months. And with Roche and Genentech, to take all of these large, multimodal maps that's helping us understand biology better and really translating that to novel targets and first-in-class programs.

Taken together, this creates a diversified sets of catalysts and also the increasing momentum, as you're seeing, month-over-month, week-over-week, as we look to take and harness all of what AI and our dramatically excellent team can do to turn that into meaningful outcomes. I'll just close by saying we are focused on building an increasing body of evidence that this approach can translate, advancing differentiated programs, unlocking new biology, and doing that work with improved speed and efficiency. And that's what gives us confidence in the path forward. The momentum you see, the work that the teams are doing, but also the system behind it and its potential to generate outcomes over time in a repeatable fashion for patients, our partners, and our shareholders.

Thank you again for your time and attention, and we will open it up now for questions.

Great. So let's dive in. I have Vicki and Ben, who will help me cover some of these questions. So the first question is from Dennis at Jefferies and Priyanka at JPM. On the REC-1245 program. Can you talk about the level of target engagement that you feel is needed to drive efficacy and where you are relative to those levels? What are common on-target safety and tolerability issues that you're hoping to avoid with your approach? And how are you thinking about biomarkers being explored?

Well, maybe I'll just kick it off and then I'll hand it over to Vicki to also share additional details. I'll go in order of second, third, and first. So what are common on-target safety tolerability issues that you're hoping to avoid? Look, first of all, we are encouraged by the favorable safety and tolerability profile that we see to date. As you saw, 90% of what we see so far are grade 1 and 2, mostly GI, and no DLTs to date. In terms of just RBM, there are areas that we would usually keep an eye on in terms of potential tox is heme tox. And to date, we have not seen any grade 3 heme tox at all. So that is encouraging. But again, we're in the middle of dose escalation. So more to come, as Vicki mentioned, second half of 2027.

And in terms of target engagement, we're already -- we have some PD data that we shared, and Vicki can share more about that. But we're seeing good target engagement. We've confirmed that to date. As we have more dose escalation, what we've seen preclinically is about 70% to 80% was sufficient at efficacious doses, but we'll be tracking that as we continue further. Vicki, did you want to add anything more to those 2 questions?

So coming back to the safety and tolerability issues, I mean, again, what we've seen so far is mostly low-grade GI tox. We'll certainly continue to monitor that as we move forward. Hematologic toxicity, which is a concern here, is something we're really not seeing at this point. Again, we'll continue to monitor as we continue to increase the dose, and we'll have more data for you there in the second half of this year. Relative to target engagement, I think the estimates are spot on. I'll add that in -- we're coming close now within the next 2 dose levels to being at the exposure levels where we saw tumor regressions in mice. So I think that's an important point as well. Obviously, we'll continue to monitor the target engagement in terms of RBM39 degradation and, again, have more data, more fulsome update in the second half of the year.

Thank you, Vicki. And just the last question was, how are you thinking about biomarkers being explored? As Vicki mentioned during the presentation, we're looking across select biomarkers. And, of course, as the data matures, we will look at relative benefits versus not across those biomarkers. More to come second half of 2026. Thank you for the great question.

All right, next question from Gil at Needham, Alec from Bank of America, Sean from Morgan Stanley, Brendan from Cowen, and others on REC-4881. Given the encouraging Phase II data for REC-4881 in FAP and ongoing FDA engagements, what are the key uncertainties around the registrational pathway? We have 3 questions here. I'll just start one at a time, so we can keep track. I'll kick it off, but Vicki, it would be great to get your thoughts. But we're very excited about the data that we see with FAP. And with every day that goes by, we engage with more FAP patients, really not just the unmet need, but how underserved these patients are is becoming even more and more apparent. We have a significant polyp burden reduction and durability that we've seen to date. I would say the main areas of focus with the FDA is what would be for any asset that's a first in disease. We have other assets in our portfolio that are best-in-class, where the regulatory approach is already very defined. For a first-in-disease, it's really around patient population, endpoint that has clinically meaningful benefit, and then, of course, dose and dose escalation. But those are the conversations. And as Vicki mentioned and I mentioned, we've already started that engagement. Anything to add?

Yes. So I think here, because there really is a lack of regulatory precedent, it's really important for us to work closely with the FDA in terms of defining the registrational path. To that end, we've already initiated that engagement within the oncology review division, also requested input from the GI division. And certainly, as we go about these, we're thinking about leveraging the rare disease framework as well. So we can derisk this program by really closely aligning with FDA on what are clinically meaningful endpoints for patients that will help us define the primary endpoint for our pivotal study.

The next question is still on REC-4881. Has there been any shift to timing for FAP regulatory? And when will we see additional data? So a couple of things. We're on track with it. We'd initiate FDA engagement first half of 2026. We're actually a bit ahead of schedule. We have initiated FDA engagement. And as Vicki mentioned, we expect us to be working with the FDA very closely, given it's a first-in-disease on our potential registrational study. And then when will we see additional data? Vicki, do you want to share that?

I'm happy to take it. In terms of additional data, we have already initiated 18 and over patients. We're already recruiting those patients. And we'll also have potentially additional data from our Phase II that we will share either here or at a forum going forward. But we're on track. Have you leveraged any ClinDev capabilities from your platform in assembling the proposed pivotal study design? Great question. A couple of areas. Number one, if you recall, the natural history that we did in parallel with our clinical program has been really, really important for a few reasons. This is a rare disease, limited literature, natural history. This has allowed us to not just understand patient trajectory, but also helped us as we think about how do we power the study, what endpoints are important, and so forth. So it gives us a much richer contextualization understanding, and then also incorporating that in terms of our study design as well.

In addition to that, for our 18 and over plus our potential registrational study, we are also going to be using our clinical development AI capabilities for recruitment. This is a rare disease. We want to be much more efficient and use these approaches to go to where patients are and recruit with speed and rigor.

Great. Next question from [ Bruce, Philip, Rishabh ], and others on the platform. How does Recursion evaluate whether its platform is improving its ability to identify and eliminate lower-quality candidates earlier in the discovery process compared with prior years? Let me answer that one first. So as I shared earlier today, we look at every segment of our platform, and we're really looking at how is it that we can design better molecules faster.

One of the things is you asked about lower quality candidates. This is where if you can actually simulate more, predict which compounds would actually have better versus worse ADMET properties. But this is where active learning and the multiparameter optimization, that complexity that we can do it in a very -- industry can do in a very sequential way, we do it in a much more efficient way. We simulate online and only synthesize the compounds that we have confidence in. That's where you see some of the numbers shifting pretty dramatically, cycle times becoming half and then also 90% less compounds synthesized. So that's just one example as to how we track it.

The other thing I would say on the biology side, look, the maps of biology give you a lot of hypotheses in terms of potential novel targets. But we pair that with really robust experimental validation. I think that is incredibly important to do both. That allows us to look at targets that no one has looked at before. This is where novel biology is coming from. But we always pair that with rigorous experimentation. And that's where the lab, the wet and dry lab, is incredibly important for us because we can do it at speed, we can learn fast, and all that data is captured to make our models better. So we are a continuous and rapid learning organization. And I will say the integration with Exscientia has really helps there, because now you have both the biology and the chemistry side sitting side by side. And we iterate and learn from that.

Okay. What are the investments you're looking to make on the platform? Compute, data generation models. Is this question from one of our leaders -- AI leaders in the company? I'm just kidding. Look, we are -- our strategy is we invest in our programs with our platform, and that platform needs to remain differentiated. So we surgically invest in areas that matter. So as an example, on the clinical development side, you've seen the investment we have made, but there's a reason. It's to make a better product. So how do we recruit faster, how do we pick the right patients.

In our chemistry and design platform, we're continuously evolving and iterating on our models. And we'll share more in due time. And then in our biology platform, you've seen the investments we're making in state-of-the-art transcriptomics models. But again, they're all with a purpose. How do we improve the target that we're putting into our programs, the compounds that are high quality, and ensuring that we execute our programs with flawless execution clinically, but then also pick the right patients and increase our signal to noise.

Maybe I'll take one more question. I want to take as many. I know we're a little bit over time. From Gil and Sean on partnership strategy. Any expected guidance for a potential clinical opt-in from partners? Will we receive an update on this?

So actually, Ben, do you want to share on that one?

Sure. Happy to. So as we continue to advance the programs along with Sanofi and begin to move different programs from Roche into the design phase, we absolutely expect to see some of those 5 programs that have hit their early discovery milestones move into the opt-in, and we're working very closely with Sanofi to make that happen as quickly as possible.

I think there's also a broader point that's really important around the partnerships. If you take a step back, and we get a lot of questions on what's our partnership strategy, where do we plan to go in the future, if you take a step back to what we do, part of our mandate is how do you create a risk diversified model for being able to be an investor in the biotech space. And so we obviously have transformative potential medicines that are coming up through our internal pipeline. But you also have to look at our partnership business and see how we've been able to advance programs and do it in a capital-efficient way and really diversify that risk and diversify what the long-term benefits of that are. So we will absolutely continue to drive that partnership forward along with our internal pipeline, and we'll balance out how we're getting the upfront payments while also still maintaining a lot of that downstream economics.

But suffice to say, Gil and Sean, we're working actively on this, and, of course, we'll share updates in the next 2 months. Last question. You've generated over $500 million in partner-related payments to date. How should we think about the forward trajectory of platform monetization, particularly the balance between near-term milestones and retaining long-term economics and wholly-owned programs?

I think it's very similar to what Ben said. Our platform is focused on generating better products. That's what we focus on, whether we do it internally or with partners. And we create optionality in terms of our wholly owned programs. Some of the programs, again, we're data-driven in our approaches in terms of could be wholly owned, could be partnered, could be outlicensed, and same goes for some of our partnered programs as well.

So with that, I'll just close by saying thank you so much for your time and attention. Thank you for all of the questions. We have a lot of momentum, a lot of important work ahead, and we continue to move that forward and excited to share more updates in the coming months and years as well. Thank you again.

Good afternoon, everyone, and thank you for joining us on the first day of Needham & Company Healthcare Conference. My name is Gil Blum, and I'm a senior biotech analyst here at Needham & Company. It is my pleasure to have with me today Recursion's Chief Financial Officer, Ben Taylor. As a reminder, any viewers who are watching through our conference portal are able to ask questions via the Ask a question box below the video feed window.

So Ben, maybe just starting with a bit of an introduction, setting the stage and just discussing some of the company's priorities in 2026.

Sure. Thanks, Gil, and really appreciate you having us at the conference. So '26 has been -- it's really, if you think about it, a year that last year was transition because we had the merger that happened towards the end of '24. And so we were doing a lot of reshaping the strategy, reshaping the financial structure, putting together the right pipeline. Now we've started to enter into that phase in '26, where we're looking at data readouts and partnership milestones on a more and more regular basis. So we already had our first proof-of-concept data readout with our FAP molecule and program moving forward, and now we're in FDA discussions on that.

We've got 4 more clinical programs that are moving in besides that -- behind that. But then also our partnership business, where we've already brought in over $500 million through our partnerships and moved 5 candidates forward in early discovery with Sanofi, brought in over $60 million from MAP milestones with Roche. We want to see a more regular cadence of all of that coming through. On that backdrop of data, we also have a lot going on with the platform. So we've built out Recursion historically was always focused on phenotypic discovery. On that biology discovery, we've built in transcriptomics.

We've added in using real-world patient data and really are focused on that multimodal integration to be able to do better biology discovery. Obviously, bringing in Exscientia changed the profile of being able to do chemical design. And now we can discover the targets as well as build the compounds, but then also quietly and you were one of the first to notice it, Gil, quietly in the background, we've also built up a pretty robust cleantech business that's allowing us to design better programs, understand the patients better. We've seen real data where we're improving enrollment by 30% to 60%. We're being able to identify sites in a better way and really understand the patient populations in a better way.

So the way that we view the business model of Recursion included 3 major components. So the internal pipeline, the pharma partnership and maybe the Recursion OS itself. Can you rank these components as it relates to your value creation? Who's your favorite child?

Well, it's funny. Our original mandate as a company was figure out how you can make a more risk diversified way of advancing biopharma. And a lot of our original investors, not only did they embrace technology, but they didn't like the classical model of biotech investing where you have this binary risk. And so that's why we have the diversified business model that we have and different ways of bringing it in. And even inside of that, if you look at our internal pipeline, some are more focused on the biology. Some are more focused on the chemistry or both. We didn't want anything to really be limiting us.

Now if you look forward, I think the partnership business obviously can be more of a steady rolling business than the biotech, which is very up and down based on data, the internal pipeline part of it. But they also have different value propositions. I mean if you look at the biotech industry as a whole, good proof-of-concept clinical data can be absolutely transformative. And it's much harder to do that with parts of the partnership business. And so we want both. We want to be able to build our platform and bring in cash inflows through the partnerships. That is scale and capabilities that would exist regardless of if we had a partnership business or not. So why not leverage our ability to do things faster and better and cheaper to build out a partnership business rather than just focus it on our internal pipeline.

So another key differentiator, at least in our view for Recursion has been the ability to industrialize data generation at the discovery stage. So are you seeing some of your competitors, pharma, anyone doing this kind of investment these days? Things have changed in the last year, I would say.

Yes, yes, absolutely. Well, I mean, this is funny. This is why I think if you look at the NVIDIAs and the Googles of the world, the reason they're such good operating partners for us is if the entire industry did business in the way that we do, it would absolutely be one of their largest segments. And so what we expect is that the rest of the industry will start to adopt it more and more because we are just showing that it can be done not only faster and cheaper, but really to get to results that hadn't been achieved before.

And so Lilly is the most notable party stepping into that and investing some of their good GLP money into building out the business. But I think you'll see a lot more. The reason you'll see a lot more is if you take all of the drugs that are currently in development or have been approved historically, you only cover about 10% of the genome. And that's not even talking about all of the diversity that each one of the genes can put out.

So really, 90% of the -- of biology has not been explored from a pharmaceutical sense. And that's because we have the same techniques. We have the same data sources. And so you're going to end up drilling down the same holes and getting to the same answers. If you want to start bridging out and actually exploring the other areas, you need new ways of doing it. You need new ways of creating and analyzing data. And so that's what we've been doing for over a decade. I think you're starting to see other people really understand that and start to adopt that. The good news is even if everyone switched over to our way of doing business today, you still have 90% of biology that's unexplored and so much to do around chemistry as well. So there's plenty of room for all of us to play in.

So related topic. So tech space has this attitude that everything is possible to do anything. But this assertion is reliant on large linguistic data sets that are offered by the Internet and human activity, which we both know is not that long term. So we don't think there's a true parallel data set in biology. Do you guys agree with that kind of line of thinking?

Yes. And it's really interesting because almost building off of that last point, the data has got to be at the core. I mean the reason you talk to ChatGPT and that can give you deep social advice is because there is a wealth of data on how humans interact socially. You just do not have that on the biological or chemical level. I mean chemistry is -- there's 10 to the 60th potential medicinal chemistries. We've barely even hit a drop in that and biology, I just made the point on. So you really need to be creating differentiated data to build models on. And the models themselves like how you do the models.

There is still a lot of capability and differentiation in the teams of people that can do it. But over time, they will commoditize more. And then you're going to be focused more and more on how do I get differentiated data to get to a different answer. And how do I integrate those things. One of the things that makes such a difference to us is we are not a point solution. We see this time after time where, okay, let's take what Recursion was originally known for, the phenotypic discovery platform. It's a great platform. It's a totally different way of looking at biology and using -- allowing the cell's biological process to be as complex as it wants to be in doing an image analysis of it.

However, even that data set is going to have a lot of fuzzy data in it as every data set does. And so what you need to be able to do is say, okay, I think I'm seeing a signal here. Do I see the same signal here in transcriptomics? Do I see the same signal here in real-world data? Can I create some experimental system that will support it as well? And it's at the integration of those different pieces that all of a sudden, you get much, much higher predictive probability. And so you have to have the data, you have to have different orthogonal data and then you have to be able to put it together. And I think that's really what makes the difference.

We've also seen a very stark difference in ability to learn and develop quickly if we're actually doing it on applied work. And so almost all of our spend, it's at least 2/3 of our spend is dedicated to applied programs in our pipeline or our partners' pipeline. And so then we have a real endpoint that we're designing and building and understanding. And so we can iterate on our models, we can iterate on our data and get better at it. I think it's really hard to do that in a vacuum because you always think your model is more predictive than it actually is when you put it to the test.

And maybe a last one on this topic. So we get asked a lot if drug development is a killer app for artificial intelligence, aren't more investors interested? Or what do you think the disconnect is?

Well, this has been fascinating for me. I mean, watching it evolve for a number of years now. What we ended up having to do was use that balance business model to be both the vendor and the customer in validating what we're doing because there was so little understanding of how it could impact different parts of drug discovery and how that would translate into development and such a massive established understanding of this is the right way to do business that we just had to be able to do it on our own. So part of the reason that we started our own internal pipeline was so that we could manage that process and validate, hey, we are getting differentiated chemistry or hey, we do have a novel biological insight and drive that forward and be able to talk about it. And so this has been a really interesting process.

I think someone who is more comfortable in the tech environment is more used to investing in sort of the potential of the market opportunity and trying to think about total addressable market and where do you go with it. I think in the biotech sphere, it's always been data based, right? Like what is the data say and what can I draw from that. And so to be able to transition from what was originally a much more innovation and tech-based shareholder register to really integrating more of a balance profile along with biotech investors has required us to create data and be able to talk about data. And that's what I mentioned in the first comment.

We're really just starting to turn over the cards that biotech investors really care about. And then I think, look, if those data points go well, it will be hard not to say that AI played a key role in it because it's everything that we do and the points of differentiation that it would be successful on, those are things that we try to solve with AI.

Do you think there's potential for like an aha moment like a drug that would go into market or some larger move as it relates to the space because one feedback that we do get is there's just no proof-of-concept that I would argue a little bit otherwise. I mean, there are marketed drug and placebo [indiscernible] comes to mind that are generated this way. But what do you think is that aha moment if such thing exists?

Yes. It's interesting. I think that initially, it will probably -- from a biotech investor viewpoint, I think until there's a couple that have really shown differentiation based on something that the rest of the industry couldn't do that people still just look at it as individual products. And to some extent, that's okay. Look, if we have successful products that are going through, the value attributed to that will help us continue driving the value of the business overall. But I think at some point, there is a aha -- I just -- I don't know if it's a wave craft. I think it's more of a distribution that happens over time rather than a single event like we saw with OpenAI. I'd love to see a ChatGPT moment. I just am more skeptical of that happening.

Moving from philosophy to practice. Collaborations, what can you tell us about expected milestone payments from your pharma partners if you have any visibility this year? And can we anticipate larger payments for foundational work kind of like what you've shown with your neural MAP...

Yes, really, really good question. So our partnerships, our primary partners are Sanofi and Roche. They both continue to go really well. With the MAP milestones coming in from Roche, we're at over $210 million coming in from Roche. I think our real focus there is let's translate the MAPs, which had never been done before and are really focused on entirely new targets in neuroscience, let's translate that into programs. And so that's a core focus for ourselves and Roche this year. And hopefully, we'll have some good events on that.

On Sanofi, we've -- that partnership is quite different because that came into it with, hey, here's a target we want to do. We, Sanofi as well as the rest of the industry have never been able to solve problem X. And that could be targeting the pocket or it could be, you know what, this is something that's a big multibillion-dollar industry, but there are known issues with the drugs in that industry. Can you solve it. And so that's a different type of problem. But the fact that we've already advanced 5 of those candidates through early discovery now -- so what that early discovery milestone means is we think we've solved the problem.

Initial testing, the preclinical and nonclinical testing has shown that, that problem appears to be solved or the problems. But now we have to make sure it's a good drug that will go into human testing. And so that's what the difference between the early discovery and the development candidate milestone is. Those development candidate milestones, which are the next ones for all 5 are larger milestones. The other really nice thing that I love is they enter operational obligations. And so that's all just profit that drops down to the bottom line. And then every milestone that we get after that is profit as well. And so we -- for each Sanofi program, each one can have $343 million of milestones, $193 million of that is pre-commercial. So this isn't massively back-end loaded And then we've got just beautiful royalties on it that would average in low double digits.

Now we haven't given explicit guidance for what we expect as far as upcoming milestones and timing. But I would say both of those partnerships continue to go well, and we're just going to keep on trying to both expand them as far as expanding more programs that are running through them, but also just executing on that pipeline and hitting milestones.

Very closely related question. Should we expect any clinical opt-ins in the near term? Are we reaching that level of fruition? This is something we specifically pay close attention to.

Yes. So from our internal pipeline, you mean?

No.

From the partners.

From internal...

Yes, yes, yes. We can answer both questions. We can answer both. So on the partner programs, so when they do development candidate, that is essentially saying it is -- they'll still need to do some IND prep work on it, but you wouldn't hit that milestone unless you're intending to take it into the clinic. And so that will have a milestone not only at the development candidate, but then also in the Phase I. And so we can start to see those come in. But it would be really surprising for us to hit a development candidate milestone and then not have them take it into the clinic.

That is actually really helpful. So I mean you touched on this, but maybe just to form it a little better. How should we view cash flows from partners over time?

Yes. So what we would like to do, there's always the question of expansion versus profitability. So the way that our partnerships are structured, we want to get paid upfront for our direct costs. So we don't want to be investing cash off of our balance sheet for our partner programs. And so we try and always run our partnerships at a neutral to mild profit until they start to hit the development candidate and beyond for Sanofi because then they can become very profitable. And then it's a question of, all right, how many more programs do you want to continue expanding? Because then obviously, you're not really reinvesting your profit, but you're delaying the margin associated with it is what effectively happens, but you're also expanding your pool of potential milestones and downstream payments. So that's what we're always thinking about.

We don't have a specific hurdle for either Sanofi or Roche. What we really focus more on, is this a good program? Like if we're successful technically in what we do, which we're usually pretty good at, is this going to be something that Sanofi and Roche are going to want to put their entire franchise behind and build up because that's when we're not only going to hit those development milestones, but that's where the royalties can kick in downstream and get really, really profitable.

So we do hear reports on pharma and large tech collapse every other week now. How can a small biotech like yourselves compete in shifting attitudes in pharma? You kind of touched on this, but recent reports on in silico come to mind.

Sure. Well, first of all, I think there are -- there is room for everyone's different business models that they want to run in it. I think in silico is a little bit different in that they probably go after a broader swath of potential partners than we do. We've been really trying to focus on what's the maximum value that we can get out of each individual partner, which actually generally means a little bit more exclusivity around each individual partner. Like I have no doubt that Sanofi and Roche pay us more because we don't have 20 other partners that are going along with it. So in that sense, I just think of them as different business models, not one is better and one is worse, but just different. We also do a lot more in-house sort of data generation and sort of proprietary biology, different aspects like that, that make it a little bit different in how we structure things.

And maybe last one, just a thinking point as it relates to additional strategies, just given the power of the platform, how do you guys view business development opportunities? You have like a relatively strong engine for vetting things.

Yes. No, absolutely. Well, the -- and it's interesting because we now have biology chemistry and sort of the clinical and real-world data side of it. We have a lot of different ways that we can contribute value. And obviously, FAP was an in-licensed program. We had -- the biology platform gave us a novel insight that no one had ever figured out before. And so that was before Recursion had a chemistry arm. And we went out and in-licensed a really nice compound for it, and that's worked out really well. And so there certainly could be other opportunities like that where we just have a differentiated insight on it. And we'll keep looking around because there's a lot of different ways that we can bring value.

This is a great segue. So let's talk a little bit about the clinical pipeline, starting with FAP. We're waiting on a regulatory update. But what do you think is the best case scenario here?

Yes. So the best case scenario is wherever we end up. I feel like we've got a lot of different ways that we can optimize around it. So if you think about it, there's a couple of different levers. One is what are the clinical outcomes. And we may be able to move some of the precedent that has been there before because no one's ever done the natural history studies that we did. Nobody has ever showed the depth of response that we did. No one's ever showed it as fast. No one's ever shown that they can maintain that response over time. So those are 4 new pieces of data coming in for the FDA that we can talk to them about and potentially that changes what a good clinical outcome could look like. And that's one thing to talk about.

But even if it doesn't, what we really want to do is understand which are the right patients to be treating, where is the highest unmet need. And we've been able to demonstrate that we can enroll these patients faster. It's very interesting actually because a lot of people don't think about it, but the FAP patients are going into their gastros like every -- it can be every couple of months, and they do this for decades. And so actually, we have a really good sense of where all the patients are and who their doctors are and how to get to them, right? And so being able to use those sort of levers to drive enrollment in clinical trial and trying to decide which of those patient categories we want to target in the clinical trial. I think those are all important questions.

There are also some around where in the treatment stage do you want to engage. But it's interesting. That's not a normal situation here because, let's say, a lot of people ask us about pre-colectomy or post-colectomy. Even if they're post-colectomy, those patients go on for decades where they're still going into their gastro every couple of months because the polyps actually just continue to expand. They go upper and lower then they'll go into the pouch, they'll go into the duodenum. And so you have to just continuously be going into the doctor. So there's not a segment of the population that doesn't have need. 100% of their polyps are precancerous. So I think it's a matter of us trying to decide where we want to target in the clinical trial and how we want to do it.

I mean some of the past experiences really did revolve around polyp counts as a key measure. But like you said, there could be ways to look at it in outcomes and does it postpone [indiscernible] colon, et cetera. So clearly, there's several potential outcomes here. When do you guys think you're going to be able to tell us?

Yes. Well, so we're in process and going through everything. I'll never get in front of the FDA, and they can -- they've been good partners with us on everything. We haven't seen disruption come out of any of the recent events. But you also never know until it's done. So we'll just -- we'll let the process continue to go and provide investors with updates as we have them to give.

And as it relates to clinical updates, when should we expect additional information?

Yes. So the 4 other clinical programs that are advancing CDK7, that's got combination data coming first half of next year. Really exciting there. The -- if you think about the CDK4/6 class, I mean, that's almost a $15 billion drug class right now. CDK7 should have broader applicability, but hitting that therapeutic index is the tight window. So that's what I'd say is a really high-risk, high-reward outcome, but that would be really exciting potential combination data first half of next year. MALT1 is also first half of next year. That's one where we've seen in other compounds a biological effect. So a good monotherapy biological effect, also combination therapy biological effect.

But all of the drugs that have been developed to date have a hyperbilirubinemia issue, which is a real problem because clinically, that would almost always be used in combination with drugs that cause liver tox. And so we think we've designed that out, and you'll actually know pretty early on, right? Like it won't take that many patients to say, are you having UGT1A1 issue or not. And so that data, we're excited about coming out. I skipped over RBM39. That's the nearest one. And I'm sure you will ask a question on it if I don't bring it up anyways.

So first half of this year, we'll be giving an update on safety and PK. It's probably an earlier update than people are used to, but this sort of underscores what we want to do is treat this like a business model, not like a pet science project. And so kill them fast if the data doesn't support it. This is a target that's never been drugged in this way, and it's a degrader. So you want to see good safety, good PK. You want to know that you're getting the selectivity and that you can keep going. And so we're going to take a look and hopefully move it on. But if it's not good, we're going to kill it quick as goes for all of our programs.

And then we would expect the PD data and anything else that we can present coming later on after that. And then LSD1, which is actually a really cool program as well that is just kicking off into the clinic. So that's probably late '27. But this is something where a lot of people have -- we know that LSD1 works but causes thrombocytopenia. And so everyone is basically taking their LSD1 out of cancer and put it into diseases where you get too high of platelet levels. If we can get that therapeutic index level right, then it could be a really nice drug as well.

And maybe another moment of philosophy just because I can't skip it. Is there any concern that once you start generating value from your clinical programs, for example, FAP, you're going to start being valued only on those programs. You'll become a rare disease company instead of a platform company.

What I'd love is just in balance. So I think that what's interesting is we actually appeal to a much broader investor base than a traditional biotech. And so historically, as I said earlier, we really appealed to a lot of the tech and innovation investors who saw what we can do and said, you know what, if you're actually able to scale that, then it could change the industry, right? Like that was a lot of the original background. At that point, it was too early for biotech investors because we didn't have enough data to be able to dig into. I think what we're seeing now is that data start to come through.

My guess would be the first drug that comes through, we're probably going to be valued more on that first drug, certainly by biotech investors. But if we start to have multiple drugs that are showing success, then it's going to be hard to draw a different conclusion than the AI is making a difference. And so that could be -- if you want a inside a single company, watershed moment, that could certainly be one of those.

An item that's near and dear to your heart. So control of spend is a recurring theme for Recursion investors. What's your strategy on those?

Yes. Well, I mean, hopefully, people noticed we took 35% out of the expense base last year while integrating the companies while advancing all of these programs. This is -- this was a huge priority for us. And the way that we do it is actually just by making much better decisions. And so what we did last year is we actually rebuilt all of our back-end systems and put in place an outcomes-based budgeting and business model. And what that allows us to do is break down and individually say for each program, this is what we're spending on it for every dollar that we have, this is where it's going. And to do that with very high fidelity.

And so what that means is we can look at it and what we did in our budget projections, for example, is we assumed that all of the clinical programs go forward. We probability weighted the partnerships and our own internal discovery efforts. But if we decide no, a program isn't going, we know every dollar and every operation that we need to shut down on the other side of that, right? And we can go in and be targeted and be fast. At the same time, we are also just going through every operation that we have. There's not a part of G&A or tech or clinical or anything else that we do that isn't constantly testing itself, can you do this better, faster, cheaper, right?

We obviously know a lot about AI. And so we've been an adopter of identifying. We've been an adopter of automation. And we actually utilize the knowledge and resources that we have across the company to help bring down that spend wherever we can. But it's literally just an ongoing, that's how we do business attitude across the company.

And as it relates to your cash runway and potential financing, how do you view this? How are you guys looking at things?

So the guidance that we gave for 2026 is less than $390 million in cash expenses. And that is not offset by any inflows that we have from our partnership business or otherwise. So that's just raw cash expense. I am very focused on bringing that number down, and I know Najat is as well to below $390 million and keeping to modify that. Obviously, part of that will depend on data and where things go. But if we look forward and we assume that all of our clinical programs are going ahead, we can see cash runway to early '28.

And so I think we would like that to be longer, and we control part of that with our spend, and we'll continue to manage that, as I was talking about. The other part is we're going to continue to figure out how we can best finance the business going ahead. I mean, in the end, we're a biopharma, and that's part of the industry. So no specific guidance beyond that.

All right. We're pretty much on time. So maybe if you have like a last point, something that you feel investors are missing or anything that we didn't discuss that you want to highlight?

Yes. I think the only point that I'd bring up is a lot of people are out of date on the story is what we find. They view us as a phenotypic discovery company or they know maybe 1 or 2 of the older programs. And we've just changed so much. I mean it's not just the management team. It's also parts of the pipeline, it's parts of the platform. It's the strategy as well, and it's continued to evolve really. And we've learned from the things that we did well and the things that we didn't do well, and we've really come out of it as a different and better company on the other side. So I hope people are paying attention.

Right, Ben, thank you.

Terrific. Thanks, Gil. Bye-bye.

3

Great. Thanks, everyone, for joining. My name is Scott Schoenhaus. I am the health care tech equity analyst here at KeyBanc. We're a pleasure to have Ben Taylor, CFO of Recursion for a fireside chat.

Ben, maybe it's helpful to sort of introduce yourself and the company for anyone that's new to your story. And you've gone through a lot of changes over the last 12 to 24 months, too. So it's worth highlighting your story to anyone that's new to it.

Yes, of course, happy to. So my own background, I was actually coming over from the Exscientia side of the merger. So I was the CFO and Chief Strategy Officer over at Exscientia and had been there for a little over 4 years when we brought the companies together 18 months ago. And before that, I had actually run the day-to-day operations at an oncology biotech. So got to see what it was like really digging in and setting up clinical trials, going out to the clinical trial sites, working with the FDA.

And it was actually really interesting because it was what led me into the AI-based drug discovery. What I realized is I and all of my colleagues around me and a lot of -- so I've been in banking before that. A lot of my clients had really been guessing because the data that you have to make good decisions is so sparse. And so we'll take data that's coming out of a few animal models or a few unrelated studies and try and make a guess as to what's going to be a successful trial or drug or patient population. And so when I was leaving that oncology biotech, I thought I wanted to do some sort of role that was more data-driven. And that really led me into this space.

The Exscientia side have been working on doing that in the chemistry space. So a lot of people focus on the biology component of biotech. But in fact, once you think you have an idea for a good target, a lot of trials actually failed because of the chemistry side of it, because either everybody knows about the potency or selectivity. So are you hitting your drug and your target and not hitting other ones. But there's so many other properties that go into it being absorbed and how is it metabolized and the toxicity profile of it. And that's something that AI really enabled us to do a much better job of doing a multiparameter optimization.

So combine that then with the legacy recursion system, which had been focused on how do I come up with new ideas in biotech because -- right now, about 3% of the genome has an approved drug associated with it. If you include all the drugs that are in development stage, you get a little over 10%. So that means about 90% of biology is effectively unexplored. And even the 10-ish percent that we are exploring, I mean, there's multiple proteins that come off of all of those genes and a lot of those drugs aren't ideal.

And so what we wanted to do was put those things together and say, how are we going to find new biological targets because you need to have new ways of creating and looking at data. And so that's where we've built our phenomics system. We do a lot of with transcriptomics. We do a lot of real patient data. And then how do you create a better drug for it.

And then the third part that we actually built internally as a part of the combined company was really what we call our ClinTech business. And this is using real-world data to analyze both patient selection, so trying to identify which patients will respond better to our therapies, but also optimizing the clinical trials. And this is where we've been able to increase our enrollment rates by 30% to 60% by just taking a different approach to how do you identify the clinical trial site? How do you work on identifying the right patients for the trial.

So we've been able to put those 3 pieces together, and I think that's been -- oftentimes, we're asked what's your differentiator? I think classically, it was always -- we have a lot of proprietary data, and we've got a lot of validation cases, both through our internal pipeline as well as our partnerships. But I think over the last 1.5 years to your point of bringing the companies together and building it out, we've actually created a really substantial differentiation by integrating all of those systems so that they work hand in hand. And we're not a single point solution.

That's really helpful for that context. Then I guess maybe talking through that, you've done some recent portfolio consolidations. Maybe talk about the strategy there? What inclines you to license your assets to other companies or shelving it altogether? And then how do you think about advancing certain other molecules like REC-617. So walk us through that strategy.

Sure, of course. So when we came together, we had a massive pipeline of about 10 clinical agents and far more than that on the preclinical level. And that was just -- it was too much for us as a single company to take forward. And we do have some large partnerships. We brought in over $500 million from our partners on that. But just focusing on our own internal pipeline, we said we need to be really focused and bring the highest impact programs forward. So what we did is ground up, look at the science, but also look at what's the clinical pipeline or clinical program for it. What's the commercial opportunity? How differentiated is that product going to be?

And then do we have aspects of our company and our platform that are really going to make a difference and improve our probability of success. And so based on any of those things, it could be the data, it could be the commercial opportunity. We tried to eliminate a number of those programs. And now we've got 5 in the clinic that we have taken forward. We've got about 15 programs in discovery between our own and our partner programs.

And the partner programs, we didn't do the same process too because with the partnerships, basically how that's structured, we get paid in advance for our direct costs associated for it and our initial milestones really cover any additional costs that would come into it until they get to development candidates or whatever the major milestone is like with Roche, we brought in $60 million for map delivery, right? So when we start to hit those milestones is when the profit starts to come through, but we don't have the same capital restraints that we do with our internal pipeline.

Maybe walk us through on the partnership side. I mean you recently achieved your fifth milestone from Sanofi in February. You have 15 best-in-class or first-in-class programs across oncology and immunology. How do you see partnerships further progressing from here?

Yes. We love our partnership business. So our main partners are Sanofi and Roche. We do also have smaller partnerships with Bayer and Merck KGaA. But as you brought up with Sanofi, we hit the initial milestones on 5 programs. What that really means is it was a program where the industry and Sanofi hadn't been able to solve a problem. Maybe it couldn't be drugged, maybe it had a serious side effect, maybe it wasn't discovered and it's a first-in-class.

And those 5 milestones all say, we think we got it. And now we have to go through some additional testing to make sure it would be a good drug to take into patients, at which point it hits the development candidate.

The nice thing about that development candidate milestone, it ends our operational obligations. So that's all profit. They're also large milestones, but that's all profit that drops to the bottom line. And as you mentioned, we're able to do that across 15 programs. And so we love to see that continue to grow and expand.

With Roche, where that started, it was originally to try and come up with new ideas in neuroscience. And so the $60 million in milestones that they paid us were actually doing first-of-the-kind maps of different neurological cells that allow us to think about the biology of neurological targets in a different way.

And so that was really exciting to see. We got an initial payment from them for $150 million upfront and then bringing in those $60 million of milestones for the success on the maps. What we'd like to do there is now start converting that over into programs, which then looks more like Sanofi. And so when you think about those programs, the average milestones are about $300 million over the course of the program.

Most of that, so for Sanofi, for example, $193 million of that is pre-commercial. So this isn't some big back-end [ biobucks ] loaded milestone pack and spread relatively evenly. And then average royalties are sort of high single, low double digits. So really nice economics behind all of that. So between the ability to do that in a capital-efficient way and really the high value of NPV that we capture, we love moving those forward.

I'll say whenever you're in a partnership, it's going to go along with the partnership as well. So there is a different risk profile to it. There's a different flow to it versus our internal pipeline, which is why we always wanted to keep the balance between the 2. If you look at our internal pipeline, like we can talk about the targets. We can do that with a partner pipeline. We can control the gating and the milestones associated with that. It's much more standardized in some of the partnership side. And so what we try and do is balance our risk profile and diversification across both of them.

Back to your internal pipeline. Maybe talk about the various readouts that you have upcoming over the next 12 to 24 months. What excites you guys? And what should we -- what should the investment community be on the lookout for?

Yes. Really, really exciting time for it. So we just had positive proof-of-concept readout on our most advanced program, which is in an orphan drug indication called FAP. And that we saw near 50% responses after only 3 months. It was durable even off drug, 50,000 patients in the U.S. and EU5 with no therapy that's approved for it. So this is a really exciting area for us to keep going ahead.

We are clarifying with the FDA the pathway for the pivotal trial there. And then we'll have additional data coming out on that first half of next year in addition to getting the pivotal trial started.

On our other programs, so we've got 4 other clinical programs. All of them have data between now and first half of next year. And most of those programs, so they all have -- well, the right way to put it is, they all have different aspects that had not been solved before that our platform has addressed. And in sort of the preclinical work and sort of the laboratory work, it looks like we've found ways to overcome those obstacles, whatever they be biological or chemical. And so this clinical data coming up will show did -- not only did the platform do what it was supposed to do and resolve those problems, but also does that make it a good potential drug. And so a lot of proof points coming up in that time period.

Sorry, just trying to unmute myself. I think an underappreciated part of your platform is the molecule patient side, the AI-powered clinical trial design and patient recruitment ability. Maybe walk us through those, Ben, and how you're seeing traction with pharma customers on that part of your platform? I think most people don't realize you have this as part of your platform.

Yes, you're totally right. So this is where it's funny because we have all of the infrastructure you need to be able to handle data in an AI type of way, be able to create and train models. So we have our own internal supercomputer and then the expertise to be able to do all sorts of different modeling systems. And so what we did is we basically in-licensed real-world data from a variety of different data providers. So this is going to be actual patient data, whether it's coming from clinical trials or claims or whatever other source. And we use more than 6 different data providers to get the information we need.

But then what we can do is create modeling systems to say 4 patients that have this gene mutation or 4 patients that have this disease and this comorbidity, who's more likely to respond? Or we can also look at it and say, where are they? Like one example we gave is for the FAP program, being able to target in and identify here is where there are a couple of dozen patients. It's not a normal clinical trial site.

And this is one that we should be at, right? And we were able to see enrollment rates increase by 30% to 60% on the trials that we're using this on because we're just being smarter about how we look at patient populations and we look at targeting rather than just relying on sort of the traditional methods that we've seen out of the CROs and some of our partners. So if you break it down, $0.70 out of the dollar in drug development and discovery is spent on clinical trials. Those clinical trials are all based on statistics.

Like we talk about biology, we talk about chemistry, we talk about a bunch of different things. In the end, it's what do I statistically need to do to prove that this drug works or doesn't. And so the better that you can identify the right patient population, the better that you can design your trial, the better the statistical plan will be, which means you will prove that out with fewer patients. And that means efficiency on the clinical trial time and cost associated with running that trial.

And so we apply it to everything that we can to understand the patient, understand the clinical environment, understand the sites. And what we're seeing is improvements that are having real financial and timing impacts to our clinical trials. Our -- you would think that everyone would be doing this, but the fact is that they're not. And so we've had a lot of interest out of our partners and trying to help them think through these questions as well.

Yes. I mean it totally makes sense. I guess taking a step back here, Ben, maybe talk about your balance sheet, capital on hand, what you have -- how long you have to cash burn for these clinical programs? Yes, walk us through all that, though.

Cool. So we ended the year with $754 million in cash, and that gives us a runway out into early 2028. This has been a really core focus point. So the companies independently were spending over $600 million a year when we brought them together in 2024. And we took just in the first 12 months, over $200 million out of that budget while still substantially expanding our capabilities.

So how we did that? We talked a little bit about sort of looking at the pipeline and thinking about what has the most impact. But honestly, we did that across the entire company. So right now, 2/3 of our budget is focused unapplied, so whether it's experimental or technology or anything else, this is applied costs going into pipeline and partnerships. So we don't have a massive part of our budget that is going towards sort of blue sky research or unapplied experimentation and tech development.

We actually believe that the best way to develop good technology is actually to have a program to develop it on because what you end up doing is you design something, you know what it's supposed to do. You also create an experimental validation for it and then you learn really fast.

So rather than sort of trying to make and train a model in a vacuum, we're able to actually like say, is this working now or not in the real world? And what we've seen is over and over again, we can get a model to work really efficiently and bring it up to speed. And that's why we've been able to take 90% of the experimental cost and work out of chemical development, and we've been able to cut the time lines from idea to to identifying the right compound in more than half, right? Like we're doing it in 17 months on average versus 42 months as the industry standard.

So we're seeing massive benefits while we still build pipeline or platform value by doing it in applied way. Long story to come back to, that is actually a really efficient way to run a budget as well. And we keep taking money out of the budget in the sense of we figure out ways to do it better, faster and cheaper. We're a tech company.

Even if we're doing biotech, we are a tech at heart. And so we need to be focused on that efficiency because almost always, efficiency actually means quality. It means you're getting to the right answer quickly and with the best possible system.

Last question for me, Ben. Where do you see Recursion in 5 years from now? Is it just a larger version of its current platform where you have more partnerships and more pipeline? You walk us through -- walk me through the vision of Recursion.

Yes. Look, I mean, I don't want to -- so the AI industry is so full of hype. We really try and not jump into that or trying not to have that be sort of the lead. I think we do have big ambitions, but what we want to do is really prove it step by step. Is our technology scalable? Absolutely. Should you scale it before validation? No. We are already partially scaled, right? We're running over 20 programs internally, which is massive for a company of our size.

If we have success across our clinical pipeline and partnerships, would you imagine that scaling further? Sure. Now what I will say is we want to manage that with the right budget and infrastructure. So if we -- if all 5 of our clinical programs went well, which, by the way, we never guide to that. We always assume that there will be some failures in there. But if they all went well, we would probably not try and take them all forward ourselves, right, because that is a really big infrastructure and clinical commitment to do. Could we commercialize products on our own? Absolutely. We have some people with expertise in that.

One of the factors that we looked at our pipeline with was, is this something internally that we can advance ourselves if we decide to. They had to have that in because we never wanted to be, "hey, we developed this great product and we can't take it forward because we need a partner." So we were really able to focus on that. But I'd say the focus of the management team, where we want to set expectations is look towards the near term. We've got a ton of value potential coming up. We're going to do it in a portfolio management strategy of killing early the things that don't work and investing behind the things that do.

Perfect. Well said. Well, thank you so much, Ben, for doing this fireside chat with me. Investors, if you have any follow-ups, please don't hesitate to reach out to me directly, and I can put you in touch with Ben. But thanks again, Ben.

Terrific. Thanks a lot, Scott. Have a great night.

Good morning and [Technical Difficulty] Healthcare Conference [Technical Difficulty] love to have you in Miami. [Technical Difficulty]. I'm Mani Foroohar, senior analyst, genetic medicines here at Leerink, and I am obviously excited to have -- I also have some technical difficulty, Najat Khan and Ben Taylor from Recursion. Welcome, guys. How are you doing?

Great place for having this.

Let's start the conversation with a little bit of a strategy question, Najat, and this is a further conversation on the last earnings call. How do you think about the continued rationalization and narrowing and tightening of the portfolio? And should we expect future updates to look a little bit like what we saw in the earnings call with fairly rapid iteration on what the pipeline looks like on a Q-over-Q basis?

No, great question. Again, great to be here. Look, I think from a strategic perspective, we're really focusing on 3 areas. One is doubling down on those proof points. Second is surgically investing in the platform where you can really lead to that proof point. And then third is carrying that ambition with discipline.

So just to answer that question, in terms of the pipeline, we have rapid go/no-go decisions across each of the programs. And there's a reason for that. We have a portfolio, which means that we want to make data-driven bets where we see the differentiation play out versus not. So I think that's number one.

Number two, you also see progress in terms of our partner portfolio. We just shared for the first time our joint portfolio with Sanofi. It's about 5 programs where we are designing the molecules for challenging targets in I&I and in oncology as well.

And then the third that you'll also expect to see in earnings, we track how is the portfolio doing, the velocity that we're starting to see in the portfolio and the platform specifically.

For instance, we are synthesizing about 90% less compounds to those that actually go to advanced candidate and to the clinic in about half the time versus industry standards. So we think it's really important for us to be objective and data-driven on the program, partner program and platform.

The last thing I'll say, and you heard us talk about the outcomes-based budget, this was a lot of work. We basically zeroed out the budget, and Ben can speak more to it. And we said, in order to meet the milestones and the catalysts we have, what is the fully loaded cost of each area.

So it really helps to have a more objective investor-like mindset to say, okay, if I don't see the data here, I know exactly how much capital allocation I can extract to either extend the runway and/or apply to another program where you have more conviction. That's the approach that we're taking.

Well, and Najat hit on a really important point there because what we're trying to do is use our technology to enable us to do portfolio management in a more classic way. So because we're able to advance programs and get better data earlier, we're able to make better decisions earlier and move clinical programs forward. And so you look at our -- currently, we've got about 7 programs internally that are advancing, and we've got the partnership programs.

You can expect us to be making not only quick decisions, but also a lot of transparency as we go through that program and advance it forward. And so it will look more like something that an investor would be used to, but all of these programs have something that our platform has contributed that we believe creates a better probability of success.

So let's talk a little bit about how that translates, how that flows through the financial statements from the pipeline. Obviously, you guys are engaged with the FDA actively with what is your current most advanced asset. Is it reasonable to assume that as individual assets move forward to pivotal studies, longer make as much sense, et cetera, that we're going to see tweaks around the margins on how you guys talk about OpEx on an annual -- on a quarterly basis?

Or is it more reasonable to expect tinkering with how you talk about OpEx assumptions on an annual basis? Like how frequently should we expect you guys to be tweaking our expectations and sort of giving us feedback to make sure we model right?

Yes. I think what we'll do is we've provided annual guidance. And when we're making more significant updates to the strategic plan, we'll give you more significant updates to the annual guidance as well. But otherwise, you can expect that we're working within that guidance to try and execute on all of the different programs that we put forward.

So for 2026, we said we expect our gross burn to be less than $390 million. That doesn't include any of the inflows from the partnerships. We do expect to get meaningful milestones coming in from those partners. We will treat that like I was talking about before, like a portfolio management area. So that we may take money out of some areas and put it into other areas. But if there's a major update, we'll give it to you.

I think you -- as you said, you're excluding partner inflows, can potentially be meaningful, but hard to predict with certainty. Let's talk a little bit about the opportunity and target landscape on the partnership BD side.

Obviously, an opportunity to inflect those numbers near term and get some leverage off of other people's infrastructure into the clinical development. There's been a lot of discussion on whether or not the tech-enabled drug discovery field is "crowded". I'm not sure what that means.

But how do you think about the dynamics in terms of the level of pricing power you can demand for your platform, how that evolves over time and the competitive dynamic with some other tech-enabled drug discovery companies, which are quite capital hungry and so clearly looking to partner compete for partnership volume with pharma as they attempt to fund themselves in what I think we can agree are fairly choppy macro markets?

Very choppy. Look, big picture, I'll start on that, and Ben, please feel free to chime in. I think the most important thing is not partnership announcements, but partnership value realization. And when I think about the landscape from that perspective, it's actually not that competitive. We don't really have a lot of companies that have actually shown that they can deliver and their pharma partners are actually paying them money talks, right?

So at Recursion, we just crossed over $0.5 billion in upfront and milestones. So just to give you an example of that, one is with Roche, which is much more focused on -- I like to call our platform like a trifecta, vertically integrated trifecta, biological AI for novel targets, chemical AI, chemistry AI for small molecule design.

We're not in biologics yet and then also clinical development AI.

So the Roche one was for novel targets in the biological space, and we just received $60 million in milestones back-to-back for 2 novel data sets that actually generate novel targets which we're working on. That's one.

And one thing that gets pretty misunderstood, I think that makes us different is not just the integrated vertical AI tech stack, but also the fact that we do all of the wet lab and dry lab. So to make these maps, we made 1 trillion iPSC-derived neuronal cells and the foundation models and the knockouts and now we're working with Roche, Genentech on the functional validation to say which insights are actually causal targets that we will collectively start programs on, very different.

The next, I'll say, as you mentioned, Sanofi. We just received our fifth milestone from them where we have 5 different targets we're working on around lead series. So this is Sanofi's TPP and us delivering on that, and we have more milestones coming up with the development candidate milestone, which is the trigger to actually onboard it into their pipeline.

So that's sort of what I tend to track, which is there's a lot of activity, but where is their impact. And having been on the pharma side for a long time prior to Recursion, that really matters to me a whole lot.

And I think this debate [Technical Difficulty] to your point, we learn from some of the best partners that we have. And these are in the areas of neuroscience, oncology and I&I, pretty big areas. And last but not the least, it is also another dual track to validate our platform. We test, learn and then we scale.

Yes. Just a couple of points to add on to that. So if you think about how we structure our partnerships, we get paid upfront or in early milestones to cover all of our direct costs. And so that building aspect that Najat was talking about is both a platform and an NPV value that we're building off of basically without having to use our own capital off the balance sheet to make it.

Now this is really important because about 2/3 of our spend is actually applied to our pipeline and partnership programs. That's including applied development on the tech side of the platform and our experimental work. So we really are gaining a lot from the scale that comes with that. But the financials are terrific, too.

I mean the Sanofi programs that Najat was talking about, per program, we can get $343 million in potential milestones, $193 million of that is pre-commercial, so not a big bio drugs deal. And then if you think about the royalties, average royalties will be in the low double digits. So really nice strong financial relationship there.

We've advanced those 5 programs through the first discovery milestone. The second milestone that's coming up is actually a development candidate. Now that's significant for a couple of reasons. One, it's a larger milestone. Two, it ends our operational obligations. So that's all profit that drops down.

And every milestone that comes in after that has no offsetting expenses to it. So it's all going to be profit that drops down to us. We have room for up to 15 programs on the Sanofi partnership. The Roche partnership, I almost hate to say it, technically, it can go up to 40. I doubt we'll get to 40, but we've got a lot of room in both of those to dive in.

Let's talk a little bit about the underlying infrastructure that you guys are building these partnerships as well as wholly owned assets on. Obviously, the company has done a little bit of rationalization around number of sites, et cetera. Is there still room to run there? Or do you guys feel like you've established like a baseline level of sort of maintenance OpEx, platform CapEx, et cetera? Or should we still expect to see narrowing the geographic footprint, et cetera?

Yes. I mean I'll start, but I know this is close to Ben's heart. Just as context, we reduced our pro forma expenses by 35% to less than $390 million, and we shared that at a conference earlier this year. Number one, I think OpEx, yes, from a G&A and so forth perspective, we've brought that down significantly.

We want to make sure every dollar is actually going to value creation. I mean this is operational excellence 101. I'm not seeing anything that exciting. We are going to continue watching that, number one. So expect that sort of discipline to continue, that has to be the case.

The second thing that we're also doing is, look, the platform is never static. In this era of AI where there's constant innovation happening, the way Recursion has stayed ahead is by actually investing in the frontier areas, but you also have to balance that with where does it really matter.

You don't want 1,000 flowers blooming. You want to make the ones that are the bottleneck in R&D. So we've taken a strategic look at that, and we're going to be very surgical in terms of where we invest in our platform. At the same time, we're also starting to see some of the velocity coming in and the efficiency from what the investments we've already made.

Like some of the stats around, look, back to the trifecta. In the biological part of the platform, once you've generated that data, which in biology, one of the biggest issues is the data sets don't exist. This is proprietary. 40 petabytes of proprietary data, that's a lot of data.

It becomes a search issue, right? You're just starting to search. You're not doing CapEx investment. You're actually just reusing that to understand, better understand relationships and validate them.

Second thing, on the chemistry AI platform, it's -- I want to underscore that again, making only 330 or so compounds to get to development candidate, advanced candidate like what goes into the clinic in 17 months versus 2,500 plus over 42 months, which is industry standards, and I'm being generous, I don't think it's one yet, but these are green shoots and it makes us -- people ask me, how do you do that? You simulate more, you make less. That's how you also get efficiencies.

So Mani, it goes both sides. We're going to invest, but we also expect to see efficiencies, what we've already built in and not just efficiencies, efficiencies that can lead to effectiveness in the clinic.

Yes. And I mean, if you think about it, we're a technology company. We should be getting more and more efficient every single time in every single new project we take on. And so as we look forward, we think about how can I make more tomorrow with less cost.

And so I think to some of the points, like our data has become more and more valuable because we're able to mine and build and actually look at orthogonal data sets and orthogonal testing systems to be able to do more in a simulated environment rather than running to experiment.

I think another important part is our CapEx spend. I mean if you look at legacy Recursion or Exscientia, both of them were in the tens of millions of dollars every single year on CapEx. Last year, as a combined company, we had $6.5 million.

We're only going to have a few million this year. And that's because we've made the investments. We know what is valuable, and we're driving that forward to push programs ahead in the pipeline.

And if I can just build on that, you also want to make sure your investment, especially CapEx investment is future-facing. So we invested a lot in this wet dry lab loop. Not everyone is talking about it, but that's an investment we made years ago.

The more important question is how do you use it effectively to make the right data sets that actually are useful to generate these novel targets. We're doing that internally, but then we also learned how to do that with the likes of Roche, Genentech.

So I think it's -- like sometimes I get asked the question, what's the differentiation of Recursion. And you can talk about many things. But ultimately, it's not just the data. Data is a huge moat. Like everybody, most models are based on public data. Having 40 petabytes of private proprietary data is important, but it's also fit-for-purpose high quality.

Models, yes, that's important. People, very important. Finding people who understand both tech and science, harder than it looks. But it's actually the integration of that vertical tech stack, right? The fact that you can go from biology to chemistry to clinical and back and forth and learn, that's where the effectiveness comes from. That's how you become and produce a more repeatable engine and not just a one-off.

Let's talk about the talent piece of that, now you brought it up. I know obviously, the debate about the struggle for talent in AI land is eternal. Sadly, no one's throwing $100 million at me. So for those who are listening, that would be okay. I think that's cooled down a little bit.

But as you mentioned that overlap of technical and analytical skill and understanding of science, especially with understanding of drug discovery, which is its own unique art and science, how should we think about the pool of talent and managing and investing in talent as an asset and where we are in terms of the competitiveness of recruiting for that piece of the technology stack, labor capital, however you want to think about it?

Yes. Previously, I had built an AI team at J&J, which was like to 300 people scale that across. One of the hardest things, Mani, was finding what I used to call and I still call bilingual talent, like proficient in both science and AI scientists who understand -- they don't need to code -- really do you need to code anymore, but they need to understand the interpretation of AI-generated data. That's really important.

Like if you think about scientists, statisticians and scientists talk similar languages, but it's still not the same case with AI and reverse, AI scientists who have the humility to understand drug discovery and development and how much of it you can't really engineer yet. Let's just be fair here, right?

So to answer your question, there are a few very rare people in the last decade that have been working in this space. A few of us happen to be by accident. I remember doing my PhD and I was doing both coding and computer science and organic chemistry, and I was consistently made fun of like take one lane.

Actually, innovation comes from the intersection of the 2. There's not a lot of people that exist like that. So I think what's more rational and pragmatic is you hire folks that have the openness, like a drug hunter, that has the openness to understand AI and doesn't sit there and say and get threatened by it, let's be fair.

And then AI scientists that actually want to learn about drug discovery and development and the time lines it takes is so much easier if you're optimizing ad revenues and so forth, right? It's like the reward cycle is so much faster. And I think the core of how you get those people to join you and to find you has to be the mission, has to be the purpose.

I think there's a lot of people, especially post-COVID, where I mean, let face it, it touched so many of our lives, patient and improving patient lives, everybody's got somebody that's a patient or they themselves are a patient. And third is, I think some of the innovation money, like things like the folds, I like to call them the AlphaFold, DragonFold, whichever fold, right?

The fact that you're actually starting to see these green shoots of, hey, I can simulate more and make less, this is something we talked about. I mean I get asked the question, oh, now you're good at efficiency. What about effectiveness? I'm like, thank you for noticing that because 6 months ago, nobody was saying there was even efficiency.

I think these green shoots are as important to talk to an investor or an analyst as it is actually to an employee, a potential employee because they're looking for who is that one company that's going to have the best shot of success because they have all of the pieces together, the scaffolding is right. And they also have the right purpose and mission.

So I will say, and you bring them in, but the journey just starts there, getting the teams to come together, not having silos, not having organizational constructs where they compete. It is not a versus. This is one big pharma. I mean, I can tell you this, like even though we had 300 data scientists at the prior company I was at, that was 2% of the R&D org.

You can do the math how big the R&D org was? 2%. How do you win? How do you have that impact? I mean, 2%, the cultural adoption and the inertia is one of the reasons people don't stay. So you got to recruit them, but you got to retain them by actually taking both disciplines and saying they're both equal. That's one of the hardest things and one of the things that I spend a lot of time on.

I think recruiting 300 AI data scientists who are characterized by humility sounds like an interesting task. Snark aside, I'm going to pivot over to the financial side of questions.

Let's -- how do you think about accessing capital? Obviously, the most nondilutive to ownership capital is partnership inflows. But how do you think about accessing different parts of the capital stack in current markets?

How do you think about use of the ATM in the future, equity, debt like instruments, partnership, et cetera? How do you think about those and rank them on the path between now and cash flow breakeven in the future?

Sure. Well, so obviously, we can't comment on future financing, but I'll give you a few parameters on how we're thinking about it, generally speaking. So partnerships, we always hope to be a good flow of nondilutive capital in. We obviously have our existing partnerships that I talked about earlier. We're always evaluating new BD and different opportunities.

I think part of that also depends on what the pipeline looks like going forward. We are going to be very disciplined. And I think you get to a very different set of options if you have 7 successful programs versus if you focus on the first one, the FAP program where we had proof of concept.

And so all of that needs to factor in. That's why we've got a very dynamic business model. We can actually pivot very quickly based on what the results are in from that pipeline and move behind that.

Now you brought up the ATM from last year. We did dip in opportunistically. It looks pretty good right now based on where everything has gone, and we've got a nice runway that actually goes out into early '28, which I think puts us into a good spot to hit a lot of the upcoming milestones.

ATMs are never meant to be a primary financing source, and we're really focused on how do you build out the shareholder register with lots of great investors. And I think we're also getting to the point where a lot of the biotech investors that traditionally wanted to see data first, now they've got data from the FAP program they can dig into, some early data from CDK7 and lots of interesting green shoots, as Najat would say.

So we've been getting a lot more attention from that side of the universe, not just the innovation and tech investors that were sort of our 5 years ago crowd that really drove us on.

I think one of the other topics I want to talk about here, you talked about the value of data as an asset. Talk to us a little bit about where you are in your relationship with Tempus, and opportunistically, how do you think about the role of other like transactions to acquire assets, access to data, expand your pool of other proprietary data sets that are necessarily available otherwise?

Like how should we think about that both in terms of that existing relationship and its financial implications, but also that is part of your strategy for accumulating your pool of data assets?

Yes. I mean it's a great question. Look, big picture data strategy, whether you're on the biology side, chemistry or the clinical development, there's no one provider that has it all. It's a little bit of patchwork, smart patchwork in order to have partnerships with the right people that really stitch together the data set one needs for the programs that they focus on.

So just as an example, like if you're going into a program ovarian cancer or non-small cell or prostate, there is a variety of different providers that are complementary. So we're going to be opportunistic always in terms of which data set. So Tempus is one, but we also have partnerships with at least 7, 8 other providers that don't get talked about but were constantly doing that.

The other thing is the space of data providers is also evolving, right? It's not static. The amount of multimodal integration that we're starting to see, because look, we do a lot of the phenomics, transcriptomics, a lot of the omics data generation, coupling that with genetics from others and also that connected to clinical data.

And then they're also generating transcriptomics really helps us with that signal to noise. It's incredibly opportunistic. We're going to stay flexible, and we're going to stay smart.

Another thing I'll say is, look, there's always a question as to how much money you spend with each partner, breadth versus depth. As we have more programs coming in, we're going to do not just breadth, we're also going to do depth. So that means we have to be smart about how to allocate our dollars. So everybody should be on the their tippy toes. We want the best data.

So when you think about depth of access to a data, to a partner that's providing a data asset, is that something that we should think about in terms of the length of the relationship as they continue to accumulate the data? Or is that a function of just transaction size? Like what does that mean?

Yes. When I say depth in terms of the data set, like I'll give you an example. You can either partner with somebody and say, I'm in oncology or you can say, I'm in oncology, in ovarian cancer patients, platinum-resistant, how many patients do you actually have? This is really important to do diligence with data partners the right way.

The top of the funnel always looks good, 10 million patients. You would start to apply the inclusion exclusion, you end up with 20, right? And where a lot of the value comes from is actually that 20. The top of the funnel is good for sort of broader causal AI networks. But then we're applying it to a specific patient population, you want to get very specific where the data sets have high quality, high depth.

That's what I mean by that. Not the length of the partnership per se, but the richness of the data because there's a lot of data missing that people are still working through. And that's where, I mean, for me, at least I judge the quality of the data and what they're doing to actually close out the missingness.

Let's talk a little bit about that dynamic. We've talked about acquiring data assets. Partnerships are in a way monetizing your own asset. Let's talk about moat. I think there's a lot of discussion.

I'm sure it's going to come on my panel later, that, well, to what extent is there an investment in building internal infrastructure, tech-enabled drug discovery tools at your pharma counterparts, either your partners or those who are not your partners, et cetera.

Other than the cultural dynamic you mentioned, which is obvious, how do you think about internal efforts at large pharma as competitors or as complements to what you guys offer as a partner?

Yes, it's a great question. Look, I will say I expect the world to be where pharma partners are going to continue to build, and they should. That actually shows conviction in the fact that leveraging AI, leveraging larger data sets is going to make a difference, number one.

Number two, and pharma has always done that. Like think about any modality, ADCs, siRNA, any other platform, they build their own, they also partner, right? So I think it's going to be a little bit of both. Some probably will be competitive. Some probably will be complementary.

But again, at the end of the day, the value proposition comes from the integration of the different layers. And in large companies that sit in different organizations. I mean when I was at J&J, we were one of the few companies that had it all together under one organization. So organizational construct matters, cultural adoption matters.

But then the third thing I'll also say is the speed with which you can also innovate. The reason why you end up partnering with a specific company, not just AI, but any other platform is the depth that they have in that area, right? I mean the 40 petabytes of proprietary data we have, that wasn't done in 6 months. It took time.

The design of actually building a wet and dry lab is not nontrivial. In some ways, like Recursion has one of the most long-standing historical platforms possible. You can look at it in many different ways. The one thing I think about, we have made a lot of mistakes/learning across the board. And you really want somebody who has really gone through those reps, who has a lot of reps.

And that is also, I think, important and complementary for any organization. So I think it's always going to be a bit of both. And the proof is going to be in the pudding in terms of do you actually have better data, whether it's in the clinic, discovery, both effectiveness and also the efficiency and velocity.

One other thing I want to add on. I love the question of, is there enough space in drug discovery for everyone to be competing. I mean about 3% of the genome has an approved drug, around 10% has something in development. That doesn't even factor in if you think about the diversity of proteins that come off of that genome.

And so we are just scratching the surface. The reason we have a 95% failure rate in the industry is because we don't have enough data, we don't have enough ability to make predictive models and really search and understand biology and understand chemistry. And so we're actually just starting to step into the much, much, much bigger part of the industry that has been primarily untouched.

That's also part of the problem with the public data sets that those public data sets, not only do they have a lot of different ways of annotating that data that makes it really hard to use in machine learning, but also it's focused on that 3%. And so you're going to keep going down that same hole unless you come up with some new ways to explore the rest of the space.

Yes. Most of the models that exist because all of the public models, anything that's open source, we can bring it into our platform, leverage our data, refine it, use it in a matter of a week. That's, once you have that infrastructure, you can do it rapidly.

But most of the models that we have found is they don't work well in out-of domain areas. Might have worked really well with kinases, but you try to go into other target classes, it doesn't work as well. So somebody has got to do the work to actually generate that data and be hyper focused on it.

And once you have it, I mean, you think about some of the other AI companies that have grown rapidly, OpenAI, Anthropic, et cetera, is based on the corpus of data from the Internet. We don't have a corpus of data in biology, chemistry or even in clinical development.

Somebody's got to build that road before you actually build a good car to drive it. So you have to do both at the same time. And that's why the portfolio and the platform strategy, but you have to be very smart about capital allocation.

Awesome. And with that, we're now over time and I'm being given to get off the stage signal. Thank you so much. I look forward to this company.

Thank you.

Thank you.

All right. Good morning, again, for I see some familiar face in the last one, but thanks for sticking around. It's good to see everybody, and welcome back to TD Cowen's 46th Annual Healthcare Conference.

It is my pleasure to be joined today by Recursion's CFO, Ben Taylor. We've got a lot of great stuff to unpack, discuss, sift through. So we have our own questions, but by any means -- by all means, excuse me, if you've got any questions, just kind of flag me or feel free to send me an e-mail at [email protected], I'll be checking my phone.

But Ben, maybe let's just kind of start by taking a quick second to update us on really the Recursion platform of March 2026. Where do you kind of see the most important areas of evolution recently? And what this kind of tells the rest of us about what the platform is actually able to do today that maybe it hasn't been in years past?

Sure, of course. Well, and I think there are a couple of different pieces that come together. We have obviously changed the platform and the business model quite a bit over the last 18 months. And what I think that really reflects is going from a single point solution focused company or companies to really trying to create an integrated platform, as well as an integrated business model.

And so what I mean by that, maybe starting with the platform first, Recursion, a couple of years ago was really focused on inductive phenotypic screening to uncover novel biological targets. And really, how do you translate that novel discovery into a clinical program and in the future in medicine. What you've seen now in our platform is we've taken that novel biology, and we've added on transcriptomics to it. We've added on proteomics to it. We do a lot of reverse genomics using real-world evidence.

But then we've also added in through where I came from on the Exscientia side, a novel design platform that's really looking at how do I take that nearly infinite chemical space and try and come up with a good chemistry to try and test that biological thesis and try and solve problems where the industry hasn't solved it before.

And then we added on to that as well, and this was really thanks to our new CEO, Najat Khan, who had joined from J&J after overseeing their portfolio strategy as well as developing a lot of their AI technologies. how do you look at the clinical side of the world, the real-world evidence, how do you design better trials, but also using that to bring back into the biology and chemistry and thinking through what's the right patient and how should I design this for the right situation.

And so what you've seen is we've come together. And that's become a single workflow across the company where we're saying, what can I use to better understand the biology, the patient, the chemistry, the clinical trial I need to run, so that I'm incorporating better predictive modeling so that I hopefully have a lower failure rate in the clinic. And so that's been how the technology platform has come together.

At that same time, we've also been really trying to do the same with our business model. And so this is something that's very different in what we're doing today, if you look at our partnership business, we've crossed the $500 million mark in inflows coming to the partnership into the company. And so that's really making a dramatic difference in how we're able to build and invest in the platform as well as a great contributor to the NPV, but then also on our internal pipeline.

We've got five clinical programs, two and two preclinical that we own the rights to. And that's not in a single indication. It's not in a single technology. It's not a single scientific principle. And so what we try and do is say, some of that will be first in disease or first-in-class. Some of it will be best-in-class.

We really want to take a risk diversified approach to how we're advancing that. This is coming back to our original mandate as a company. Was actually to take some of the binary risk out of biotech and make it into more of a business model. And so that's really starting to come together in a new way.

Okay. Great. So I know you mentioned Najat. So we've spoken with her a number of times. It's great to kind of see the company, I would say, 2.0 under her leadership. So kind of tied to the evolution of the platform, I guess, what are the most important kind of things to flag from a strategy leadership standpoint in the Najat era now? And maybe kind of tied into that, how we should think about even expenses, right, like cash burn kind of tied into all of this. As we look at your growing and expanding pipeline, how is Najat kind of approaching what Recursion is able to do with that pipeline through the lens of kind of burn moving forward?

Well, I'll start answering in reverse there. So everything that I just talked about, we did last year for about $400 million in gross spend, so not including any of the inflows from the partnership. The comparison in pro forma 2024, that was $606 million. So we took $200 million out of the expense base, while actually really expanding our capabilities and building out our clinical pipeline.

And I think those two, the previous answer in this one is a great analogy to what Najat's focus on the businesses. So how do we get down to those things that matter most and be disciplined about making decisions around them? Not to invest in areas that have a low potential return, not to invest in things that don't have a clear clinical or commercial pathway. Like we have so many things that we could do. How do we focus on the things that we will do best and that we can have confidence? This is a bet that we should be making. And so that's been something that she's really come in.

Obviously, she's got fantastic operational and scientific experience. And so really bringing the rigor into understanding how we're making our scientific decisions, what we're investing in on the technology side. She's both a coder and a chemist by background, and so that's been really a fact-based, data-driven, disciplined management style.

And you mentioned some of the partnerships now, too. And I think this is -- this is something we are kind of constantly trying to unpack on our side, both as analysts trying to actually build out some of these models and understand when and where the value is coming from. But also on the investor side, too, trying to get at really what is differentiated about like what Recursion is able to do versus some of the other AI players in the space when you're talking to some of these pharma partnerships. So maybe we can use the Roche PhenoMaps conversation as an example here.

So I know you recently announced a kind of another milestone within that. So can you maybe just walk us through -- first of all, kind of just the economics of that particular partnerships and that particular partnership and how representative it is of Recursion's ability to kind of work with pharma. But then also, again, maybe extrapolate a little bit to what you're hearing in your conversations with pharma, whether that's Roche or others and how we should think about the capacity to partner with AI companies?

Yes, absolutely. And we get this question a lot because obviously, you see things like Lilly making very large investments into the AI space and as well as a number of other companies. I mean, Roche and Sanofi, who are our two largest partners have obviously made a lot of investments themselves. But they continue to invest behind our partnerships.

So with Roche, in particular, to date, we've brought in $210 million from that partnership largely around the construction of Novel PhenoMaps towards neuroscience. So this is looking at cellular systems, transcriptomics, whatever other data sources we can create a map from to really understand what are some novel targets in neuroscience. This has been a corner of the industry that has had very, very few new targets over the last several decades.

And so what we're finding is there are a lot of new potential targets that are coming up when you take data creation differently and then look at it differently. And so that's been the collaboration with Roche where we've had the $60 million in milestone payments coming in from those two maps. Right now, we're really looking at how do we convert some of those ideas into design programs. So actual work on creating the medicines from those ideas.

And that has continued to move ahead very, very well from both sides. I think what's interesting, how we structure our partnerships is really, we try and get paid in advance for the direct costs associated with our development of the technology and or programs and then maybe have an early milestone like with our Sanofi collaboration. We have now hit five discovery milestones. And what we're really doing there is each one of those programs is an example of there was something along that target that had not been solved by industry before.

And so this is a collaboration with Sanofi to say, can we do something better? Can we get to a target product profile that no one's ever reached before? The initial discovery milestone looks at it and says, we believe we got there. Now we have to verify that it's a drug that's going to move ahead. That's the next milestone each of them. What those collaborations, whether it's Roche or Sanofi lead to is once we've completed our aspect which hopefully either comes up with a new target and/or new molecule, then they would bring it in.

So in Sanofi, those five programs, the next milestone is development candidate. That is effectively profit coming towards us because it ends our operational obligations and Sanofi carries forward and does the clinical development moving ahead. With Roche, what we're doing is converting those ideas into the design programs, which would then follow a similar trajectory.

Now getting back to your original question of what does this mean for the industry? How is this differentiated work with us because we're not only creating great models. There are actually a lot of companies out there who can create great models. What you need to be able to do, you need to have a data set that actually powers it. So we've got tens of petabytes, over 50 petabytes with our own proprietary data that has been created by us in a way that is usable for machine learning and other AI techniques. That's so important because most of the public data out there is -- it's effectively dirty data.

It's very hard to use. You need to do a lot of cleaning before you could use it for machine learning. And it's also going to point you towards those same things that have already been drilled into. One important thing a lot of people forget about, right now, about 3% of the genome has an approved drug for it. If you look at everything that's in development, you're going to be closer to about 10%. That's a tiny part of biology that we're playing around with.

And almost all of the data that's in existence is focused on that part where it's already been something developed. And so if we want to start to address that other 90% that's still basically a blank sheet, we need different techniques to go into it. And so this is where our data creation, our models are being able to do this on an integrated basis makes such a difference because any one model system that you create is going to be messy. There's going to be a lot of noise.

And so one of the things that we found in that integration that we talked about earlier, is being able to take multiple different modeling systems and compare them to each other, you can start to take some of that noise out. As the only other way to test it is experimental. And so what you actually want to do is create a great new novel modeling system, have other modeling systems that could actually help you say, is this noise or is this real signal. And then as late as you can get to that experimental validation.

Can I ask just maybe one last one here before we dive into the pipeline itself. How this answer shifts if instead of me asking about pharma, I'm asking about the big tech players. Obviously, NVIDIA made some waves in the past couple of weeks here. How should we kind of think about their impact on AI drug discovery players such as Recursion, but like where is their interest in all of this?

And you're mentioning a lot of kind of recognition that a lot of the data sitting out there is dirty or maybe not as useful as maybe a lot of us on the outside would think just based on the sheer quantity that's available. So is a lot of this kind of identifying what's good and what's not, building out some of the model systems to be able to fill in some of those gaps? Like how does the answer shift when we're talking about NVIDIA or an AMD versus -- and their health care initiatives versus like a Sanofi or Lilly?

Yes. So actually, everything that they're doing, whether it's NVIDIA or Anthropic or different groups, in the end, it's very helpful to us because a lot of what they're designing for is actually either workflows or greater compute capability. And so the -- if you break it down, the underlying modeling systems, still need independent creation and validation. But that workflow that goes along the top, and this is all of the agents and the different aspects that are being built are basically as good as the foundations that you put them on top of, but you need that foundation.

And so we look at some of the work that's going on there and fully embrace it and bring it in as quickly as we can because it just makes our job easier, but you still have to create those underlying systems. I think NVIDIA is a great one. I'm sure everyone saw some news. It was sad to see them go as an investor, but it was actually completely separate from our corporate partnerships. So we have an ongoing corporate partnership with NVIDIA that's doing great. And they shifted their portfolio strategy, and they're obviously focused on quite large investments right now.

But we continue to be a really good partner. They -- we have one of the fastest -- I think it's technically still the fastest, though Lilly will take it, supercomputer and biopharma. That was all an NVIDIA collaboration, and we continue to build on top of that. That compute is essential to what we're able to do. I mean the fact that we have it in-house means not only can we do it faster and we have a lot of cost savings, but every incremental improvement that they can make on to that, all of a sudden, it amplifies our ability to use our modeling systems to search the data to really power how this all works together.

All right. So I want to now dive into the pipeline itself because there's a lot going on here. So maybe let's start with 4881 in FAP, right? So you recently put out 25-week follow-up data as is the MEK1/2 inhibitor. I guess, first, really quickly, how did that initial update last year kind of measure up to what you were hoping to see also in the context of the current standard of care and really now lay out for us what are the next steps for this program over the coming months?

Yes. I think the data itself was about as good as we had hoped. So we saw dramatic reductions in polyp count. This is a disease where because of a genetic abnormality in the patient that they were born with, they will continue to have cancerous polyps grow in hundreds or even thousands of numbers throughout their intestinal tract.

And so what you're trying to do is control that polyp growth because as the physicians look at it, if the polyps look like they are becoming cancerous, that's when they're going to do resections. And what our goal is, is to minimize that impact to patients, not only try and control the polyp count and burden, but also limit the need for resections and other surgeries for these patients. And so we saw a more dramatic reduction in polyp count than it's ever been seen in the area.

Also, we had the patients and we took them off drug for a 3-month period. And what we saw is those responses were stable over that period, which is a really massive statement because this is a drug that, in an ideal situation, you'd be giving to the patient chronically for their life. And so you want to see not only that mechanistic action, but also the durability with it. So what we are looking at now is we're engaging with the FDA to figure out what the right next steps are.

I think the big question mark coming out of the data is just what does the regulatory path look like. And so because this is a potential first in disease drug, we're engaging with the FDA to find out what the right trial design will look like going forward. We also did a large natural history study getting back to some of the things on how we can make a difference. We did two, actually, one that was following patients to see. Do these polyps naturally go away? They don't.

The other question, we were actually able to create a foundation model within a couple of days that looked at about 250,000 clinical records and said, how are these patients actually being treated? What are the doctors writing in the notes about these patients? What is the actual standard of care? Because nobody knew, right? This is not something where you've actually got a stable standard of care. And so that was incredibly important in us being able to look at what are these patients likely to see in the clinic? What's their experience going to be? What other drugs are they going to have? All of those different aspects.

Okay. So presumably now the update from these conversations with the FDA, first half would give you a sense of what -- when you could potentially start a pivotal study and what that study would actually look like. Okay. So then I guess kind of tied to this because I know you all have talked about the evolution of the cleantech platform within recursion platform itself.

Is it fair to assume that a lot of those investments when you're looking at starting a pivotal study for FAP, does that mean that you could find those patients faster, enroll them faster and maybe get to an actual pivotal data that much sooner? I guess what's that kind of look like through the lens of recursion today?

Well, we definitely hope so. And we started to put out data. So cleantech is a relatively new platform. We basically created it over the last year for us. But we've already started to see results. So what we've seen is enrollment going about 30% to 50% faster than the baseline that we had previously, which is really exciting, and that's basically, and we gave an example of this in our earnings. We're able to dive in and understand every one of the clinical sites, the patient populations and basically design our sites and design our outreach to be able to reach those ones that have the most potential to get us better patient populations.

This is something you would think you would do along with the CROs, but I can tell you now having been involved in it many times over, it doesn't usually happen. And the way that we're doing it is actually quite different. So that's one of the ways that we're driving it. The other is being able to say who are the actual patients who I think are going to respond to best out of it. And so that's really looking in and that can actually lead us down different indication pathways.

We gave this example with our CDK7 molecule, where if you talk to any of the KOLs, if you look at any of the papers, everything drives you towards estrogen receptor positive, HER2-negative breast cancer, generally in the CDK4/6 refractory area. We said, okay, fine. The preclinical data supports that, that should work. But what else could we do? And so we looked at different patient populations. We looked at different CDK7 mutational profiles for real-world patients and how that affected them in different in different indications. And what we found is, actually, it made a pronounced difference in ovarian cancer.

Now this is really interesting for a couple of reasons. One, there are a number of ovarian cancer patients who are also estrogen receptor positive. So maybe there's a potential link to why that could be more successful in breast cancer. But also, this is from our own anecdotal data, and I consider all Phase I trials anecdotal. From our own anecdotal data, we saw a partial durable response in a fourth-line metastatic ovarian cancer patient on monotherapy in our own trial.

We didn't expect to see any responses because this is effectively a cytostatic mechanism rather than cytotoxic. So it's a cell cycle disruptor and transcription inhibitor. So that was really exciting. We were seeing it from multiple different angles. There is a logical mechanistic angle. There's a real world inductive evidence that it should have an impact, and we saw something coming across anecdotally in our clinical trial. And so we pivoted the clinical program to ovarian cancer. There are other indications that we could go after and expand into, but we felt like that was a good place to start.

And I want to ask a little bit more about the CDK7 drug. But just to follow up, so the 30% to 50% faster enrollment that you mentioned, is that in one specific drug, one specific program? Or is that kind of just across the board? And I guess I'm curious that 30% to 50% faster than what's the comparator here, like what you had been seeing in your own trials before you did this investment or what you've seen kind of broadly in each of these spaces?

Yes. So multiple different trials, both is the answer on the -- what's the comparator. So we looked at -- basically, when you work with a CRO, they will give you an estimate of what they believe is likely to happen. We also do our own work and look at historical clinical trials and look at the enrollment rates from historical clinical trials. But then we looked at the trials that we were running before and after we implemented the change. And it's the exact same signal across all of them. But that range, the 30 to 50 is probably three or four trials.

Okay. Got it. I love those kinds of metrics.

Yes, we'll have more coming.

Great. Okay. So then on CDK7, when we're talking about ovarian cancer population, I guess, how does the data you've seen so far now and you talked a little bit about moving into this kind of set you up for what is -- maybe start with what is the time line now for this program and when we can kind of expect next data readouts here? And how has the data you've seen so far kind of informed where it makes the most sense to go forward?

Sure. So CDK7, what we've done there, and I mentioned about the ovarian cancer. This is a drug that will almost always be used in combination. And that's true of most any oncology indication now. And so we ran the monotherapy and put out the data most recently in December. And what we wanted to do is then start the combination dose escalation and then expansion, the Phase I/II basically in combination. And so that is ongoing now.

And the initial focus is on ovarian cancer. It is designed as a basket trial that could be expanded into other indications as well. Our guidance for data is first half of '27 for initial data coming out of that. And so that will be -- it's one of those high-risk, high reward, but would be really exciting to see the impact coming in. CDK4/6 is a $9 billion-plus commercial market right now. And we think that the opportunity for CDK7 is actually a lot broader.

And I guess this kind of gets at something I was asking earlier, but when we're talking about -- so you have the CDK7 update coming first half of next year. How are -- beyond FAP and CDK7, the rest of the pipeline now, I know Najat has spoken a little bit more to framing some of these updates as kind of go/no-go decisions, right?

And I think for a lot of us that it feels like, okay, there's clearly a bar internally you want to see. And if it's not met, then you kind of move on. But how is kind of that reframing of some of the earlier-stage pipeline impacting like some of the strategic burn for asset, for example? And like as we kind of look at the next 18-ish months for recursion, like what are kind of the most important pivot points within the development of the pipeline through that line

Yes. Well, and this gets back to the very beginning, everything we have has a near-term go/no-go, and that will be data-driven. There is no drug that we feel like, hey, we're comfortable with really gray areas here. And that's an important distinction because there is no single drug or a partnership that makes up the value of our company. And so we look at this as a portfolio management approach.

Every drug that we have has a reason that comes out of our platform that we think it has a better probability of success. And so that's why it's there. But if we don't see the data supporting what we want, it will be terminated tomorrow. And so even though we brought 35% out of our expense base over the last year, it's not enough. It's still a high burn.

Now if all seven of the drugs that we've talked about in our pipeline are successful, I think that drives you to not, oh, how do you figure out how to manage all of that burn, you probably just need to start to do out-licensing or different activities like that.

And so what we have right now is a business model with a lot of option value in it. We need to finish running the current experiments that are ongoing. But if any point, we see negative data, those drugs are gone. And that's something that we will hold to. We hope investors watch and hold us to it as well.

Okay. So we've got the update for FAP first half of next -- this year, excuse me, we've got the CDK7 data first half of next year. We'll also see some updates on 7735, if I'm mistaken and 102 in the second half of this year, if that's right. So maybe talk us through what the go/no-go for those programs look like and where you kind of see the bar like we need to see success on this metric, this metric for each of those. Otherwise, we need to kind of reinvest in the ones that we have more confidence.

Yes. So all of them need to realistically exceed traditional benchmarks for your safety, your PK, your PD, whatever it is that we're looking at, at that particular stage. However, I'd say if you look at like a MALT1 that was specifically designed to take UGT1A1 inhibition, which causes hyperbilirubinemia out because you're going to use that drug with a BCL-2 or a BTK inhibitor, which are going to have liver tox. So you have to take that out. All of the drugs that have been developed for it have that.

And so not only does it have to look like a really good drug, you have to get rid of that. LSD1, similar in the sense of -- we've seen mechanistic effect from it, but everything in that class causes a lot of thrombocytopenia. And so if we see a lot of thrombocytopenia, it's gone, right?

And that is an on-target effect, to be clear. But that's what we're trying to use our system to better develop is how can you actually minimize the on-target side effects by designing a better drug. And so without going through every program, each one has something specific in addition to the fact that it has to be a good-looking drug.

Okay. And I know we're basically out of time here, but -- so maybe last question. We covered a lot of the pipeline. We've covered the burn, we covered the renewed strategy here. So where are you kind of now -- as you look at the recursion to 2026, where are you feeling like the disconnect -- biggest source of disconnect is between where a lot of investors, a lot of analysts are trying to understand the value for Recursion and where ultimately you see a lot of the day-to-day value being driven?

Yes. So I think there's a couple of parts. I think a lot of people still think of us as sort of a single point solution company. We still get people asking us about how are the Phenomics program is doing, which, of course, we are focused on every part of it, but it's a corner of the company versus where the company is now.

And so I think that's lost. I think the value of the partnerships. I mean if we were a smaller biotech, each one of those programs were probably something that you'd see on our pipeline with a logo of Sanofi next to it or whatever, right?

And so they're really fantastic programs. We can't talk as much about them because they're partners, but they're fantastic programs, and we've got great economics. So I think understanding the risk diversification that we have across our pipeline and platform and understanding our partnerships.

I think with that, we are at time. So Ben, always great to see you. Thank you, everyone, for listening in. Got a lot more to come. Great.

Thank you.

Good morning, everyone, and thank you so much for joining us. I want to start by briefly framing where Recursion is today and its journey and evolution. Over the past decade, recursion has built something truly special. A differentiated platform, pioneering the integration of large-scale biological data generation, machine learning and compute to better understand the complexity of biology.

We have also deliberately strengthened the foundation in chemistry, NAI through the acquisitions of Exscientia, Balan and See, creating a truly powerful foundation.

Today, we're at an important inflection point. We're harnessing everything that we've built to date to do 2 things. Number one, translating insights into evidence. Evidence that this platform, the use of AI end-to-end can generate medicines that matter. And we're doing this both across our wholly owned portfolio and through our partnerships with strong momentum across both fronts, and I'm excited to share some of the updates today.

In parallel, we're also continuing to advance the platform itself. Today, we have what I like to call a trifactor, that's required to make impactful medicines. AI-driven biology, AI-enabled chemistry and AI applied to clinical development.

We continue to invest to ensure we're defining the standard for how AI is applied across the full life cycle of R&D. And look, as we look across the sector, we are encouraged by the broader momentum in the field, new models,new players, new partnerships being announced, but the industry is clearly entering a new phase where value is being defined not only by the models you build and the collaborations that are announced but by actually translating those. This is the hard work into capabilities, into real application and measurable impact. The important question now is not only what you build, but what you can unlock. And that's the chapter Recursion is in.

Our focus is on unlocking that value, using AI end-to-end consistently to generate better targets, better molecules and advance programs faster with repeatability. The ultimate goal is to deliver medicines that matter.

So this quarter reflects that focus. We're making progress across all fronts. First, on the clinical side, our first positive proof of concept with FAP. On the partnership side, a fifth milestone with Sanofi, reflecting our growing joint portfolio, tackling highly challenging targets , we're excited to share more about that today and the continued evolution of our end-to-end AI platform.

And last, but certainly never the least, disciplined execution, which is something we talked about at JPM, which has now extended our cash runway into early 2028. Look, there's a lot to cover today. So with that, let's jump right in.

Today, we'll be making some forward-looking statements on this call, so please refer to our filings for more information.

All right. We always at Recursion, start with the end in mind. And in that case for us, like I said before, it's medicines that matter that are truly differentiated. But in order to do that, you have to use the right data, models, compute and more. So look, there's a lot of talk about data. But what really matters is data that's high quality and fit for purpose. And at Recursion, our foundation has been building high-quality data at scale, not just 1 type of datasets, but multimodal across the board. This is where pioneering the lab in a loop, pioneering the wet and dry lab has become incredibly important so that we not only generate data, but then we generate purpose-built models that we test, learn and improve.

The other thing I want to say is we sit in a sweet spot of being able to leverage both public data and our proprietary private data. That's incredibly important to ensure that our models are impactful, insightful and unique. And on top of that, and I've mentioned this before, the importance of not just having the ingredients, but actually having a team who knows how to use it well.

Teams that are buying both fluent in science and in, but I want to add a third line. It's also important to have reps under your belt to know what good looks like and having talented teams that have reps is one of our core differentiators. But the ultimate secret sauce, I will say, is how it all comes together.

Having an integrated end-to-end operating system that is a continuous learning loop all the way from novel biology or novel insights through to the clinic. Look, for many of us that have actually made medicines and have focused on this, which is a humbling effort, we all know that improving 1 decision in R&D is simply not enough, it is the compounded impact of better decisions across molecule biological insight all the way through the clinic, that is what makes the difference. That's how you truly change not just the outcome but also the time and cost on how you do things. And that's what we are focused on at Recursion.

So what does that result in? First of all, in our clinical development, we have a diversified portfolio. We are very encouraged by our first AI-enabled clinical proof of concept with FAP, which has the potential to be a first in class for but we also have additional programs behind that. In addition to that, in our discovery portfolio, we also have another diversified set programs -- and specifically, I will just touch on the partner piece where we have brought in over $0.5 billion in upfront and also milestones, and we'll share some additional updates today.

I just want to say every single milestone we achieved is not just -- it improves the economics, but it's also a validation of the platform and a validation that we are learning fast in terms of what works, what doesn't to make our platform ever more intelligent.

In addition to that, let's just talk a little bit about the platform. I'm going to share this slide every time we have an earnings because this is so core to what we do. Number one, being end-to-end, like I said before, is critical. You have to connect biology, chemistry to ultimately the patient, which is really where the rubber hits evolve. That's where we are going. The other thing I also want to say is it's important to innovate not just on data generation, but also your models.

So we have state of the art and I'll talk a little bit more about this foundation models, not just in phenomics but transcriptomics and pulling those together in emerging virtual cell efforts that we're also focused on. We are also continuing to innovate on additional frontier models in the chemistry space as well as our newly built clinical development AI platform.

Again, it is that integration and how you harness it to unlock value that matters the most.

Next slide. So in terms of our strategic pillars, we have 3 main areas that we're doubling down on in this new chapter. Number one, tangible proof points. This is so important both from our clinical portfolio as well as our partner programs. Second, like I said before, in parallel, continuing to invest surgically in our platform ground and in areas that will enable us to have more of those proof points.

And third, but certainly not the least, pairing that bold ambition that we have with disciplined execution, how do we do more with less. So let's go through each of these. If you go to the next slide, 1 area that's really important for us is we like to track what are our wins and learnings as we go through each of these. So you'll get used to seeing that as well going forward.

First, in our first pillar, which is really focused around making progress around clinical pipeline as well as our partner programs. First, FAP, this is really, really important data for a disease that has no approved therapies to date, durable and meaningful poly burden reduction. Second, today, we'll highlight our Sanofi collaboration. Just as a reminder, this is where we're tackling challenging targets in I&I and oncology and leveraging our AI component chemistry component of our platform to design novel compound.

And here, we just achieved our fifth milestone to date. We'll do a double click on this, but this is an example of the repeatability of our platforms, especially around using AI to develop chemistry molecules and small molecules. Second pillar is really focused on our platform. And I want to highlight 2 things here. As we look across the portfolio, we look at green shoots, as I like to call it, proof point where we're actually seeing that we can do things better and faster.

So 1 example is, again, in our AI-enabled chemistry platform. When we look across the portfolio, we're synthesizing 90% fewer compounds than what we see in the industry. So about 300 versus 2,500 compounds in test. This is because we are predicting more and making less. This is where in silico approaches should be guiding us, and we're seeing that happen. And we're doing this 2x faster. So instead of taking the industry 42 months, we'll see on average, it takes us 17 months. We're going to keep pushing on this.

The other area, lets talk about biology, we talk constantly about the amount of unknown biology and what we're trying to do is generate and we have generated first in industry maps of biology, these huge Atlases where we are trying to uncover unknown biology. This is in partnership with our great partners at Roche, Genentech 2 back-to-back maps that were just accepted, and now the team is hard at work in translating those maps into novel biological programs.

And our third pillar, momentum with discipline. Look, we have a lot of things we want to do, but we have to do it with discipline and good financial stewardship. Financially, of course, but also operationally -- and we're really excited to share that, first of all, we've seen a 35% reduction in pro forma operating expenses year-over-year. This has come from multiple areas, sharper focus on our portfolio, yes, but then also optimizing our G&A and improving our platform efficiency, which an example of it, you just heard about in the last slide in terms of the number of compounds we're synthesizing, our speed, et cetera.

And the other thing that we're excited to share today is extending our runway to early 2028. All right, so let's dive into each of these pillars a little bit more. Starting with our wholly owned pipeline. Look, when we look at the number of programs here, we have a diversified portfolio. There are different types of differentiation across each of these programs, and I want to categorize it in 3 ways.

Number one, their programs with novel biological insight on our platform. Number two, there are programs that have emerging biology, interesting biology, which is unconquered not validated yet, and we have developed optimized programs. And then the third is really focused around areas that have validated biology with a significant unmet need that still exists from a patient perspective.

So you've seen this slide before, we always track which components of our platform are we using across our various programs. So let's dive into a little bit more around the 3 categories, starting with the platform-derived novel biological insight. All right. Two programs that exist in that category. One, FAP-4881. First of all, I don't need to say again, but like the reason why there's such a significant unmet need. There is nothing approved for these patients. This is a disease that is hallmarked by hundreds of polyps. Each and every 1 of which is precancerous and has 100% risk of CRC colorectal cancer by the timing of 40. More than 50,000 addressable patients in the U.S. and EU.

The Recursion differentiation is using the Phenom early version of the Phenom platform to ascertain in an unbiased fashion that MEK1/2 inhibition could actually work and FAP. We have just completed our Phase II study. We had a positive clinical POC, which we just shared in December and I'll share a little bit more about the data, just to recap for those who might have missed it. And 1 of our core next steps, and we're on track is to initiate FDA engagement on the registrational path first half of 2026.

We also have another program that has similar elements from a differentiation perspective, RBM39. RBM39 look is going to be potentially important in genomically unstable cancers. And from the patient population, as you can see that -- that impacts a wide patient population. The differentiation for conversion in our platform really came from uncovering this MOA and the connection it has to CDK12, which is known to be important for DDR modulation for many decades challenging to target because of the similar homology with CDK13.

Right now, that program is in these 1 monotherapy dose escalation, and we expect to share an early Phase I update on safety and PK first half of 2026. So later half of this year.

All right. Let's go to the next category, an emerging biology that unconquered biology and where we can optimize program. There, we have CDK7 and ENPP1. And you'll see what we're doing from a optimizing the program perspective is both on the chemistry side and also on the clinical development.

So let's start with CDK7, CDK has been known for a long time to be an important central master regulator, both of cell cycle control, but then also a transcription, which are with a wide variety of patient populations that are addressable, given its centrality in oncology.

From an Recursion differentiation perspective, others have tried this target before. And 1 of the key challenges has been optimizing the PK/PD authorizing the therapeutic index. That's where we have leveraged the second element of our platform, the active chemistry in order to optimize the molecule, especially around gut permeability.

We also are leveraging our platform in order to figure out which patient population should be going to that could potentially impact the most from CDK7 inhibition. Progress right now, we finished our Phase I monotherapy dose escalation, maximum dose has been selected, and we are in progress of the combination study, which is focused on ovarian cancer second line, and platinum-resistant with more data expected first half of 2027.

And again, apologies, we're working very hard at Recursion, which is why I have lost my voice, but I won't try to make it through the rest of this presentation. All right. The next program that's also in this category is focused on ENPP1. ENPP1, loss of a certain mutation leads to challenges in bone mineralization thereby leading challenges in fractures, pain, et cetera. Again, another lifelong disease that starts very early in the patient's trajectory life trajectory.

The Recursion differentiation here is focusing on a molecule that can actually be oral because what's available today for patients and also some of the efforts in investigational agents is around enzyme replacement therapy that requires a huge patient burden in terms of injection, subcutaneous, sometimes multiple a week. So what we wanted to do is design a molecule for ENPP1, which, again, challenging target, especially in this space for hypophosphatasia, which can be suitable for chronic dosing.

IND-enabling studies ongoing for this program right now, and we expect to have a go-no-go decision second half of this year on this program. All right. A third category. Look, these are some targets that have validated biology but have significant unmet need that exists. So let's take Malt-1. Malt-1 is validated from a target perspective in B-cell drivers. But some of the challenges really have been around limitations around tolerability. So we, again, leverage our Recursion platform to really design molecules that could design a way some of the UGT1A1 and other targets that have been seen, which is going to become increasingly important with combination with BTK inhibitors and others, which is what will be the ultimate efforts in this space.

So via Phase I monotherapy dose escalation ongoing, with early Phase I update data again on safety and PK monotherapy expecting first half of 2017. Another program that is a similar theme is LSD1. LSD1 is known to be an epigenetic regulator really trying to prevent or inhibit some of the differentiation that you see in solid tumors such as small cell lung cancer and also AML with some valid data seen in AML recently.

And the differentiation again here is, can we design out some of the challenges around tolerability, which has led to some DLTs and not be able to dose up high-notch as thrombocytopenia. This 2 Phase I monotherapy dose escalation is in start-up. And next steps is to have early Phase I update on safety and PK monotherapy expected second half of 2027.

Again, we expect to start to understand if some of the tolerability improvements we're trying to do, can we actually see that early on. This is our theme around early go, no go decisions to really understand is the design playing out in the clinic.

And another program that's in preclinical and a late preclinical is our PI3K 1047 mutant selected. PI3K in general is an important oncogenic mutation linked to resistance and relapse, et cetera. And I'll walk through a deep dive in terms of some of the latest data we have here, where again, remember, we use our platform to design a molecule that would be much, much more selective, over 100x selectivity over wild-type PI3K, which leads to some of the tolerability challenges that leads to dose interruptions and reductions and more to come there, but that's an IND-enabling study. Again, go no-go decision, second half of this year expected before we consider a Phase I initiation.

So I know that was a rounded trip around our portfolio, but I would love to actually double-click on 1 of our later stages, which is Reid on -- and then also 1 of our earlier stage and potentially entering up clinical pipeline, which is our PI3K program. So let's go through the REC-481. I'm just going to do a quick update here. For this program, we had our clinical POC late, late last year. And a couple of things to note. No approved therapies -- what we saw in our Phase II 3 months on treatment with 4-milligram QD of this MEK1/2 inhibitor, significant polybutenreduction, about 43% medium. -- highest -- 1 of the higher polypore reductions to date of the patients responded.

In terms of the AEs that we see very much in line with what you see from MEK 1/2 inhibitors, majority were Grade 1/2, Rash CPK and no grade 4/5 to date. What we also saw, which was even more encouraging was when these patients were then off treatment for 3 months. And remember, this is a chronic disease. So the on-off element is going to be very really important for us to understand. And we're the first to actually look at on and off in this disease area, we see continued durable protein reduction, in some cases, actually deepening and with a significant amount of the patients actually respond.

So this is a really important -- when I said at the top of the call, like it's important to not just have insight, but how do you turn those into something that's meaningful for patients and then ultimately new medicines. So I won't recap in terms of the insight to Proofpoint, but I'll focus on what's next. We're on track, as we discussed late last year in terms of the FDA engagement initiating that first half of 2026 to really discuss the registrational study design. In addition to that, we have already started the enrollment of 18 and over cohort. As you remember, some of the data we shared was for 55 and over, so we are already progressing on the 18 and over and then also advancing dose optimization efforts really inspired of what we saw with the durability data that I shared in the last slide.

So we expect to have additional clinical data first half of 2027 as well. So stay tuned, more to come.

Now let's move to another exciting program that we have in our pipeline. This is our PI3K 1047 mutant selective. So look, for PI3K, I'm sure you're thinking, Jacques, there are multiple PI3K. Why are we working on PI3K. First of all, this is a very, very important target across multiple solid tumors. The current PI3K inhibitors have been constrained, and we have some data that we'll share shortly, hyperglycemia, metabolic toxicity, dose interruptions, dose reductions, limited treatment duration.

All of that means, is there an opportunity to do better by patients. There's an unmet need that still exists. So what is our differentiation? And what is our thesis is really focusing on the 1047 mutant selective, which has 100x more selectivity over wild type, thereby having the potential to minimize risk for AEs. And in order to do that, we designed a molecule that can allow us to have that exclusive selectivity and with that, let me just actually walk you through something that's very exciting from a platform perspective. For this program, we started off with X-ray and structures, where we have proprietary structural insight. And that led us to leveraging our MD simulations, and this is what compute becomes really important. Our molecular dynamic simulations revealed a novel pocket. We then use our generative 3D modeling efforts and machine learning in order to design molecules, novel scaffold or this novel pocket.

And we were able to use other approaches or other ML approaches to really rapidly design our cycles so you get exclusive potency, but then also selectivity. Remember, it is that selectivity that leads to the tolerability challenges we talked about. And I want to take a moment, like just look at the lower bar here, in order to design this compound, we designed 242 compounds, 13 cycles in 10 months. This is what we want to see from a green shoot perspective at the platform. Can you do it better? Can you do it faster? And this is what we're tracking across our entire portfolio. I can tell you, compared to industry standards business. And this is what gets us excited. The data that we can actually do things better faster.

So then the next question is how does this molecule do? So we'll shift some preclinical data that we haven't shared before. First, let's look at how it does from a tumor reduction regression perspective. So if you look at the left-hand side over here, what you're looking at here is a dose-dependent tumor regression for our compound, which is in -- and we actually also looked at some of the compounds that are in markets, such as piqray and also Scorpion's compound, just to get a sense of how we're doing, and we see significant tumor regression, not just reduction but regression with this compound, comparable to what you see with scorpion and much better than what you see with Piqray. But given the standard of care, we also wanted to see the performance versus standard of care. So a surge with CDK4/6 inhibitors, which is come as a standard of care today. And what's exciting to see here is the synergy. Monotherapy, yes, you see reduction and regression with our compound, but you actually see synergistic efforts with the standard of care.

This is very encouraging. We actually have additional data. We only have so many trips we had space for where we also looked at other encouraging assets in the space such as CAP. And we saw improved tumor regression with low dose of our acid versus high dose cap. So all in all, this is encouraging from an efficacy perspective for this compound. But then we also wanted to look at tolerability. So here, what you're seeing is animal models from both naive wild type and then also obese diabetic animal models as well. On the left-hand side, you see we don't see any impact on hyperglycemia markers, in naive wild-type mice versus what you see with Scorpion and Piqray as well which is encouraging.

This is what we are designing the molecule to do. And then if you go to the right side, a little bit complicated, but we like to share data. Also in obese diabetic rat, you don't see hypoglycemia or the metabolic liability even at super efficacious dose for our asset versus scorpion and Piqray as well. So again, taken together, this is encouraging. But like I always say, the rubber hits thrown in the clinic. So what does this mean from a clinic perspective?

Look, current PI3K inhibitors, focusing on HR-positive breast cancer, they do have tolerability limitations, 65% to 85% experience hyperglycemia, large percent actually also have dose interruptions, dose reductions, some of this is driven by the hyperglycemia they're experiencing. And we also did some real-world analysis as well given our clinical development AI platform, but really thinking about what the target product profile would look like.

And you see the discontinuation about 3 to 6 months. And that's not a very long time. So I think the potential here, and we'll have to see, a, how the compound does through IND-enabling studies. So that's where we are today, is can we expand that patient population in twofold. Number one, in breast cancer, not just in patients that are nondiabetic, but also patients that are prediabetic and diabetic, if this trajectory of hyperglycemia markers and not having impact holds, that's about 50%, 50% in breast cancer.

And then the other is there's also a broader patient population such as colorectal and endometrial, we can also explore. And 1 thing I'd be interested to also look at is can patients because of the better tolerability stay on longer, longer treatment duration to really maximize the impact of these therapies. But again, clinical validation of improved tolerability requires is critical to confirm this expansion thesis. So if you go to the next slide, more to come. But again, we keep looking at these arcs. What was the insight, what do we design the molecule, what are the early proof points so far that you saw with preclinical data? And what's next right now is the go-no-go decision for Phase I which will be second half of this year.

So currently the study is an IND. All right. That was just our first pillar. Double click. We'll also do a little bit more around our partnerships. I'm really excited to share the progress we're making because remember, proof points can come from most of your internal portfolio, which is what we just focused on, but then also from our amazing partners that we're working with on actual programs.

To date, we have already achieved over $500 million in total cash inflows from our partnership, both upfront and milestones. And we've actually laid out some of those recent ones with the momentum that we've been achieving recently. But I want to emphasize something that sometimes gets lost, each and every 1 of the programs that we're working on has a potential for over $300 million in milestones and tiered royalty per small molecule program, some of the royalties are up to double-digit royalties.

So this is significant economics and also validation opportunity for Recursion. All right. We are very, very excited for the first time today to unveil our joint portfolio with Sanofi. I mean, Sanofi has been a fabulous partner. We learned so much from that exceptional team both across I&I and oncology. And what we're showing here is the multiple programs that we're working on filing with multiple early discovery programs as well. And you see, just like our internal pipeline, this is also a diversified pipeline. it is focused on challenging targets in I&I in oncology with molecules that have the potential to be first-in-class and/or best-in-class with programs that address very specific unmet needs, so thinking with the clinic and their mind, and to date, we have advanced 5 lead packages that has been delivered by Recursion across 5 of these programs and accepted by Sanofi today.

That's about $34 million in milestones to date in addition to the $100 million in upfront, so $134 million so far. And I just want to say, we have a lot of important work ahead of us with later-stage discovery milestones over the next months. And look, Discovery is parabolistic, we know some will work and some of these programs won, but it is the repeatability and the ability for our platform to have multiple shots on goal. That's incredibly critical for us. That's what you see with our internal portfolio. That's what you see as we work humbly with our partners to also advance important programs for patients in areas that are challenging.

So just double clicking on 1 of these. How do we get there? And remember, these are challenging targets and we are leveraging our platform. And I just want to explain 1 aspect that I think is really important. Our platform is not about 1 data 1 model, 1 asset. It's about the confluence of a suite of end that you use for the problem at hand. So we start with the problem first and then you have flexibility and optionality across our models to get to the best outcome.

And so again, our latest program, where we just got a milestone, our fifth milestone that we just achieved in the oncology program really focused on leveraging a, these are targets that are data poor. So we leverage both our physics-based approaches as well as our machine learning approaches, physics space to really understand the protein flexibility better by novel target novel pocket, and then leverage our machine learning algorithms in order to rapidly do our design made test cycle and find highly potent molecules that are now progressing to the next stage. Very exciting progress here. And stay tuned, more to come.

But this is truly what proof points look like, actually showing value that will matter for the medicines that we are working towards. But look, None of this can happen without a unique and differentiated platform. That is an ever important work in progress. So I want to just do a snapshot of the 3 components of our platform, starting with biology to Insight. I mentioned about the proprietary data that Recursion has been building for a decade, 50 -- over 50 petabytes of high-quality multimodal data, and I want to emphasize a multimodal piece.

Biology is complex and having diversity of data and having at-scale data sets complete to the extent possible, whole genome knockout overexpression. That's the kind of model -- data that you need to then build foundation models that are state of the art. We have a fantastic team that's working on this, whether it's the Phenomics foundation models or the transcriptome foundation models and combining those as the fusion of those models that are going to be really, really important in biology because we all know we need to connect input to output, genetics, transcryptomic, proteomic phenomic patient data. That's the effort that we're focused on.

And how do we leverage it? That's the so what matters, is creating these novel proprietary data sets. We call them biology maps, and we have those internally across different therapeutic areas. We also have it in neuroscience in GI or onc, with Roche Genentech, and that's what those insights is what fueling our discovery pipeline.

The next area is focused on leveraging AI for chemistry, novel small molecules can tell you this is harder than it looks. And we have used our insilico approaches to generate over 100 million molecules. One emphasis -- 1 point I want to emphasize is the point around synthetically aware design, as 1 thing to design molecules that are interested, but if you cannot make them, then that limits or you can make them, but the CMC is very challenging that really limits that end in mind. We always start with that end in mind, the target product profile. What can be a true drug that matters. We do that across our partnerships and our internal portfolio. And like I said before, 90% of these molecules are generated prioritize by our models. One thing that we're doing increasingly, not just leveraging automation but also a genetic orchestration so we can get things done better, faster and in a more unbiased approach.

And I mentioned this stat before, but I can't wait to mention it again. Look, we, on average, across the portfolio. So with PI3k, you said 242 compounds, 10 months, but across the portfolio, we like to be transparent around our data. 330 compounds is what we synthesize on average versus 2,500, 5,000 in industry, and we do it in 17 months on average versus 40 months plus for industry. This is -- these are the kinds of things that we track, and that's going from target all the way to advanced candidate. And as a result, we have over 10 development candidates across our internal portfolio and getting to that line with our internal and partner programs as well.

And last but certainly not the least, is an area that I get a lot of questions about as well in terms of our newly built emerging clinical development AI platform. What we have done first, and again, just like we did with our biology platform and chemistry, you got to build a really good data foundation. $300 million plus real-world lives that through both some internal work, but then also the great ecosystem integrated data partnerships. We're very opportunistic around that. So some of the early results, I mean, you can read the bullets here, but the 1 that I would point the attention to is enrollment rates.

Look, we are in order to execute on programs, you have to enroll in a very efficient and intelligent way. And some of our early results for some of the programs, we're starting to see 1.3 to 1.6x improvement. We are also just improving the operational piece that goes underneath it in terms of just starting studies faster by up to 3 months. All of this accumulates, remember the point around the compounding impact of decisions across the platform, this is how you do -- this is how you define drug discovery and development, leveraging AI.

And let me just give you a sneak peek as to how that works on the enrollment front. So we start with the 300 million patient lives. Our platform can actually generate a heat map, just like you see for biology or chemistry in different ways. But here for potential patients across and we're showing the U.S. here across the country. Then we go into deeper resolution at a state level and then at a ZIP code 3 digit code level and then at a site level. And what's really important here is we can also get data around the site experience or they're running that trial. And this is -- you can probably guess from which program, ovarian cancer trials. And how many competing trials that exist? That becomes really important. You don't want to fish in the same pond that can need to delay. And then beyond that, we can also get what how many patients do these sites have and then you can do a filter in terms of your inclusion, exclusion and what's relevant for the type of patients that we are looking for in a specific study. That filter does not get happen enough, I can tell you traditional approaches.

I call this, we talk about precision medicine, precision biology, precision chemistry. This is precision operations and starting with the patient in mind. With that, thank you for being with me for some time. I want to now hand it over to Ben Taylor, our CFO, to actually go through some of our financials.

Thanks, Najat. So 2025 was a year of financial transformation for the company. As a part of the integration, we decided to rebuild all of our corporate systems from the ground up. This was really important because we wanted to be able to apply the same level of discipline and rigor to our strategic decision-making that we do to all of our scientific decision making.

And so we looked at how every dollar in the company goes towards a specific quantifiable outcome. And that's how we were able to achieve the efficiencies that we did over the last year while still advancing a portfolio of 5 clinical programs hitting multiple different partner milestones, really investing behind the growth in our platform as well.

And all of that comes back to focus on those investments across our pipeline and technology portfolio that have the best risk return that are going to give us the most impact for the investment that we're making. And so that's how we were able to come back and have a 35% year-over-year reduction from proforma '24 to '25 and even come in 10% below the guidance that we originally provided in May of last year.

So we ended the year with $754 million in cash. Looking forward, our 2026 cash operating expenses are expected to be under $390 million. Cash operating expenses is a non-GAAP measure that we're going to be using to give you guidance. We have a lot of noncash expenses in our P&L. And so we wanted to provide something that showed what our cash profile might look like going forward. And so this is coming directly off of our cash flow statement. If you look at operational cash flow, and then you add back our inflows from partnership and transaction costs, you'll be able to get directly to this guidance number that we're using.

In addition, last year, it was really exciting to see that we crossed the $500 million milestone in cumulative partner inflows. We expect to continue to achieve those going forward. And in fact, we hit our first milestone earlier this month already. And so we do include probability weighting of some of those milestones in our cash flow projections going forward.

That's actually the really exciting part for me is not only were we able to exceed our efficiency expectations, but that actually means to extend out our cash runway. And so we're updating our guidance to go to early 2028 as of now. And with that, I will hand it back over to Najat.

Thank you so much, Ben. We'll wrap it up by just saying, looking ahead, we have a very broad set of catalysts that are coming up, and it's going to be a busy next 18 to 24 months. We'll see if I can recover my voice soon. In terms of this year, like I said, we're on track for our initial engagement with the FDA on REK481. We're looking forward to that and also initial data, early safety and then also PK for RBM39 and go no-go decisions for PI3K ENPP1, which are both in IND-enabling.

We'll also have additional data for 4881 early next year and then combo data expected for our CDK7 program as well as more early safety and data from MALT-1 and LSD1, recall for both of those, we design the assets to be more tolerable. So these are going to be important. And I know the partner catalyst looks like a small box here, but I wish I could physically expand it. because that's going to be very important. Partnerships with Sanofi, as we just discussed in terms of multiple programs has been progressing into more later stage development candidate and other milestones as well.

But in addition to that, these maps, these maps where novel biology is really -- would come from extracting that into new programs with Roche, Genentech, et cetera. So really, really important work that continues. And we continue to invest and push the boundaries in terms of our platform, defining what industry and standard really looks like for making medicines using AI, and as Ben just mentioned, pairing all of that important work with disciplined execution. We've really pivoted towards an outcomes-based budget where we test what every dollar value creation every dollar can drive, so doing more with less.

So I'll close by saying, thank you so much for the time. And also, our focus will always remain on value creation for patients. They're the ones that we ultimately serve. Patients are waiting and also for, of course, our shareholders. So thank you again for listening. And with that, I'm going to pivot to the Q&A section. And I'll also have our CSO, Dave Hallett, joining us as well, in addition to Ben Taylor in order to address some other questions.

All right. From Sean at Morgan Stanley and Priyanka as well. Thank you for the questions from JPMorgan and from Brendan at Council, so many people.

Questions around REC 4881 understanding what potential registrational pathway may look like upon alignment with the FDA, how we're thinking about providing a regulatory update and updated patient population?

So a long question. I'll break it into pieces. In terms of the regulatory update, as I mentioned, for us, we're on track for that engagement, initial engagement with the FDA first half of 2026, all ends on that for that. That's going to be really important in terms of discussing their potential design for registrational study, patient population, endpoints, we have a very compelling data set in terms of the durability and then also polybutenreduction.

In addition to that, I didn't cover it today in the interest of time, we also have the natural history data as well. So coupled with that, it's going to be really important for us to have conversations with the FDA. So that's point one. Point 2 is also around the updated patient population. So as I mentioned, 18 and over that arm is already recruiting as well as we're also looking at dose optimization schedules just given what we saw with our durability data.

So more data on that coming first half of 2027. Look, as we have meaningful updates across both fronts, as you've seen, we've done webinars, ad hoc. We like to be real time and transparent when we have more meaningful outcomes and updates will absolutely be sharing with the Street as well.

All right. Next question is from Alex from Bank of America. It looks like the cost cutting measures, cost optimization cost-cutting measures really started to take hold in Q4. Any one-offs that helped in the quarter? Or are these levels sort of the expectation for the go forward?

Ben, do you want to take that?

Sure, happy to. Yes. Thanks, Alex. So if you think about it, I agree with you Jon, it's really about efficiency more than cost cutting. So we have hit a point where we have gone through all of the integration, I would assume that, that is all complete. There's no big one-offs in the system. But what we really trying to do is come in with attitude, where we want to continue to find ways for every dollar to make more of an impact in the following years and months than it did previously. And so when you come in with that attitude, all of a sudden, you start to find ways to do more with less.

And that's where we'll -- we expect to be able to continue growing our pipeline, investing heavily behind our platform and moving things forward while still hitting those cost targets that we put out there.

Great. Thank you, Ben. I mean the only thing I'll add is, Alex, I think it's piece around rapid go/no-go decisions and how we are doing that, just the mentality in the mindset and also understanding and just taking a step back, the variety of areas we're working on and what is the value proposition across the different areas, which evolves as you generate more data, almost thinking like an investor, I think, is really important being agile around capital allocation, and that's what we will continue to do, of course, being driven by data. .

Great. Next question. What's the rationale in terms of the divestment? Do you plan to seek other technology partners? Does NVIDIA now have proprietary insights from the models you've trained, et cetera, et cetera.

Okay. I think it's going to be is a great question. Thank you so much. It's going to be important to decouple 2 parts. One is the investment from NVIDIA and 1 is our collaboration, our technical collaboration with NVIDIA. The technical collaboration with NVIDIA continues. I mean some of you might have just seen, we're going to be highlighted in a lightning round for NVIDIA's upcoming GTC presentation with high raise really being the Recursion being a pioneer in how to leverage automation, this wet and dry lab. This is not just words. This is actually in action. This is how we do millions of experiments a week.

The other piece is also our collaboration with NVIDIA around Boltz-2, 1 of the fastest computer in life sciences. That -- I mentioned the examples on PI3K around Sanofi using machine learning, using molecular dynamics. All of that is underpinned by our supercomputer. Our partnership with NVIDIA couldn't be any stronger. So that continues. In terms of the divestment, this really was, if you look at the public 13F filings from Q4 of 2025 is really a shift in NVIDIA's investment portfolio, to more larger on-strategy, supercomputer data center, et cetera, efforts. And so that's really a investment portfolio shipped, and we were not the only company. There are other decisions made as well. It's a collective shift from a portfolio to more on strategy, investment, large, large $1 billion-plus investments.

So those are to be 2 areas to be decoupled. The last thing I'll also say you also -- sorry, there are so many questions. Also ask a question, are we seeking other technology partners. We have a strategic partnership with Google as well in terms of cloud compute. We have the partnership, as I mentioned, with NVIDIA on machine learning and models, et cetera, but also on-prem compute, and we will continue. We are -- we've always been 1 of the pioneers in really bridging the world of tech and science, and we'll continue to do that.

All right. We'll take 1 more question here. From George's, with the recent positive preliminary efficacy data for REC 4881 in FAP and the achievement of your fifth milestone with Sanofi, what specific metrics or historical comparison from your current clinical portfolio, best demonstrate that Recursion is improving the probability of clinical success or speed of development compared to traditional discovery methods.

I'm going to hand it over to Dave Hallett to get us started, and I'm sure we can also add some more comments to there.

Good morning and good afternoon to those of us in Europe. I think I'll maybe start from the Discovery perspective, I think at during the last presentation has highlighted a number of themes. One is about the repeatability of kind of delivery. I think I specifically highlighted in the burgeoning Sanofi kind of pipeline that we're kind of -- that we're building together. This is kind of a repeatable platform that's kind of delivering both best-in-class and kind of first-in-class challenging targets.

Above JPMorgan and again, this presentation, I think we've highlighted the speed of delivery. If you look at the metrics that we're delivering in terms of numbers of novel compounds that we synthesize and test and the speed that we're getting to these development candidates. These are, I think, further demonstration that the role kind of technology plays in kind of accelerating that delivery. The proof is ultimately in the clinic. And clearly, we're very excited for patients in terms of FAP.

I think this is the -- the first example from our platform, where we've been able to kind of demonstrate a compound that came from Recursion has shown clinical proof of concept and obviously, the goal over the coming months and years is to show repeatability in that frame as well.

Thank you, Dave. And just to maybe add a little bit of a broader perspective, looking at the person five-plus clinical programs, a diversified portfolio on the clinic side, a diversified portfolio on the discovery side. And in the time and effort it takes to build a platform. I mean these data sets didn't exist, the models didn't exist. All of that I just think taking a big step back, we are not a 1, 2 asset biotech, and we are a tech buyer for a reason, which is the piece that Dave just mentioned really well, which is what we're really trying to focus on is the repeatability, the scalability making all of this much more engineering focus using whether it's genetic agents or automations to do things better and faster, taking toil out of the system so we can supercharge our scientists more and more to do the hard work.

And I just want to emphasize the hard work of drug discovery and development. drug discovery and development inherently is probabilistic. Most things don't work. We have a 90% failure rate. So we know that multiple shots on goals is going to be important. So that's the kind of fortitude and resilience as needed in the space, and we're adding an area to worlds coming together in tech and bio haven't really come together before and not just building models that are interesting but actually apply models that unlock value.

And so just to tie it together, we are constantly looking at metrics and stats. The team knows I call it green shoots, whether it is the number of compounds we sense, just 90% less than the industry, the speed with which the cost of our IND, we do the same thing in the biology platform. We do the same thing with the clinical development, as you saw me share where we're seeing improvement in enrollment and so forth.

We are -- there's so much work to be done. But this is what, quite frankly, gets us excited. It is hard, but incredibly challenging and rewarding work. So thank you all for your support to our partners, to our shareholders. but most importantly, to patients that are willing to take a bet on us in our programs and that are waiting, and we are working as hard as possible to really forge a new era of how medicines are made for patients that are reading. Thank you again for joining us today, and we look forward to sharing more updates in the coming months.

Hello, everyone, and welcome to this next session on the final day of the 28th Annual Needham Growth Conference. I'm Ryan MacDonald, and I lead Needham's Healthcare Technology research efforts. And in this session, I'm pleased to be joined by Recursion Pharmaceuticals CFO, Ben Taylor. Ben, thanks for joining me.

Pleasure to be here. Thanks, Ryan.

Yes. Thanks for coming. All right. So for those who are joining us, first of all, thanks for joining us. And second, we are going to have about a 35-minute fireside chat conversation here. I've got a number of questions and topics that Ben and I are going to run through. But if any of you have questions for Ben that are listening in, please put into the chat box, and we'll make sure to get that asked and answered if there's some time at the end. But with that, Ben, let's get started. So for folks in the audience who might be newer to the Recursion story, I'll give you us an overview of the business.

Sure, of course. Well, and I think one of the really important parts is actually taking a step back and thinking about why we're here at all. So there are well over 1,000 biotech companies out there, and we actually have quite a different mandate for most of them. So a lot of our foundational investors, so the Baillie Gifford, the SoftBanks, and the [indiscernible] of the world came in and wanted to invest behind us because they were looking for something that broke away from the binary risk model of biotech and really got more towards a balanced business model.

Now all of those investors are big tech-focused investors as well where they want to see transformational change in the industry coming out of technological advancement. And so if you think about our operations, everything we do tries to integrate in AI and automation. So we do both what we call dry lab and wet lab. So in silico work on the tech, we've really been leading the industry on a lot of the advancements and how do you image cellular systems and how do you think about biology in different ways? How do you create novel chemistries in different ways.

All of that is using AI to power it as a new tool and modeling system. We also are world-leading a lot of our experimental capabilities. So we've been able to do things by using a lot of automation that has allowed us to create new data. We have over 45 petabytes of proprietary data that we have generated internally in our own labs, which is a huge differentiator and allows us to drive a lot in those modeling systems.

And so those 2 things have come together to allow us to have both an internal pipeline. So more like a traditional biotech where we just announced great data on our most advanced asset, but we have a number of them coming through as well as a sizable partnership business.

So we've brought in over $0.5 billion worth of cash inflows out of our partnerships, and that's predominantly through our largest partnerships with Roche, Genentech and Sanofi.

Excellent. And before we dive sort of deeper into the business, I wanted to at least kick things off and just kind of talk about the recent succession plan that was announced with Chris stepping into the Chairman of the Board role out of the CEO role and him being succeeded by Najat Khan, obviously, a familiar face with investors and someone who's certainly not new to Recursion. But if you think about sort of this transition and sort of -- what are some of the key initiatives that the [ Jots ] excited to work on? And does her vision differ materially from sort of Chris' leadership today?

The mission, absolutely not. I mean if you look at where we're going long term, I think that's exactly the same spot. If you think about that mandate I was talking about, I mean, one of the problems in the industry is we have a 96% failure rate. And beyond that, all of the -- so I've been in the industry over 25 years, all of the innovations over that period, we are still only drugging about 10% of the genome. I mean that means there's about 90% of biology that's out there that we don't have a good way to understand.

We don't have a good way to drug. And so it's really untapped potential. And so that mission and vision of where we want to go in creating new ways to achieve that and improving the probability of success has not changed at all. I think it was a really natural evolution between Chris and Najat and Najat had come in and ran our R&D operations for 18 months before the transition. So they both worked well together and got to sort of build the business plan around it.

So it's been a very smooth transition from all of those perspectives. And obviously, Chris is still on the Board and helping to give us a lot of advice and guide strategy. So if you think about where we want to go, some of the changes that Najat really brings in, it's a lot of operational tone and discipline and looking at how we go about advancing into the next level of the company's trajectory. So Najat was at J&J for many years. She oversaw their portfolio review committee across the entire company. And so obviously, was helping guide J&J's decision-making process on what to advance and not.

During that time period, the value of their portfolio just massively, it was about a 3x on the growth in the portfolio. So that was a huge change. At the same time, she was also their Chief Data Science Officer. So she built from the ground up J&J's AI capabilities, how they think about data and data science. I think it's very telling that she did want to still come to Recursion. And that's really because we are an AI-native company, we just have so much more freedom to be nimble to operate in a different way and execute on it. And so during her first 18 months, we did a lot of work changing around the portfolio already and aligning the products that we wanted. We did a lot of work sort of building different ways that we collect data and analyze it.

Now going forward, I think what you'll see is a real transparency and discipline to how we make decisions, lots of go/no-go decisions so that it's not that classic biotech model of I'm going to spend $50 million or $100 million or $150 million, and then I'll get an answer. It's like, can we spend $5 million or $10 million and get to that same point. So you'll definitely see more of that. You'll see us be very disciplined also with the cash management. I mean, she and I have worked very closely together over the last year to take over $200 million out of the pro forma spend from 2024 to what we give guidance as for this coming year.

So that's a 35% reduction in our costs. We haven't actually changed what we expect to do. So we announced our strategy last May, and we haven't backed down from any of those pieces while still being able to fund that efficiency. And that's really a fundamental belief that we are a tech-based company. We should be finding more and more efficiencies. The gains that are going on with agents and with compute and with all of those different aspects benefit us directly. And so we're going to keep taking advantage of that and implementing it across the company.

Yes, it really has been great to see sort of the natural evolution and maturation of the company, particularly as you and Najat have entered the business and sort of kind of helped to guide strategy as well. So a very exciting point in the overall life cycle of the Recursion business. I want to take a step back a little bit and talk about sort of the market dynamics, macro dynamics. You talked about -- obviously, we're in the very early innings here within tech, AI, data and how sort of pharma companies, biotech companies utilize this.

And I'm curious to get your view on how -- from the conversations you have with your partners, how that large pharma sort of view on data strategy is shifting overall? Obviously, Lilly has made some announcements around opening up and sharing access to its AI models. The news of the week for them was NVIDIA, who's obviously also a partner for Recursion. They're going to kind of market and sort of partnering there. What sort of -- what's the market dynamic? How is it changing? And do you feel like this is a tailwind to the Recursion business or a validation, if you will?

Absolutely. I mean I take the step back and say, okay, everything that the entire industry is working on now covers that 10-ish percent, whatever the right percent is that we actually are -- have some understanding in biology and chemistry right now. I mean that other 90%, we need new tools. We need new ways of looking at data to be able to penetrate into it. And it's fast. I mean any of us can say we don't have our arms around disease right now.

There are so many different illnesses, ailments, waste just to be more healthy that we currently can't do. And so what you need to do is start to come up with new ways to create and analyze data. And that's what we were a very early pioneer in. And I think Lilly is definitely taking a great step. I love the fact that they're building the supercomputers and the labs. I would expect more people to do it because if you actually take the results that we're seeing, and it's really important to remember, we're working with AI, but we create things, right?

Like this isn't a concept that we're doing. We create a chemistry or a biological target that you can go to a lab and test. And what we see is we are achieving things that traditional methods didn't. So it is not surprising to see other people say, "Hey, I want to start to look in that area." I'll say there's a learning curve to it, and I'm sure the industry will get their arms around that learning curve, too. we're going to continue running fast. But even if people do also try and really run fast and do put a lot of investment in, the amount that we are not doing right now is really mind-blowing.

Yes. I was going to say so there's plenty of opportunity like -- because like you said, there's only that 10% right now of the targets. But for those who like look at this and say, well, okay, some of these large pharma to try to build it themselves instead of maybe partnering, you talked about the learning or alluded to with the learning curve and how steep that is there. Can you just refresh investors like on the defensibility and like how hard it is to build -- Recursion is built?

Sure. And I mean, there's a couple of different components to it. I'll start with the data because I feel like that's something that a lot of people understand. So publicly available data is -- and actually, this is true of almost all of the privately held data as well. So most of the data that pharma companies have and other groups. It's done on a -- almost always an individual project-by-project basis. So imagine you, Ryan, decided to run an experiment, you're going to define the data that's collected, how it's collected, the format that it's kept, even things like how you're going to refer to the target name could be different than if I ran that same experiment, and it often is.

And so when you think about that publicly available data, one, the annotations are wildly different. The formats are wildly different. Two, there's been a long-held practice of basically publishing the positive and not the negative data. And the negative data almost teaches you more. And so those 2 things really work against a lot of the public data sets in being very strong. We have seen one success really come out of the public data sets, one strong success, which was AlphaFold. And AlphaFold and we obviously worked with MIT and NVIDIA and Boltz-2, which was sort of what we would look at as the next generation of that in bringing binding affinity as well.

But the reason that, that was possible for Google to go out and do originally was because for the last 30-some years, there has been a public database where people were collecting unified data on certain types of protein structures. And so there was more than 30 years of data that was annotated in a consistent way and available publicly. That's really the only part of biology where you'll find that.

And so that's why there haven't been more breakthroughs in other areas. That leads us to say, well, we needed to develop our own data in-house. And so we always look at publicly available data. Oftentimes, our partners will share data with us as well, and we'll try and use that as sort of a starting point for some new projects. But in the end, what we've had to do is really create a lot of that data in-house. And that was going back to that wet and dry lab piece. So for over a decade, we've been creating our own data in-house. We have about 45 petabytes worth of our own internally generated data.

And all of that is done in a format where we can use it to inform models and drive exactly the sort of outputs that we want. I think in addition, there's a level of understanding what data is important. When you first start collecting data, it takes a lot of time because you don't know what data is important. And now we don't need to continue generating the same amount of data to get the same result. First of all, we may already have data that gives us the answer. Second of all, if it doesn't, we probably need much less new data to get to the answer than we did 5 or 10 years ago.

So I think that's a really, really important piece on the data side. On the modeling side, there are a lot of people out there that can write algorithms now. I mean algorithms can write algorithms now. And so the important part is actually how do you validate it and then how do you layer on top of it things? Like I'm going to talk about something that's really unexciting to most of the world. But like we -- so we generate all of these different chemistries and have this generative AI. We were using generative AI a decade ago before it was a big term. But you can do that all day long if it doesn't create meaningful molecules, it doesn't matter.

And so like one of the things that we've designed is everything that our system generates has to be what is called synthetically aware. Now again, it sounds really boring, but that means it has to be able to be manufacturable in a reasonable way at a reasonable cost and something that is actually drug-like. And so these are things that you have to layer on top of it. And that takes a lot of time and trial and error to get right. The final piece is we're the only company that has that end-to-end spectrum. So we can create a new biological idea.

We can create a drug to drug that new biological idea. And then we actually use a lot of real-world data, patient data to be able to better understand that patient, better design a clinical trial. And bringing all of those components together, we actually use them interchangeably internally. So it's not like that's a handoff on the progression. We use our development clinical data at the very beginning of a process as well. That integration of those systems and being able to think about all of the different types of problems that plague a drug is really important.

Now I want to be clear, we do not solve all of those. We are a very applied company. And this is another -- the last really important differentiator that I'll talk about. We're much more focused on how do you create a system, an experiment that improves the probability of success of this drug or this program rather than trying to solve it for all of biology or all of chemistry or those sort of things. That is just -- it's unlikely to happen given how little we know.

And so what we want to do is really focus that in. So 75% of our spend is on our pipeline programs and our partnerships. And so that's all applied learning platform development and growth that we're doing.

Well, as you said, like you have -- the great thing about the Recursion business is that it differentiates -- how it differentiates from your typical biotech is like this platform is validated through your partnerships, you have great strategic partnerships with Roche, Genentech, Bayer, Sanofi, BMS. And so I guess, how is the evolution of those partnerships as they've matured and those relationships trended? And then what's the pipeline for similar partnerships look like over time? What do -- prospective partners want to see out of [indiscernible] ?

Yes. Really great question. So we've brought in over $0.5 billion from our partnership inflows already and actually have seen a lot of momentum in the most important element of that, which is the milestones. So not just getting in some great upfront payments. We love upfront payments, but the milestones are what tells you if the programs are actually working. And that's also where the partners are also get invested.

So for Sanofi, for example, we've hit 4 different program milestones for Sanofi just recently. And what that means is those are 4 different drug programs where Sanofi and the rest of the industry hadn't been able to solve the problem themselves. And so we come in that initial milestone is really around it looks like we have solved the problem. That was the design goal. Now we have to do a bunch of additional experiments to make sure it's a good drug to take into the clinic. But those are the sort of validation points that not only get us really excited and show that the platform is doing things that outside parties care about, also brings in great money, but it gets them very excited.

And so Sanofi and Roche, which are our 2 largest partnerships, regularly talk about us at their investor meetings and their different presentations, and they're invested at the C-suite level to those partnerships. So that's really the sort of thing that I look for is, are we hitting milestones, which means the technology is working? And do we have strategic engagement from the top of the company. And we're getting both of those to come through.

Now one of the nice things, we have fantastic economics on both of those programs. So Sanofi, for example, $343 million in potential milestones per program, $193 million of that is pre-commercial. So this isn't some big back-end loaded [indiscernible] thing. And then average double-digit royalties for each program. So really, really nice economics. Roche is structured similarly once we start doing design. There's just this big end of it. So we've brought in $60 million in what are called MAP milestones, which are us doing something different with cells and cellular imaging that can highlight new ideas in neuroscience.

So they've paid us $60 million in milestones just for achievements on that. And so -- you'll see that start to translate into drug programs as well. We've started to do that in oncology with Roche already and doing it in neuroscience should be very exciting. As far as new partners and what the other groups are looking for, it's an interesting question for us because, obviously, those successes draw a lot of interest. People are looking for a partner who can solve problems they can't do internally. One of the things that we're always weighing though, is the economics for those 2 relationships are very good. They're both very large. So for Sanofi, we can go up to 15 design programs. For Roche, it's up to 40.

And so we have a lot that we can still go deeper with them just on the existing contracts. And to some extent, part of the value proposition that we get is semi-exclusive value, right? I think they like the fact that a lot of other large pharma don't have access to our technology. And that's why we get such great economics. Certainly, I worked at a SaaS company for a while. We are getting far better economics than we ever got as a SaaS company, and we want to keep that up. So -- we look, we are in discussions. We'll always be and certainly could do additional partnerships. We could also do expansions of our existing partnerships, really a lot of room for growth.

Yes. Excellent. I want to touch on some of the recent data readouts you've had, so -- and forgive a tech analyst if I'm mispronouncing some of these things. But maybe starting with the ongoing study for REC-4881 for patients with FAP. What's the therapy aim to achieve for -- maybe just kind of give us the table set for their investors. And then can you give an overview on the latest results from the Phase Ib/II study? And what's the next key milestones for development that we should be keeping an eye for?

Yes. Well, so FAP is an orphan indication. And all that means is it's a smaller patient population. So it gives you regulatory advantages and some different commercial differences. But it is not the normal type of orphan indication. So it's about 50,000 patients in the U.S. and EU5. So it's a relatively sizable patient population in comparison, for example, to most cancer markets. So most cancer markets are sort of 10,000 to 30,000 or 40,000 patients.

So it's larger than most of them. And it's a really tough indication because it's chronic lifelong disease that if left untreated, will with almost 100% certainty progress to colorectal cancer. Most of the patients in it are familial. So they're getting it from one of their parents or some other family member. And so they know who they are. So that's also another big difference. Oftentimes in orphan drugs, you have to go out and find patients, these patients are seeing their physicians 3 or 4 times a year. Most of the time, they're getting -- they might be getting colonoscopies 3 or 4 times a year to have polyps removed.

They generally will have colectomies somewhere in their 20s, and that can progress on to more and more resections over time. So it's a really, really difficult disease, high patient and economic burden. And what we saw in the results that we just announced in December was we were able to reduce polyp burden by -- which the polyps are what relates to the cancer, all of them malignant, polyp burden by 43% median within 3 months. And just to give that -- put that in perspective, there is no approved therapy for this besides surgical intervention.

And the most compelling other clinical results were a 20% reduction over the course of a year. And so this was very strong in the reduction and very fast onset. Plus we actually took patients off for 3 months because we wanted to see if it was a sustainable result. And what we saw is it was. So the median reduction stayed about the same. It actually grew a little bit over that 3-month period, which is really exciting because you think about -- in an ideal world, you don't have to take a drug every day for the rest of your life. You can sort of take it intermittently for the rest of your life on and off periods and that would be part of the goal. We do want to look at things.

We only drug for 3 months. Does that deepen if you continue drugging for longer periods. And also, we're currently going back to the FDA to talk about the pivotal trial design. There is a pivotal trial design that we know that we could use. What we want to do is talk to the FDA and see if there may be a more -- basically a faster, more effective clinical trial design that we might be able to do. No one's ever had data like we did. We were also able to use our tech in another way.

So we have a clinical development technology business. We created an LLM over the course of a week to search through about 300,000 patient records and create the first-of-its-kind look at what is actual standard of care for these patients because there's never been a study of what does the patient journey look like? What are clinicians actually doing? And so we were able to use our technology to really get a great understanding of that and also look at what normal progressions would look like. And we saw, on average, polyp count increases 60% a year. And so that reversal puts more context on. So what we want to do is take that data to the FDA, have a really data-driven discussion with them about it and see what we can do.

Yes. And just what's your feedback or what's your sense of like the FDA under the current administration? It seems like they've been a lot more forward thinking in the use of data and AI and trying to find ways faster. I mean are you getting that sense that's sort of not just press releases and actual and like real discussions, if you know?

Yes. I mean we continue to have positive discussions with the FDA. And obviously, we love the initiatives that they're taking in AI. Always -- I'll never get in front of the FDA. And until we've got agreement on whatever they want to see, we'll continue talking to them about it.

Excellent. Excellent. The other program I wanted to at least touch on and talk a little bit about is the CDK7 inhibitor and the [ REK program ]. Could you give an overview of the program itself and then highlight any recent milestones and time lines for the future? -- investors should keep an eye on?

Yes. And it's funny because how we're -- one of the important things that we want to do in getting people to think about us as a business model is I'm personally really excited about CDK7. It was part of the Exscientia side. I come over from the Exscientia side of the merger, and that was a program we really grew up over the years. But we want to take all of the personal out of it. We've got 4 drugs that have really important data points over the next 12 to 18 months, CDK7, MALT-1, LSD1, RBM39.

And all of them have had important changes to -- or important impacts from how our pipeline looks at the biology and the chemistry to, we hope, improve the probability of success and solve really important problems. And so what we want to do is all of them are early stage. All of them could be blockbuster drugs. Let's let the data lead us.

And so as those readouts come out, we will inform the investors as quickly as we possibly can. If it's good, we'll be off to the races. If it's not, we'll kill it immediately. And so the nice thing about the way that we are able to do business is we can take a portfolio management approach to it. And that's really different from this is my lead, and this is my second compound.

Makes sense. Okay. Look, you've done a really good job on the strategic and some of the science questions. We're going to give you some CFO questions now, right down the middle of the lane here. Let's talk about the balance sheet, obviously. You reported or announced earlier this week sort of with JPM Healthcare that you're expecting to end 2025 with $755 million of cash with an expected runway of out to 2027.

How should we think about the rate of cash use for program developments? And just talk about the level of confidence you have? And maybe just what are the biggest swing factors from an OpEx perspective that could either shorten or maybe even extend that runway?

Yes. Really, really good question. And it's funny because talk about another boring thing that no one wants to hear about. We've spent the last year rebuilding all of our back-end systems. So I'm here in Salt Lake City right now, and my old team is here because we went live on all of the instances January 1. And it is actually transformational how you can manage a business.

This is why it becomes -- I am answering your question. What we've done is transform the entire company into what I'd call an outcomes-based model. So we can actually measure every dollar, every time commitment of people, everything else to an actual outcome that the business is trying to produce. And we did that across everything. It's not just our pipeline programs. It's like, okay, what technology development are we doing or what G&A activity are you doing, right? Like there is an actual outcome that we're trying to achieve. And so what we can actually do is look at this and say, okay, this is exactly what we're spending on X, Y, Z program. How can we be more efficient on that? We can, in real time, go from our financial projections to our financial reporting, and that's a simultaneous system.

So if you think about that question on what happens to the business model over time, you can be very, very nimble on saying, okay, CDK7 worked. You already have that scenario planned out and you know exactly what additional resources you need and what's going to turn on. CDK7 didn't work. I know exactly what's got to turn off. I know what of that is direct variable cost versus fixed cost that we need to figure out if that fixed cost makes sense anymore and how we're going to do it, right? And so that sort of flexibility allows us to be incredibly data-driven.

So if you take that step back and look at it from a business model perspective, now you start to think about, okay, we have our 100% wholly owned pipeline, and then we have our partnerships. Our partnerships are basically prefunded. So that business is made so that it's meant to be breakeven or slightly profitable from day 1, where we get upfront payments that cover our direct costs and then the milestones, if we have operating obligations, we'll continue to cover that. But like with the Sanofi programs I was talking about, those 4 programs, the next milestone is what's called development candidate.

So that's when they in-license it and then they'd start to take it into the clinic. One, those milestones are larger, but two, they end our operational obligations. So all of the payments after that and forever more would be profit to us. And so that's where we try and create these partnerships where we're getting the money upfront, so it's not capital off of our balance sheet to do the development. We actually get a lot of platform benefit off of doing those programs. And then once they're hopefully in-licensed, they become a profit stream going forward.

On the internal pipeline view, we have a lot of different scenarios based on how many compounds are moving forward. So if FAP is the only program that moves forward in the other 4, let's say, [ die ], then we would maintain the best economics on FAP and do the investment behind it. We'd obviously also be shifting that investment to bring new candidates up to bring that balance. If FAP continues on, which it's expected to into the pivotal and then we have those other 4 programs all be successful, we are unlikely, very unlikely to take all 4 of those programs ahead ourselves because you're talking about a cost burden.

So then you can think about out-licensing, you can think about co-development. There's a lot of different optionality that comes into it. But we don't have to make that decision today, and we won't. We're going to constantly be looking at what's the value of a program, any program, there's none that are sacred. What are our capital needs and expectations and what's the market value that we could get for it.

Excellent. And then as you think about the cash collections, the milestone payments that you get from those partners, what assumptions are you making for the inclusion of achieving some of those milestones within your current cash runway projections? And I guess, what level of visibility do you have in terms of the timing of that?

Yes, sure. So there is a -- what we found is the design programs themselves and the map builds are pretty consistent in timing, like within a few million dollars and within, I don't know, 6 months, we can probably tell you what the different stages will be and when the go/no-go will go. There is some variation, but it's pretty clear. We've done about a dozen development candidates so far. We've seen the data play out enough times. So we use that visibility to predict new programs as well.

So we've got enough data to be able to start to use that. And then what we try and do, almost as shame to say this, we actually use industry standard probability weighting rates for our internal models. Now obviously, our entire business model is on changing those probability rates to be better. But I would rather take a conservative viewpoint than say, you know what, our probability is whatever the industry standard is. And then we probability weight the milestones over time of that and we offset our operational burn. But we only do it for our existing programs. I don't believe in baking in new business development or things that you just don't have visibility on.

Makes sense. Makes sense. Music to conservatism on the probability way, music to investors here. So that's great. And maybe just like lastly, before -- as we finish up here, how big of a priority as you think about capital allocation is M&A. I mean, obviously, a very great platform you have today, lots of great data that you're generating every minute, every hour. But any -- how are you thinking about M&A? Or what areas would you kind of look at?

Yes. It's something we do think about a lot. We definitely don't need to do any M&A, but there's sort of 2 buckets where it might make sense. One is if there is a compound. So FAP, it's the only drug in our portfolio that we didn't design in-house, but it was basically -- it was sitting on the shelf at a pharma company, and they had no idea it could be used for FAP. No one in the world knew that a MEK inhibitor could be used for FAP until we showed them.

And so we went and in-licensed it a number of years ago. This was earlier on in the company's history. And it's obviously doing great, right? And so there is a possibility of finding new uses for drugs that people don't understand because they don't understand the biology enough. And we'll keep our eyes open for that. And we have had absolutely a number of pharma and other partners come by and say, "Hey, you want to take a look at our -- what's in our closet and see if there's any value there." So we'll keep our eyes on that.

The other aspect is really thinking about something where it builds our capabilities or utilizes our platform. I mean we won't do anything that doesn't directly involve our platform. That's just not our business. There's 1,000 other pharma companies out there or biotech companies that could do that. What we want to do is there has to be novel insight coming out of us. We only work in proprietary compounds for our own development. We only want to be doing something that's differentiated. So if there were aspect of technology or some other component of the platform that was really differentiated. And we thought that we could drive a lot more value than whatever it is getting right now, that might make sense.

Awesome. Well, Ben, we've come up on the time. By my count, you were closing a fiscal year, you completely redesigned your internal processes and spent a week doing conference presentations, investor meetings, I think you've earned yourself on that.

Good -- thank you. I don't think I will get one, but I appreciate the sentiment at least.

Well, thanks again for joining me today, and thanks, everyone, for listening in, and we'll leave it there. Have a great weekend, everyone.

Terrific. Thanks a lot. Bye-bye.

Hi, everyone. Let's get started. Welcome to the 44th Annual JPMorgan Healthcare Conference. My name is Priyanka Grover, and I'm part of the JPMorgan Biotech team. Today, our next presenting company is Recursion and presenting on behalf of the company is CEO, Najat Khan. Thank you.

Thank you, Priyanka. Good morning, everyone. It's a pleasure to be here today. It's a very exciting time for Recursion in terms of the recent momentum that we have and also our path ahead. So as I walk you through some of these slides today, I'm going to focus on and walk you through three specific topics. First, how we're doubling down on translating the insights that we see into proof points that truly matter. And we're just coming on the back of our first platform-enabled clinical proof of concept. So I'll share more about that.

Second, we're also going to focus on how we are surgically doubling down on certain areas in the platform that are grounded in impact. They really focus on the bottlenecks that we see in R&D. And third, we're pairing that big ambition with discipline, discipline in our execution, discipline in our financial stewardship and also discipline in how we operate. So with that, let's dive in.

Before I get started, please note the forward-looking slide, forward-looking statements on the slide. All right. So Recursion mission is bold and it's also very patient-centric. Our mission remains unchanged, decoding biology to radically improve lives. Two things to note here. First, is patient focused. All of our decisions and our investments are focused on making better medicines from patients that matter. And second, we also talk about decoding biology.

Biology is one of the most unknown areas and therefore, to decode biology, we needed to take a fundamentally different approach to discovering and developing medicines. To do that, what we have built is unified AI-native intelligence platform that translates complex science that unknown science and biology into medicines that truly matter. And we focus on medicines that are better that we do it faster and at scale because patients are waiting.

We often get asked that question, what makes Recursion different. It is not one dataset. It's not one asset. It's not even one model. What makes us different is the fact that we have built proprietary multimodal data at scale, this is a huge gap in the field today. We've done that with our wet and dry automated labs in Salt Lake City and London, Oxford.

In addition to that, the models that we build are purposeful. They are focused on the questions that actually address making better medicines. And a point that doesn't get appreciated enough is our talent. I've said this for a long time, bilingual teams and cultures, scientists that understand AI, and AI scientists that actually understand science. That's incredibly hard to do, but it's core to unlocking our differentiated data models and compute.

The other piece that I want to spend some time on is how this comes together. That is incredibly different. We are the first AI native company that is an end-to-end platform, from biology, chemistry all the way to clinic. Now you may ask, why is that important? For many of us that have actually made drugs. We know that you have to improve decisions at every part of the R&D value chain. It's insufficient to have a better model in biology, but not to be able to execute on the trial in the clinic.

We use AI to advance and accelerate decision-making and insight across the entire value chain with the end goal only one end goal to make novel medicines that matter. So what are we working on? Let's talk a little bit about Recursion by the numbers. We have 5 clinical programs in our portfolio. And as I just mentioned, we had our first AI-enabled clinical proof of concept with a potential first-in-class profile. This is a disease where there are no approved pharmacotherapies.

We also have 15 programs in discovery. And I just want to point out both partnered and wholly owned. But on the partnered front, we have pulled in over $0.5 billion in upfront and milestones. What does this mean momentum, clear, tangible value being created and each of these programs that we're working on with our partners, whether it's Roche, Genentech or Sanofi, have an average of over $300 million in potential milestones with up to double-digit royalties per program.

In addition to that, we also have $755 million in year-end 2025 cash, providing an expected runway until the year end of 2027. We have done a lot of work operationally to get there, and I'll get into that more. So a question we get asked often with this new chapter in Recursion, what are going to be the priorities. First and foremost, doubling down on translating insights to proof points that ultimately become new medicines. That comes from our wholly owned portfolio and that also comes from our partnered programs. I'll double-click on that more.

Second, we are going to be surgically focused in terms of innovation in our platform where it makes a difference, taking insights, connecting patient data to ensure that our insights are actually actionable. We've also built a clinical development AI platform that is very new in this industry. And then the third area of focus is pairing bold ambition with disciplined execution and underpinning all of that is our exceptional people. So let's dive into this a little bit more.

We had our first clinical POC with 4881. We saw meaningful rapid and durable polyp reduction with a safety profile consistent with that class. Where are we going next? The next step is to align with the FDA on a registrational study plan. In addition to that, we have 5 more programs coming with clear go-no-gos and differentiation derived from our platform. We also have external validation of our platform from our partners, 4 milestones achieved with Sanofi, where we're taking hard targets and designing small molecules using AI.

With Roche, Genentech, we have delivered 6 AI-powered massive biology. Remember the point on unknown and unknown biology, we're changing that. So what to expect next? For the programs that we have with Sanofi taking that and progressing it further into the value inflection point, later-stage discovery milestones and development candidate. And then for our maps, taking those maps, taking the novel insights and converting them into new partner accepted discovery programs.

In terms of our platform, we have been able to drive unprecedented scale in phenomics, now we're layering in omics data and patient data to derive high-quality targets. In our chemistry and clinical development portfolio we are going to go full on at scale. And in the third pillar, in terms of pairing bold ambition with disciplined execution, we've done a lot of hard work here. We prioritized the portfolio only 6 months ago and already have our first clinical POC enabled by our platform. We also streamlined our operations significantly.

Expect to see more of this clear go-no-go decision, clear disciplined investment in our platform and leveraging agentic agents to really improve not just how we make decisions on our programs and our platform, but just how we operate as a company. So expected to 2026 cash burn of less than $390 million, this is a 35% reduction in pro forma since 2024.

All right. Now I'm going to double-click on some of these proof points. So let's start with our wholly owned clinical pipeline. I'll double-click on our FAP program, which is REC-4881, but just to say, in addition to that, we have several programs with clear go-no-gos, which fall under 2 themes of differentiation. One, novel targets that have never been studied before, such as RBM39 in solid tumors or in fact, REC-4881 as well, where the insight of MEK1/2 inhibition and the loss or the rescue of APC was not known.

Then the other theme that we have are targets that are important, but challenging from a therapeutic index perspective. This is where our chemistry AI design platform comes in. So examples such as LSD1, CDK7 and more. Every single one of these programs leverages our platform. I won't go into it in detail, but it shows you how much the platform has evolved across biology to insight, insight to molecule and molecules to patient. And we leverage and track how we use our platform across every single program.

Sometimes it's focused in on novel biological insight. Sometimes it's on design and sometimes it's both. So let's dive into FAP, lots of words on this slide, but 3 important points to note. FAP is a disease with significant unmet need, over 50,000 patients U.S. and EU, no approved pharmacotherapies surgery and debilitating surgery is the current standard of care. And it's a progressive lifelong disease, starting in adolescents throughout the entirety of your life.

So our platform derived a novel insight where we actually used our phenomics approach to knock out the gene that actually drives FAP, APC and the images of cells with an APC knockout versus that of healthy looks different. This is where foundation models come in. We screen thousands of compounds and in an unbiased approach learned that this allosteric MEK1/2 inhibitor did the best job of reversing disease to healthy. We then licensed this asset from a pharma company, the other piece that we also use our platform is our new clinical development AI platform.

This is a rare disease with no approved therapies. We wanted to better understand the natural history and progression of the disease. Very important to power studies, but also in regulatory conversations. We built the largest, longest running registry with the University of Medical Center in Amsterdam, to understand that these polyps that are the hallmark of the disease, remember, every single polyp is precancerous with 100% risk of CRC, colorectal cancer, if not treated.

They don't spontaneously go away. In fact, they progress in a relentless fashion, which is why many of these patients get multiple several surgeries a year, and we learned that. Also in the U.S. We built an LLM in under a week, and we're able to scan across 260,000 U.S. physician notes to really understand and contextualize the current standard of care. This is what modern clinical development looks like. And let's talk about the data.

From a safety perspective, in line with what we see with MEK1/2 inhibitors, majority grade 1, 2, derm and rash are being some of the main drivers. In terms of our efficacy data, 75 patients -- 75% of patients responded. Rapid reduction in polyps in 3 months, 43% median. This is above and beyond what others have seen before. And what was even more exciting and a little surprising is the durability. When these patients are on drug for 3 months and then they're off drug for 3 months, 4 mg QD, the polyp reduction is sustained. That is incredibly important for a chronic disease.

So what's next? I love this arc. We go from insight to proof points and defining the registrational path. That's the core next step with the FDA, and we continue to optimize our dose, looking at broader age ranges, et cetera. All hands on deck. Next, let's talk about our partner discovery. This is another core pillar for us in terms of our proof points. Look, we brought in over $500 million in cash flows and a very fast momentum in terms of milestones that you can see here. But I want to uncover that a little bit more. So let's talk about Roche and Genentech. We often get asked the question, what is a map. Well, let's talk about it. For Recursion to build these maps, this proprietary data that doesn't exist. These are 2 examples in neuroscience. Recursion in our Salt Lake City Labs, we culture and develop over 1 trillion iPSC-derived neuronal cells or 100 billion microglial cells. We then and the wet lab work continues, do perturbations across the board, the entire genome, overexpression to be able to replicate some of the disease states that are important. And that's when our dry lab kicks in.

We build foundational models, AI models to be able to drive insights from those maps. That's what a wet dry lab driving insight really looks like. And what's the focus for us next? We've actually delivered more than 2, 6 of these maps with multiple milestones coming from Roche, Genentech. The next step is to drive those insights into novel programs. This would be a first in this industry. And we are working hand-in-hand with our great collaborators at Roche, Genentech in order to do the functional validation to truly translate them into programs.

Let's move to Sanofi. In Sanofi, we're using the chemistry design AI segment of our platform. Here, we're working on incredibly challenging targets, targets that we worked on for a long time unsuccessfully, and we're designing molecules using our AI platform. And this is where novel approaches. This is where the investment in the platform is critical. A lot of these projects are data poor using active learning approaches in order to make progress. So we have achieved 4 programs, milestones to date and more late-stage discovery milestones coming up. So stay tuned.

All right. Our second pillar was being incredibly judicious and surgical in terms of where we invest in our platform. This space is moving fast, and we want to ensure we're investing in areas that are grounded in impact that get us to more of those proof points that we're all waiting for. So let's start with the biology side. And we talked about genomics, incredibly data rich. What we want to now do is to layer on more omics data. So in a matter of months, the team built the state-of-the-art transcriptomic models that is now enabling us to connect our lab insights to patients' translation. This is incredibly important as we think about novel targets.

And why a novel target is important. Remember, that's what generates our first-in-class programs. This is generalizable across many different datasets, single-cell, BulkSeq, patient data, in vitro data and when you build good models, they tend to be very data efficient, 50x less training needed. More work to do here, but exciting progress.

Next, let's talk about our chemistry engine, AI engine. We have doubled down here significantly. You saw some of the milestones from Sanofi. You've seen some of the programs we have in our clinic and actually in discovery as well. We have generated over 100 million molecules. And what's really important is these are synthetically aware. They're not esoteric structures that you can't actually make with a partner or CRO and over 95% of these compounds are not handmade that AI generated and prioritized.

On average, just to give you a metric, we're synthesizing about 330 compounds to get to advanced clinic and advanced candidate in about 17 months. The benchmark, 2,500 to 5,000 in 42 months. That's how we're trying to do things faster as well. And then this is our newest part of our platform. $0.70 on the dollar in R&D to make a drug is actually spent in clinical development. And everybody spends a lot of time talking about discovery. We want to win in all parts of the value chain. So we have built our AI-driven clinical enablement platform where first, we pulled together the data mode and the data foundation, over 300 million patient lives.

And we have 2 main areas we're focusing on: one, how do we pick the right patients and the right indications for our programs to really improve the signal to noise. And the second, how do we just execute flawlessly, how do we run these programs fast? And you're already seeing some great wins there in terms of 10% to 40% increase in some of the eligible patient populations. So we have a more unbiased approach to protocol development based on actual real-world data versus assumptions and then also an increase in enrollment rate.

Much more to come in over the coming earnings, I'll be sharing more details about what we're doing here. This was just a quick flash. So looking ahead, we have a wide range of upcoming milestones. So starting with our wholly owned portfolio, REC-4881 engagement with the FDA and starting those conversations, the first half of 2026 is going to be very important with more conversations coming up to align on our registrational path study.

Second, we also have our monotherapy early safety and PK data for RBM39 program. We also have some go-no-go decisions for programs that are entering the clinic, PI3K 1047 Mutant Selective and also ENPP1. And for next year, we also have additional data we expect from our REC-4881 program as we do different dose schedules and then also the age -- the broader age label. In addition to that, we also expect some combo data from our CDK7 Recall, this is in combo in second-line platinum-resistant ovarian cancer and then also monotherapy dose escalation from MALT1 and LSD.

In parallel, significant partner milestones that we're also focused on that I mentioned before. And to round it all out, as I mentioned, $755 million in year-end 2025 cash with an expected runway until the end of 2027. And also a cash burn that we're watching closely so that every dollar goes into value creation around our proof points.

So with that, thank you for your time, and looking forward to your questions.

All right. Thank you so much for the presentation. So I'll ask the first couple of questions, and then I'll ask the audience. So feel free to raise your hands, and they will get a mic to you shortly. So just my first question happens to be on the REC-4881 program or FAP. What do you think are the underappreciated points of data from the REC-4881 December update?

Yes. Thank you, Priyanka. Great question. Look, I think there are 3 main themes of underappreciated area. I think one is really around the unmet need. I touched on that, but maybe to go into it a little bit more. This is not a rare, rare disease, 50,000-plus U.S. and EU patients, but what's more important is the lifelong nature of this disease. Most patients start in adolescents with hundreds of polyps in their colon. Majority of them end up getting a colectomy in their 20s or 30s. That is a huge quality of life impact, as you can imagine.

And having spent some time with physicians to treat these patients and their patients themselves, the questions are very consistent. How can I delay surgery? How can I prevent surgery? Next step, 90% of those patients end up even after you get a colectomy, the polyp keeps growing in your rectum and upper GI so you get your rectal pouch removed. And then over time, sometimes you also get a Whipple procedure. So just the progressive nature of the disease and the debilitating surgeries I think that's not fully appreciated. And this is why and the fact that there is no approved pharmacotherapies there's a huge unmet need for these patients. So that's point one.

I think point two is sometimes the novelty of the insight from our platform is also underappreciated. There was no prior knowledge that like a MAP kinase inhibition would actually reverse what you're seeing with APC loss. That was not known. It was never studied in the clinic. And I think that is something that's incredibly important to double down and gives you first-mover advantage to actually have novel insights and you can do that at scale.

And then the third, look, from a data perspective, there have been other studies in FAP before, but the fact that we see such a rapid response in 3 months, pretty significant from what others have seen and the durability given the relentless nature of this disease, I think are very important points to keep in mind. We have a lot of important work to do with the FDA and regulators and so forth. But we are excited and all hands on deck.

So you touched on this, but in 1H '26, you'll engage with the FDA to define a potential registration pathway, and you'll include real-world evidence assessments. What findings from these real-world evidence assessments would you have the Street focus on?

Yes. I mean -- so a couple of things. So first of all, we'll have our first set of preliminary conversations with the FDA first half of 2026 with more conversations after that to refine and finalize. This is where I think the FDA has provided a lot of guidance in terms of real-world data. If you look at the last 18 months or so, and even before that, I would say, it really helps to contextualize, especially for a rare disease, especially for a disease that there's no approved therapies in terms of what's the natural history. We're not talking about a couple of literature articles. These are, as I mentioned, the longest and largest running registry, over 200 patients over 20 years, which helps us also quantify and understand elements of how to power the study and so forth.

So I think from a supplemental contextualization is going to be very, very important and something that the FDA has encouraged for rare diseases and, quite frankly, also in oncology and the analysis in the study meets those specifications as well. So for the last decade, we've talked about more integrated evidence generation in our clinical programs. I think this is an example of that where you have real-world data to contextualize and of course, clinical data for efficacy and safety.

Any questions from the audience?

First of all, I want to ask like 2 questions. Yesterday, we also saw that NVIDIA was working with Lilly to build their AI labs, so there is a very obvious trend that the MNC is trying to learn AI from the discovery and let AI companies are also learning how to do the clinical trials and potentially sales and distribution. So these 2 model, which one do you think will be easier to pick up? That's the first question. And then secondly, we also observed another trend that like so many AI companies in order to generate cash flow, they are partnering with the MNCs. So in terms of investments, if the best assets are being partnered with the MNC from an investor perspective, like should we invest the MNC or should we invest in the AI company and why?

We clearly thought about this for a long time. Great questions. Look, in terms of the NVIDIA and Lilly announcement, I think this is -- we think it's great for the industry. I mean there's been so much talk about AI, but no one really doubling down to make the investment that it takes. I think it also says a couple of things. Number one, the need for actual high-quality data generation, the need for wet and dry lab, which Recursion has been a pioneer in. That is actually playing out in terms of having a lab that can do both. The second part that I think it's also important is it sets companies like Recursion that were ahead of the curve and also invested early in terms of having that data moat that's needed to be able to capitalize on this trend.

I mean, AI is not going away. It's really around how and who and what value creation you do. We were at the event yesterday with NVIDIA with Jensen and we just got awarded a Spark Award from him. Just -- he said something that was really important that often, your audience or your market might not even know what's the need, and that's really how NVIDIA started. And I think you're starting to see a progression and acknowledgment of the investment that's needed in the space, both wet and dry lab.

And I think you said MNC, but I'm going to take it as pharma companies, pharma companies, whether it's platforms, AI or even ADC and others there's always a build and partner approach. It's usually been an end. There is so much white space when it comes to biology and chemistry in the space. So again, our focus is going to be to continue creating value. That's what's going to set us apart. That's why you see our first pillar. It's translating those novel insights into proof points that matter to make medicines. So that's your first question.

Second question, as you were asking, in terms of cash flow, I think you said that some of these companies might be partnering some of these programs. I think you have to be really, really judicious in terms of which partner -- which programs you partner versus not. We're open to both approaches, but some programs might do better in our hands and in some programs, which our partners could actually accelerate some of what we're doing, we're very open and agnostic to that.

Yes. Maybe just a couple of points to add on to that. One, I think it's a common misunderstanding. We don't change the what. We change the how. And so we're still developing medicines for patients. We may be able to do the discovery in a different way. We may be able to do the development in a different way in the future, maybe we do the commercialization in a different way, but it is the how, not the what. And so I think that a lot of the fundamentals of the business model between what you're talking about in the innovation-driven companies and the MNCs.

We'll probably still continue to play out just the how will change dramatically because we are getting differentiated results. I think another important part as we look forward is trying to figure out those new places that we can go into because a lot of biology and chemistry and even on the patient side is really unknown. And so seeing Lilly make the investment that they're making is actually a statement of -- we know we want to try and get into that 90% or whatever percent of the biology we don't understand today. We need some new ways to do it. Let's make an investment in there. And so I think we've taken a first step very early on that's progressed us towards that. I would expect other people to continue doing that.

Thank you for the question. Just moving on now beyond REC-4881 in the pipeline, looking at REC-1245, the RBM39 degrader, you're going to have early safety and PK data in 1H '26. What do you think -- what is possibly the size and scope of that data readout? And could we see that data in like a medical conference or a possible press release webcast like we did in December?

Yes. No, great questions. Yes, look, this is a -- the scope of it is a Phase I monotherapy dose escalation expected to the size to be very consistent with what that is. What we're really looking for is RBM39, just as a reminder, came from our phenomics platform. It's a novel target. Phenotypically was similar to CDK12 that's been hard to drug because of the homology that you see with CDK13, whether it's transcriptional stress or DDR modulation are the areas that we're really focused on. So step one, novel target, we want to be able to assess safety and tolerability. That's going to be really important. And by the way, this is a degrader molecule. Second thing, PK, we are going through various dose levels, so we also want to look at the exposure that we see. And yes, in terms of next step, that's our first step. And as we learn more, much more to progress further.

Beyond these 2 programs, what would you have the Street focus on for 2026? I know even 2027, you have some very interesting catalysts upcoming, too.

Yes. For 2026, in addition to our wholly owned portfolio, also focusing on our partnerships. This is why I spent a little bit of time talking through it, whether it's with Sanofi, some of the milestones, we've got great momentum, 4 milestones as of the last year or so, progressing those further to later-stage milestones potentially even to development candidate, which then triggers onboarding that asset. The other would be for Roche, Genentech. Maps -- these AI-powered maps of biology are such a novel way of getting to that 60%, 70%, 90% of known -- unknown biology. The next step is translating those into programs. That would be a first and really critical source of first-in-class programs that I think a lot of us are looking for. So those are other milestones as well I would pay attention to.

I'd also say, I think 2026 is the first year that you really see the business model play out in the way that it was meant to be. A lot of our foundational investment was to change this from being a binary risk model to something that's a more diversified balanced business model. And so now you see us with 5 programs in the clinic with over $0.5 billion brought in from partnerships. We're actually able to make very data-driven decisions because we can look at it and say, there's not a single program that defines us as a company or our value. And so we're able to be disciplined, and that will certainly come through.

I think that's a really important point. We have a portfolio of programs for a reason. We have 5 programs and we're going to be making just rapid go-no-go decisions, so that we can double down on the areas we have the most conviction in. You saw us do that 6 months ago with our portfolio prioritization, that's always something we'll continue doing.

Any question from the audience? All right. Your cash runway is anticipated to be sufficient through year-end '27. What specifically is assumed from pipeline development partnerships, which we've kind of touched on through our conversations?

Yes. So that fully funds everything that we've talked about in the milestones that Najat went through. What we have assumed is a probability weighting of milestones from our partnerships. So we just looked at our existing partnerships and said what are all the things that we know could move forward and apply probability weighting, but besides that, it's just our operations.

All right. Final question for me actually. So now with the recent changes, you are now CEO, Najat, how are you thinking about the next chapter of Recursion?

Yes. I mean, very, very excited about the next chapter of Recursion. It really is the pillars that I talked through, we have a bold mission, and we are doing it in a very differentiated way. So how do we really, number one, translate that into the proof points that is really going to continue setting us apart and drive towards novel medicines that matter. I mean this is why I do what I do, like what it's always going to be patients that have better medicines for them.

The second is, we have always been at the frontier of innovation when it comes to our platform, when it comes to AI and being super surgical on which areas, can we win in, can we partner in? And can we show the impact in our proof points. That's going to be another area because this is how you turn a company where even with pharma and some of the announcements that were made is really that integration of tech and science, tech and medicine. That's how we bring that together.

And then the third would be, look, I'm a big fan of ambition, but then also discipline, discipline in our execution, discipline in our financial stewardship, and you've seen us do that already and also discipline in operations. That's where using agentic agents to just improve how we -- not just the decisions we make, but how we operate as a company, taking the waste out, taking the toil up. And last but not the least, look, people and culture have always been core without a great team, you're nothing.

And here, we have the pleasure of having a team that's super motivated, mission-driven, but also unique in their capabilities. I mean this is what pharma will look like, right? In the next, hopefully, very soon. But the combination of scientists that understand and appreciate and respect AI and have the rigor to use AI the right way. I want to underscore that enough. And then also AI scientists who feel like they understand the complexity of science, drug discovery and development, which is incredibly humbling. So it's an honor, and I'm humbled to be able to lead a company with such a mission and purpose.

Well, I am looking forward to see how the next chapter goes for you guys. Thank you so much for being here, and thank you, Najat, and Ben, for being on the stage with me.

Thank you.

Thank you.

Hi, everyone, and welcome. I'm Najat Khan, Recursion's Chief R&D and Chief Commercial Officer and incoming CEO and President. Let's jump right in. Today is an important moment for Recursion, for our teams and most importantly, for the FAP patient community. We are here to share updated safety and efficacy results from our ongoing Phase I/II TUPELO study evaluating REC-4881 in FAP.

Before we begin, please note that today's presentation includes forward-looking statements about an investigational asset. Actual results may differ materially, and I encourage you to review our SEC filings for a full discussion of risks and uncertainties. What you will see today is proof points of the arc we've been building towards at Recursion from an unbiased platform insight to emerging clinical evidence, and we are very pleased to showcase that through the positive Phase Ib/II results from this ongoing study.

First, the unmet need in FAP is undeniable. FAP affects more than 50,000 patients across the U.S. and EU5. And as often is the case in rare diseases, the true prevalence may be higher once medical therapies become available and the diagnostic rate goes up. Most patients face a lifetime of escalating interventions from frequent surveillance, endoscopies and repeated excision of precancerous polyps and adenomas, which ultimately leads to a series of ever more debilitating life-altering surgeries such as the Whipple procedure. Today, there are no approved pharmacotherapies for FAP. And the natural history work our team completed with the University of Amsterdam Medical Center reinforces what clinicians already see that this disease is relentlessly progressive.

What makes this program so uniquely meaningful for Recursion is how it started. MEK1/2 inhibition was not a target anyone was pursuing for FAP clinically. It emerged from an early version of our unbiased AI-enabled phenotypic platform, the Recursion OS, which highlighted selective MEK1/2 inhibition as a mechanism capable of rescuing the abnormal phenotype of APC deficient cells. And because APC loss of function is the central driver of adenoma initiation and precancerous polyp growth in FAP, the hallmark of the disease, this finding pointed us to a mechanism that could directly address the underlying biology of the disease, not just the symptoms or the downstream consequences.

And it is that mechanistic insight, which ultimately led us to in-license REC-4881 from Takeda, a high-quality allosteric MEK1/2 inhibitor that was deprioritized after being studied in a solid tumor setting. We then redirected the molecule towards FAP, a disease, as I mentioned before, with no approved pharmacotherapies. It's a reflection of what at-scale AI-enabled biology first platform can surface when you don't start with a preconceived hypothesis. And now we're starting to see that insight translate into the clinic.

At week 13, the majority of patients experienced rapid reductions in polyp burden. And importantly, at week 25, this is 12 weeks after stopping therapy, most patients demonstrated durable reductions. That off-treatment durability is particularly encouraging and exciting when we anchor it against the natural history, which shows the rapidly progressive nature of the disease when left untreated. Safety remains consistent with MEK1/2 class expectations with primarily grade 1/2 events, low grade 3 rates and no grade 4, 5 events observed to date.

The trough treatment-related adverse events include dermatitis acneiform, CPK increase and rash, all in line with MEK1/2 class events. And finally, our ClinTech and real-world data capabilities have been instrumental for the study, informing eligibility, contextualizing the single-arm data set and helping ensure we really understand the disease in the real world, not just through trial snapshots. Overall, this represents the first clinical validation of the Recursion OS platform, demonstrating our ability to translate unbiased AI-enabled insights into differentiated medicines for areas of extremely high unmet need. Moreover, REC-4881 is the first precision medicine approach truly directed at the causal biology of FAP.

As a next step, we look forward to engaging with the FDA in the first half of 2026 to discuss a potential registrational pathway. In parallel, we're also expanding the eligible patient population from greater than 55 years to greater than 18 years and continuing to optimize the dosing schedule with the ultimate goal of bringing a much-needed therapeutic option to this community.

But before we transition fully into the REC-4881 data, I want to take a moment to acknowledge the momentum, the broader momentum across Recursion, momentum that comes from the work of many teams and from the trust of our partners that have been placed in us. Across both our internal and partnered programs, we're starting to see real green shoots, places where AI-enabled biological insights, design work and innovative clinical work are beginning to translate into differentiated programs. These efforts don't just advance individual assets. They also help us learn, iterate and strengthen that Recursion OS platform itself with every single cycle. Our partnership momentum is accelerating.

Across all of our collaboration, we have now generated over $500 million in total cash inflows. For example, with Roche and Genentech, our teams have now created and advanced 6 Phenomaps, 4 in GI oncology and 2 in neuroscience, and the teams are hard at work to translate those maps into novel programs. With Sanofi, we've achieved 4 milestones in the past 18 months, and we anticipate additional possible milestones ahead. This momentum gives us a strong foundation and reflects a company-wide commitment to rigorous science and tech-first strategy and execution.

Now let's dive into the Recursion OS platform. The Recursion 2.0 platform we utilize today has grown significantly since the early days of the company. It is a true end-to-end engine focused on applying AI where it can have the biggest impact in R&D. This includes multiple modules from integrating AI-powered deep biological understanding with advanced chemistry design and AI-powered clinical development strategy and execution.

Now let's dive into how the platform was leveraged from one of our earliest programs. Let me briefly and transparently describe the contribution. This program began with an AI-enabled biological insight. We used an early version of the Recursion OS platform, our unbiased phenotypic screen in APC deficient cells, which consistently revealed a signal. Selective MEK1/2 inhibition restored key aspects of APC-dependent biology. And as I mentioned earlier, this insight led us to in-license REC-4881, a high-quality allosteric MEK1/2 inhibitor previously deprioritized in an unrelated oncology setting.

And why did we do that? Because this profile aligned with the biology our platform consistently surfaced. It was this biology-first insight that allowed us to direct it into a disease with no approved therapies. And since then, we have advanced the program using the full power of the Recursion 2.0 platform, including our ClinTech platform real-world capabilities and AI approaches, all of which have strengthened the translational path and inform the current clinical development strategy.

So let's dive into that a little bit more. I often get asked about our ClinTech platform, which has been a ground-up built over the past year. Let me briefly highlight how our capabilities have shaped this program, which has been essential to really understanding the real-world patient experience. Now look, natural history data is fundamental in rare diseases, and regulators are increasingly expecting it. Yet for FAP, historically, there's been very little quantified evidence on how patients progress outside of the clinical trials. To fill that gap, our ClinTech team deployed large-scale real-world evidence capabilities, examining more than 1,000 U.S. FAP patients and over 250,000 physician notes using a custom LLM-based algorithm. This gave us a clear picture of the FAP patient burden in everyday practice, the frequency of interventions and the relentless progressive nature of the precancerous polyp growth.

We then extended this work through an academic partnership with Amsterdam University Medical Center, where we collaborated on the analysis of one of the largest and longest-running FAP registries in the world, nearly 20 years of follow-up in about 200 patients. This allowed us to quantify for the first time at the scale, the true natural progression of polyp burden in this trial relevant population. Across both data sets, the story is striking and consistent. Untreated FAP shows predictable year-over-year polyp growth, a benchmark that is important to contextualize the magnitude of potential benefit we're about to show you with REC-4881. This is ClinTech and its best augmenting clinical development with unbiased real-world insight, informing our single-arm design in ways aligned with FDA guidance and anchoring our future regulatory conversations in the lived reality of patients and providers.

So today, I'm excited that you'll hear from our very own FAP clinical lead, Dr. Beth Bruckheimer. Dr. Bruckheimer brings more than 2 decades of experience across all phases of drug development and previously led the largest Phase II trial ever conducted in FAP. She will walk you through the disease, our program and the latest efficacy and safety data. Following that, I will join our Chief Medical Officer, Dr. David Mauro, and we'll host a fireside discussion with 2 outstanding experts in the space.

First, Dr. Jessica Stout, a principal investigator for TUPELO and a double-certified GI specialist who has extensive experience managing complex GI diseases, including the long-term multi-organ care required for FAP. She will be joined by Dr. Alfred Cohen, former Chief of Colorectal Surgery at Memorial Sloan Kettering and former CMO of Cancer Prevention Pharmaceuticals. Dr. Cohen brings over 50 years of experience in colorectal cancer treatment and research with a career-long focus on improving quality of life for patients with inherited and GI diseases.

With that, I'll hand it over to Dr. Bruckheimer to walk you through the data.

Thank you, Najat. It is my pleasure to present the results from the ongoing REC-4881-201 trial and our FAP natural history analysis. So let's start by reviewing some background on FAP so we can better understand the significance of our results. Familial adenomatous polyposis is a rare orphan disease caused by autosomal dominant mutations in the APC gene. As a result of this loss of APC, patients will develop hundreds of thousands of adenomas or polyps in their colon and rectum by their late teens and early 20s. In the absence of surgery, these patients have 100% risk of developing colorectal cancer by the age of 40. These precancerous adenomas progressively accumulate and there is little evidence of spontaneous progression.

As Najat mentioned, currently, there are no approved pharmacotherapies for FAP. Patients face a lifetime of endoscopic surveillance, frequent excisional interventions and life-altering surgeries that impact quality of life, morbidity and long-term outcomes. There are greater than 50,000 patients across the U.S. and EU5 with FAP where the only option is the standard of care surgery and excisional interventions. However, this does not slow disease progression. Given this unmet need, our goal is that REC-4881 may fill this gap and provide another treatment option for patients with FAP to slow disease progression and therefore, improve quality of life.

FAP is a lifelong continuum of disease progression and intervention driven by chronic polyposis. As mentioned, in adolescents and early adulthood, patients typically develop hundreds of thousands of precancerous adenomas in their colon and rectum. As disease burden increases, most will require a colectomy to remove the colon and manage disease progression and cancer risk. While this addresses immediate cancer risk in the colon, it does not stop adenoma formation in the remaining rectum, pouch or duodenum. Post-colectomy, patients still require decades of repeat endoscopies, excisional procedures and approximately 50% of these patients will eventually undergo and require removal of the remaining rectum pouch in order to manage that uncontrolled polyposis. This, too, is a life-altering surgery that impacts quality of life.

Progression continues in the upper GI tract, where approximately 90% of FAP patients will develop duodenal adenomas, and these can often be difficult to manage endoscopically. Around 6% of these patients will undergo duodenectomy or Whipple surgeries. Of these, these are some of the most significant life-altering surgeries associated with high morbidity and mortality. So to better understand the extensive patient journey and unmet need, we used our ClinTech platform that consists of large databases of electronic health records. We applied custom large language foundation models on approximately 1,000 FAP patients with more than 0.25 million physician notes.

This analysis showed the continual polyposis progression, frequent surveillance and extensive excisional and surgical interventions that patients face to manage their disease burden. These notes revealed insights such as polyp burden is too extensive to be easily removed endoscopically to multiple revision surgeries were needed post-colectomy or even a Whipple surgery was needed to manage duodenal and ampullary polyp burden. So together, from the literature and our ClinTech initiative, you can see a high unmet need for patients with FAP and the need for therapies that can meaningfully slow disease progression and improve a patient's quality of life. So this is where REC-4881 comes in.

As mentioned, FAP is driven by a loss of function mutation in the APC gene. This drives cell proliferation and adenoma development over the course of a patient's lifetime. Our goal was to identify molecules that could rescue the loss of APC guided entirely by a cell phenotype rather than any prior knowledge or assumptions on the mechanistic hypothesis underlying the disease. Using version 0.1 of our unbiased AI-driven phenotypic platform and screening thousands of compounds, we identified REC-4881 specifically to rescue from the loss of the APC gene. As you can see in the heat maps, REC-4881 demonstrates a strong pheno opposite effect, suggesting it can rescue from this loss of the APC gene.

Indeed, in cellular phenotypic assays, REC-4881 treatment reverted cells from a disease phenotype to a healthy phenotype following the loss of APC. And now on the right-hand side, screening through a large panel of kinases, we also demonstrated that REC-4881 is a potent and selective allosteric MEK1/2 inhibitor. So overall, what is unique is that our platform led us to identify MEK1/2 inhibition as a mechanistic strategy to exploit a therapeutic vulnerability due to the loss of APC in FAP. We identified a molecule that had been discontinued after development in a solid tumor setting. It was then in-licensed and redirected to FAP, making REC-4881 the first MEK1/2 inhibitor ever to advance for this disease clinically. This insight emerged solely because the Recursion OS platform looks at biology differently by reading phenotypes at scale.

We feel that REC-4881 is uniquely positioned for FAP with regards to the mechanism of disease and then the mechanism of action of the compound. The loss of APC results in aberrant signaling, including the Wnt to MAP kinase pathway that drives proliferation and disease progression. REC-4881 by inhibiting MEK1/2 selectively blocks the activation of ERK, thereby suppressing cell proliferation and the downstream consequences of APC inactivation. We believe that REC-4881 has the potential to be a first-in-disease and best-in-class therapy for FAP. It is pharmacologically active at a 4-milligram dose, and it has a beneficial PK/PD profile that is really well suited for the treatment of FAP. This includes a long half-life of 48 to 60 hours and then [indiscernible] profile that allows for enriched unmetabolized exposure in the GI tract, the primary side of the disease. So this is really important.

So with that data in hand, we evaluated REC-4881 in the APC Min mouse model for FAP. In our preclinical studies, REC-4881 demonstrated a significant dose-dependent reduction in polyp count. This even outperforms celecoxib, which can be used off-label for the management of FAP clinically. On the right-hand side, and more importantly, REC-4881 also reduced the percentage of these high-grade precancerous adenomas in this model.

So now let's move on to the results from our ongoing Phase Ib/II TUPELO study and our natural history analysis. The TUPELO study was designed to evaluate the safety, tolerability, PK, PD and efficacy of REC-4881 in patients with FAP. As shown on this slide, outlining the Phase II portion of the trial, patients needed to have a confirmed APC mutation, they were post-colectomy and had to have adenomas or polyps in the upper and lower GI tract. At baseline, patients undergo an upper and lower endoscopy to measure that baseline polyp burden.

Polyp burden is defined in the study as the sum of all polyp diameters. Following baseline assessments, patients were treated with a 12-week course of 4 milligrams REC-4881. Today, we're focusing the data on that 4-milligram QD cohort where 14 patients were in the safety evaluable population, meaning that they received any amount of REC-4881. Of these 14 patients, 12 were efficacy evaluable, meaning they had measurable disease at the end of baseline endoscopy. They received at least 75% of study drug during the first 12 weeks of treatment and had at least one post-baseline endoscopic assessment.

For this study, efficacy is defined as a percent change in polyp burden from baseline to a time point post REC-4881 treatment. So back to the patient journey on this trial. After receiving 12 weeks of treatment, patients undergo another upper and lower endoscopy at week 13. And this is where we measure that primary efficacy endpoint. Following that point, patients go off treatment for another 12 weeks. That is followed by an upper and lower endoscopy at week 25. This is to determine the durability of response. So the effects on polyp burden 12 weeks after REC-4881 treatment has stopped. So let's recap.

Overall, we assess the safety and tolerability of REC-4881. Efficacy is measured as a change in baseline polyp burden for this portion of the study. And also just a little hint for what's to come in a few slides. In parallel to the REC-4881-201 study, we also analyzed real-world data from Amsterdam University Medical Center's FAP registry in order to contextualize the data from this open-label trial to the natural history of FAP. So let's start with some of the baseline characteristics from Part 2 of the trial. The average age of this population was approximately 62 and the primary disease site for the majority of these patients around 80% was the duodenum. This is really typical for this age group as they have more duodenal disease. Duodenal disease tends to be a little more indolent and occur later in life.

We also use the Spigelman staging system to classify upper GI disease severity in these patients. And we saw that the majority of patients were Spigelman 3. And lastly, we had sufficient polyp burden at treatment initiation where the median total polyp burden at baseline was 132 millimeters. So now moving on to safety. The 4-milligram dose of REC-4881 is consistent with the MEK inhibitor class of compounds. The vast majority of treatment-related adverse events or TRAEs were grade 1 or 2, the most common of which were dermatitis acneiform, CPK increases and rash. Again, these are primarily grade 1, 2 in severity. We observed 2 patients who had Grade 2 asymptomatic left ventricular ejection fraction decreases. There were 3 Grade 3 events, 1 dermatitis acneiform and 2 blood CPK increases. Of note and what is really important is that there were no grade 4 or 5 TRAEs and no unexpected safety signals.

So turning to efficacy. The week 13 waterfall plot shows a clear and rapid treatment effect with REC-4881. 75% of evaluable patients demonstrated a reduction in total polyp burden by week 13, where the median reduction was 43%. More importantly, of these, 60% of patients achieved reductions of 30% or greater. We also saw consistent activity throughout the GI tract where polyp burden reductions occurred in both the upper and lower GI regions separately. So together, these results suggest that REC-4881 is capable of impacting disease burden across the entire GI tract in patients with FAP.

As you may recall from our previous presentation in May, we had one unusual result with a nonresponder. Following that data cut and as described in the footnote 1 on the slide, we did a quality review and identified that this patient had a suboptimal bowel prep at baseline. So to ensure for an accurate and like-for-like assessment, polyp burden was reevaluated using video review restricted to the clean regions of the GI track at all time points, so baseline, week 13 and week 25.

What we are really excited about is on the next slide. And this shows that the effects by REC-4881 are durable. Even after stopping therapy for 12 weeks, 82% of patients remain responders at week 25 with a median reduction of 53%. Furthermore, 73% of patients maintained a durable response of greater than 30%. These results suggest that REC-4881's biological effect persists well beyond the dosing period. And we feel the next slide really exemplifies that. Here, we show the trajectory of the responses to REC-4881. On the left, during the treatment phase, you can see a clear and immediate downward shift in polyp burden in nearly all patients. However, what is really striking is what happens on the right-hand side after treatment stops. During that 12-week off-treatment phase, most patients maintained a reduction in polyp burden. And more importantly, in some cases, these reductions further deepened. This pattern supports the potential for REC-4881 to deliver a durable biological effect that is meaningful and persisting and even deepening beyond that dosing period. So we're very excited about this.

As mentioned, since the 201 trial is an open-label trial, we really wanted to contextualize this in terms of the natural history. So in collaboration with Amsterdam University Medical Center, we analyzed real-world data from their FAP registry. And this is done in a cohort of 55 patients with a similar profile to our inclusion and exclusion criteria. In contrast to the reductions observed with REC-4881 treatment, the natural history analysis shows that 87% of untreated patients experienced annualized increases in polyp burden. This was a mean annualized increase of around 60%. And what this data and analysis really underscores is the impact of REC-4881 treatments. So natural history shows disease progression, whereas REC-4881 treatment shows potential disease regression.

So rounding out these analyses, in addition to polyp burden, we also assess Spigelman stage, which as described earlier, is used to really determine the impact on upper GI disease severity. At week 13, 40% or 4 out of 10 of these patients demonstrated a 1 point or greater reduction in Spigelman stage compared to baseline. Encouragingly, at week 25, following that 12-week off-treatment period, all 4 patients maintained at least a 1-point stage decrease from baseline. Additionally, leveraging again our natural history analysis with AUMC, we can show that in the natural history of disease under routine care, Spigelman scores typically increase and do not decrease. So this suggests that REC-4881 can impact not just polyp burden reduction, but also reduce upper GI disease severity.

So now that we've gone through all the details of our analyses, let's summarize our key takeaways and next steps. REC-4881 represents a major milestone for Recursion. It is the first clinical validation for the Recursion OS platform with the potential to be both a first-in-disease and best-in-class molecule for FAP. Recursion's unbiased AI-driven phenotypic platform uncovered that selective MEK1/2 inhibition rescue cells from loss of APC and revealed that REC-4881 could be a potential targeted therapeutic candidate for FAP. We are the first to test MEK1/2 inhibition clinically for FAP, where we have demonstrated activity of REC-4881 at the 4-milligram dose.

From a safety perspective, the profile is consistent with MEK inhibitor class with a low rate of Grade 3 events and no grade 4 or 5 events reported. Efficacy was rapid and substantial. By week 13, we saw a 43% median reduction in polyp burden with 75% of patients responding. Current investigational agents have reported only 17% to 29% reductions over a 12-month time period. In addition to polyp burdens, we also observed 40% downstaging of Spigelman at week 13, which is, again, really critical for looking at upper GI disease severity. What is more important as an outcome from the study is that the REC-4881 response is durable at 12 weeks post therapy. So by week 25, again, after patients have been off treatment for roughly 12 weeks, we saw a deepening of the effect with around 53% median reduction and the response rate was even higher at 82%.

Other investigational agents have yet to report any durability of response. And then lastly, since this is an open-label trial, the clinical activity was further contextualized by the results from our natural history study, which showed an 87% annual growth in untreated FAP patients. Together, these results show the power of our platform to generate novel mechanistic insights leading to the identification of REC-4881 as a targeted pharmacotherapy that may address the underlying biology of FAP.

So what is next for REC-4881? There is a high unmet need for a pharmacotherapeutic option for patients with FAP that can really slow disease progression, reduce the impact of frequent surveillance, those invasive procedures they experience over the course of their lifetime. So looking ahead to fulfill that unmet need, we are expanding our trial to the broader FAP population 18 years and older and exploring alternative dosing regimens to see if we can broaden that benefit risk profile for REC-4881. Secondly, we plan to engage the FDA in the first half of 2026 to initiate discussions on a registration path. This strategy involves leveraging our data generated thus far, looking at the magnitude, speed and durability of effects of REC-4881 treatment and then combining that with our real-world evidence to support that potential registration path.

And then to close out and last but not least, we would like to extend a heartfelt thank you to all the trial participants, their families and clinical sites who made the research we presented today possible. So thank you.

Now I am pleased to pass the baton over to Dr. David Mauro for a fireside chat to discuss these results and their impact on patient care. Over to you, David.

Thank you, Beth, and good morning. I'd like to introduce myself. My name is David Mauro, the CMO at Recursion. I want to switch gears and welcome our expert panel of physicians who specialize in treating FAP. I'm pleased to reintroduce Dr. Jessica Stout and Dr. Alfred Cohen.

Dr. Stout is an assistant clinical professor at the University of Utah and serves as one of the principal investigators on the TUPELO study. Dr. Alfred Cohen brings decades of experience, having served as the former Director and CEO of the Markey Cancer Center at the University of Kentucky and was also the former Head of the Colorectal Service at Memorial Sloan Kettering. Thank you both for taking the time to join us today.

Let's begin by discussing patient management from different perspectives. Dr. Stout, from a medical standpoint, could you walk us through some of the patients' typical journey from FAP diagnosis to their initial treatment? And also, what are the main nonsurgical clinical concerns you have when managing these patients over the long term?

Absolutely, Dr. Mauro, thank you guys very much for having me here. So I think it's important to keep in mind the very intimate relationship that patients with FAP have with their gastroenterologists. I mean I see these patients at minimum yearly to do procedures. Oftentimes, I meet them in the middle of their journey somewhere. Maybe -- most of them have already had surgery, and I am just continuing on very frequent endoscopic surveillance for them and removing small polyps, hopefully, small polyps every year or whatever the right interval is for them. I think one of the most significant unmet clinical needs for these patients, there's an obvious answer here, but I want to save it for last.

To me, one of the most significant unmet clinical needs is rigorous upper GI surveillance. Because colon cancer is essentially 100% in these patients with FAP, I think we do a really good job pulling the trigger and saying, "Hey, it's time for surgery. And then after surgery, as I mentioned, I'm still performing proctoscopy or some form of endoscopy yearly. And we do such a good job in FAP at preventing colon cancer that now duodenal cancer has actually taken over as the leading cause of morbidity in patients with FAP.

Another significant unmet clinical need, I think, is that patients really should be referred for genetic counseling. This should be standard of care, but studies and then my anecdotal experience is that the minority of patients with FAP get referred. I think knowledge is power and I also think having genetic counseling and genetic testing early on with these families who know that there's an FAP gene mutation could help to expedite and make their screening process better and maybe avoid surgeries.

Finally, I wish I had effective chemo prevention to offer my patients with FAP to reduce their polyp burden and their overall morbidity and mortality. I mean I think that's why we're all here, right? Dr. Cohen, you made a great point when we were chatting earlier. We say reduce the polyp burden, but not all polyps in the GI tract are necessarily precancerous. However, in FAP, essentially all the polyps that we see are adenomas, so sometimes we'll say neoplastic or we'll say dysplastic. All of these are words to mean precancerous. These polyps are actually little ticking time bombs that will eventually become cancer if we don't remove them endoscopically or surgically. There's a huge need for something better to offer these patients and chemo prevention would be such a great adjunct.

Dr Stout, it sounds that these patients are living with a lifetime of disease that need constant management and sort of the relationship that you are building with these patients is over decades. Is that true?

Yes. I mean I am early in my career. When I took over here at University of Utah as one of the main gastroenterologists who was managing FAP, you can look back through, as you said, decades of records. And I review decades of records, these patients have been coming to see us for so long. They and their family members. And again, every year, they'll be seeing us having some sort of endoscopic procedure. And they have very complex medical and surgical histories sometimes.

Great. Dr. Cohen, similar question to you. I wanted you to focus though a little bit more on sort of the surgical management of these patients over the course of their disease. And also, in particular, could you speak to some of sort of the key decision points in when you would trigger surgery along the way?

Thank you, David. The issues when patients are referred to surgery are complicated by personal preferences by patients and family members as well as medical drivers that determine the need for surgery. As an example, if a patient is totally asymptomatic, but their mother died of colon cancer at age 28, not a common situation from a generation or 2 ago. Those patients, independent of their polyp burden in their large bowel, they're going to drive for surgery earlier.

But once you sit down and talk with patients about the reality of having what's called prophylactic or preventative surgery to avoid cancer and they hear about the reality of what their bowel function is going to be like, there is much less interest in going ahead when you're an adolescent or a young adult or you've just gone to college, just taken a job, just gotten married. And the thought of having multiple bowel movements a day, 6, 8, 10 bowel movements a day or having soilage in your underwear, nighttime soilage, other complexities that impact adversely on your quality of life, patients are much more interested in trying to delay the need for this life-altering surgery. So that has been involved with multiple "polypectomies," but we are in a good position to identify patients now at higher risk for cancer.

And so those are patients with FAP, but who have what are called advanced adenomas and that has been discovered over the past 25 years that patients with adenomas over 10 millimeters or on biopsy, high-grade dysplasia, which is just a microscopic finding or some others such as the villous feature, again, a pathologist microscopic finding. So it's possible not just based on having hundreds of polyps, but severe numbers of precancerous polyps that drives patients to surgery. So knowing that, what we're going for, what we're thinking about, David, as we move ahead is not only controlling the overall polyp number, but the numbers of these precancerous polyps.

And if we can prevent such precancerous polyps, we can delay the need for these life-altering procedures to something that's much more convenient for the patient, et cetera. So it's an exciting opportunity, and it's really the unmet medical need is, yes, we want to prevent cancer, but preventing cancer by taking out your large bowel rectum and duodenum is not an attractive option for patients. So if you have something that's an adjunct to the endoscopic management that Jesse does all the time, that will make a tremendous difference in terms of the lifetime of management of these patients.

Dr. Cohen, one quick follow-up. You talked a little bit about prophylactic surgeries. How often in your career have you dealt with patients that have had multiple prophylactic surgeries? And maybe the only other prophylactic surgery that I'm aware of in -- with regard to cancer management is mastectomies for breast cancer. Can you talk a little bit about the comparison in terms of how a surgeon might approach or thinking about a prophylactic surgery for breast cancer versus those for FAP patients?

It's a good distinction, David. So in breast cancer, once a family with an identifiable BRCA gene mutation has been identified, there is no good way of knowing at what point the patient is going to develop breast cancer. We might -- we know that the large majority will develop breast cancer, some at age 25, some at age 50. The difference for us in the FAP world is we can identify patients who are progressing rapidly towards invasive cancer, which, of course, we want to avoid by identifying these high-risk precancerous adenomas and trying to remove those endoscopically if possible.

In the -- from my perspective, the goal in the lower GI is protecting the rectum. So it is quite possible in doing prophylactic surgery, whether it's patient preference or based on more advanced adenomas, if one can control the disease in the rectum using a pharmacotherapy agent, then one can take out the large majority of the colon, but leave the rectum. And that leaves -- that prevents most cancers but provides a much improved quality of life. In the upper GI, the goal, of course, is to try to avoid a Whipple or duodenectomy. These are normally done for cancer but most Whipples in FAP are done just by virtue of many precancerous adenomas, and that's a disease that's associated with morbidity and mortality. So again, a drug that can obviate delay the need for any type of surgery would be greatly beneficial.

Dr. Stout, just a follow-up question maybe from your perspective as a PI on the study. And when you look at the data that Beth just shared, what is the most compelling takeaway from REC-4881 from an efficacy and safety perspective, just in conjunction with sort of the unmet need that you just mentioned? And what's its potential as a therapeutic option? I mean we have some ways to go. But when you think about FAP patients and the need for a pharmacotherapy beyond the relentless interventional and life-altering surgeries that occur today?

Great question. I think the most compelling piece of data that Beth and that you highlighted is the durability of REC-4881 from week 13 to 25. The fact that the biological effect persists after the dosing ends. And right now, there's no end in sight as to how long this effect may last based on the current data that we have. I mean that's so exciting. I think this initial trial data could position the drug as something that's taken daily or perhaps will find it's effective in a cyclic fashion, which might be a really alluring option for a patient who's experiencing a side effect, et cetera.

And also just to follow up on the conversation that we were having. I mean, my job is to hopefully prevent patients from going to Dr. Cohen. I mean, again, we've kind of acknowledged that colon cancer is such high risk in most of these patients that they need a colectomy. But I -- in some of these very aggressive phenotypes, I'm bringing these patients back as often as every 3 to 6 months. I mean, can you imagine having endoscopy that often. And so hopefully, some sort of chemo prevention could at least lessen -- at the very least, lessen the burden of endoscopy for these patients and then allow us as endoscopists to hopefully refer for surgery more often. But to run back to your question, the durability of the REC-4881 is very exciting.

Yes. That's really clarifying. And that's one of the reasons why we're looking to further optimize the dosing schedule, have multiple options potentially for patients given the chronic and progressive nature of the disease. Maybe just a quick follow-up on that. From a patient perspective, I had a chance to shadow some GI specialists as well, just to understand the disease burden a little bit for FAP from the patient perspective. Can you speak a little bit to that, especially starting younger and progressively over time, what would be most meaningful for patients here? So...

As much as I like to think that my patients love to come and see me, I am sure that they would be happy to have less frequent visits with me. I also just think it's important to consider it's not just that they're coming for endoscopy, but the patients in this particular region of the country, too, there are so few centers that consider themselves centers of excellence for FAP. And I have patients who drive or even fly over 500 miles to come and see me. So I think that's really important, too. We think about the endoscopic risks and the amount of procedures they need, but can you just imagine the amount of time they spend.

Yes. And especially having a solution that's an oral daily or cyclical, as you mentioned, really changes the burden on the patient significantly. Dr. Cohen, maybe just a similar question.

Let me just follow up...

Yes, please.

Try to imagine what it's like being 20 years old and sitting in a chair and a doctor is saying, well, you don't really need surgery, but you're going to have to have an endoscopy every 6 months for the next 50 years. You'd have to lift them off the floor after you told them that. So it is a lifetime disease. And anything that can not only prevent cancer, prevent the life-altering surgeries, and the morbidity and mortalities associated. But anything that had impact on the interval of surveillance endoscopy, if you can document that is a responder to your pharmacotherapy, that go instead of every 6 months, every year or every 2 years, just like we've done with colon cancer endoscopy. We used to do it on a yearly basis for prevention now 3, 5 and 10 years.

No, I think that bleeds very nicely into the next question, Dr. Cohen. Just wanted to understand a little bit contextualizing what the patient is going through, 20-year-old, 30-year-old, I can't even imagine a Whipple procedure that you mentioned that's usually preserved the oncology setting to some of the results from REC-4881 from the study you heard about today. What stands out to you the most from a therapeutic perspective for patients?

Well, I always go back to 25 years ago when celecoxib was approved as the only pharmacotherapy and the benefit was 30%. And everybody said, wow, but the physicians quickly realized there was very little, if any, benefit from that type of regression. So as you think back to the slides that were shown about the benefit and the durability, if we can show that the majority of patients taking the drug are going to have these types of major and durable regression that will clearly translate into a major benefit from the patients and from a commercial point of view.

Let me just add one other thing about the importance of durability is a lengthy study from the Amsterdam group showed 5% of patients become noncompliant. So no matter how personable the gastroenterologists may be, the patients after a couple of years, say, I'm okay, and they stop coming. So something that can be an adjunct so that, again, coming back to frequency of surveillance, patients getting surveillance endoscopy every 2 or 3 years rather than every 6 months, much higher compliance.

Absolutely. And that comes back to the patient burden, too, that Dr. Stout, you mentioned in terms of sometimes having to drive 500 miles to be able to see their physician, having an oral therapy changes the game. I think about so many other therapies. Remember when there was nothing approved for psoriasis and today with the management of care for psoriasis is incredibly different.

[indiscernible] impact on these families if they know that their children at age 10 test positive mutation, but they're not having -- they don't have to go through the same management algorithm that they went through that there's going to be some drug. It can be put off until they are married and have a life there and not having to undergo endoscopy or life-altering surgery at age 18 or 20.

One of the things that as I was looking at the data that was impressive is -- in my mind, I was looking for an efficacy profile that can demonstrate activity in both lower and upper GI and also nice durability -- so as I look at our data in comparison, it seems as though the potential to even increase the level of polyp burden reduction on. I think it has a relatively high potential, but don't want to put words in your mouth, but would like you to provide your thoughts. Dr. Stout?

Yes. I think what we're seeing so far with this drug just in the few clinical patients that we've performed procedures on already is very exciting. Just visually seeing that there are fewer polyps or that the polyps that were there before. You remember -- oh, yes, I remember that little guy at the 5 o'clock position, so many centimeters into the duodenum, wow, it looks a little bit smaller. I mean I think this is all quite exciting.

Important that the natural history is such that, as was mentioned, 90% of patients end up with duodenal disease but the bulk of those also have whether they have a residual rectum or the so-called ilial pouch will develop adenomenous disease in the remaining bowel. So having a drug that impacts both the retained rectum or the pouch and the duodenum avoids the need -- potential need as patients age of having a permanent bag or having to have a Whipple procedure. So the combination of effectiveness at both locations, I think, is absolutely a critical observation.

Yes. I mean if we're actually interfering with the natural history of this disease, which is progressive, that's also a very exciting thought.

Yes. No, great conversation. Maybe just to round it out, as you think about the future landscape of FAP treatment, we've touched on a little bit of this, right, elements of Dr. Stout, as you said, can you actually bend the curve in terms of the natural progression of the disease, which is relentlessly progressive. Can we have impact upper and lower GI? Can we maybe reduce the polyp burden reduction where some of these -- in some cases, they're carpeted with polyps. So it's even hard to discern the ones that are larger versus not or even to do -- have surgical excisions that are as accurate as possible. And the list goes on and durability and the patient need and the patient burden and how do we actually improve that. Any other components that come to mind as you think about the future of FAP treatment, the unmet need is undeniable, but any other things that we should keep in mind also as we progress this program through the next stages as well? Dr. Cohen, maybe starting with you.

So we are looking at development of a program. And I think that's where we are headed. That's something that has acceptable minimal side effects that does not need to be taken every day for the next 50 years, but can be potentially done on various intervals, the exact to be worked out probably once it's on the market.

Thank you, Dr. Cohen. Dr. Stout?

Yes. I think we use the word precision medicine so often, right? But if I just think about the future landscape of FAP in this perfect world eventually, future management, I think, will increasingly incorporate this genotype phenotype correlation in order for us to individualize surveillance intervals and also the surgical timing and then chemo prevention strategies. APC genetic testing helps predict polyposis severity, the rectal involvement, what Dr. Cohen talked about multiple times, other non-GI cancers like desmoid tumors, et cetera.

And I think that if we can incorporate genotype phenotype correlations, we'll be able to give personalized treatment decisions for these patients. And again, what does that actually mean? Currently, when I perform endoscopy for a patient with FAP, it's not incredibly clinically relevant for me to know the details about their genetics other than they have an APC gene mutation. And maybe I can make guesses about whether the mutation is on the ends or in the middle of the APC gene based on the number of polyps I see. But what I hope and foresee in the future is that someday I or maybe future gastroenterologists may use the patient's genetics to make precise endoscopic decisions about polyp excision, follow-up, when to refer to surgery.

And then most importantly, again, the reason that we're all here is hopefully use that information to figure out who we need to start on chemo prevention and maybe even someday, we'll have enough options for chemo prevention that we can even choose which drug to use based on the patient's genetics and their phenotype. That's where I see this going in the future.

Thank you. I couldn't agree more. Well, thank you, Dr. Stout and Dr. Cohen for joining us today. On behalf of Dave Mauro and I and the entire Recursion team, thank you for joining. Much more to come, and we look forward to continuing the dialogue.

Good morning. Welcome to the Jefferies London Healthcare Conference. My name is Dennis Ding, biotech analyst here at Jefferies. I have the wonderful pleasure of having Recursion Pharmaceuticals up here with us. We have Chris Gibson and also Najat Khan here with us. So thank you so much for coming.

Thanks for having us.

Thanks for having us.

So before we jump in, maybe just give some opening remarks in terms of some of the transition here, Najat. And just give us an update in terms of what's going on in terms of like the CEO and what kind of new thinking that you would be providing the company as CEO? And any kind of priorities for you over the next 12 to 18 months?

Go ahead.

Yes. No, first of all, thanks for having us. Great to be here. Yes. I mean, I think from a transition perspective, like very, very excited that we get to continue to partner. Chris will be the Chair of the Board. I've been at Recursion for 18 months. So I really had an opportunity to get to know the company joined before we did the Exscientia special combination.

From a perspective of going forward, it's going to be doubling down on 2, 3 areas that we've really been focusing on. One is harnessing all of our insights from a platform perspective and showing the proof points that I think the entire sector is looking for in terms of differentiated therapeutics that we can develop, leveraging AI in multiple different approaches across our platform, biology, chemistry, clinical development. I think we're one of the few companies that has that end-to-end integrated AI tech stack.

The second piece is going to be we -- the space is moving really fast in the United States. Like I've worked across a lot of modalities and platforms. This is the fastest one that I have seen so far. And so being able to double down in terms of our platform where we can win versus not is going to be another area of huge focus for us and for myself. And then the third thing is pairing that ambition with discipline, right? It's really important for us to have meaningful impact for patients and also demonstrate shareholder value.

So as you've seen even in the last year, we have been very disciplined in terms of our runway and also operating costs reduced by 35% without compromising any of our external catalysts, whether it be with internal programs or with our partnerships. So those are some of the 3 areas that we will continue to showcase further. And look, talent is scarce in the space, having the right talent and culture that's, as I like to call it, bilingual, understands AI, understands science, knows how to apply the 2 together. That integrated approach is critical to show those proof points, whether we do it with our wholly owned programs, and we have a readout coming out next month for REC-4881, one of the first programs from our platform and then also through our partnerships, such as with Sanofi and Roche, where we've had really good momentum in terms of some of the milestones that you've seen. But yes, very, very excited, much more to come. But Chris, do you want to share some thoughts?

No, it's just been fantastic building this company over the last 12 years and then being able to handpick a successor. And it took several years to recruit Najat and then the last 18 months partnering has been really fantastic, and I think the company is just in extraordinary hands.

Perfect. And then if we take a step back, just to talk about the platform, right? There are many AI drug discovery platforms out there. So what makes Recursion unique? And maybe talk about Recursion 1.0, talk about the Exscientia acquisition and like how that adds to the platform? And then I guess, talk about Recursion 2.0 now moving forward.

Yes. Maybe I'll kick off just with one of the basic kind of tenets of what we've been building at Recursion over the last more than a decade is one where we believe that biology is extraordinarily complex, chemistry is extraordinarily complex. And that most of the data that's going to give us the answers probably does not exist in the public domain. And so you see this big race today. There are hundreds of companies in our space and thousands of companies around the world who are building AI models based on public data. And what you end up with is convergence and commoditization of those tools because there's no differentiation in the underlying substrate for the tools.

I think what Recursion has done quite differently and frankly, our combination with Exscientia last year, we were attracted to them because they had done this similarly for chemistry is build the underlying data set at scale, the positive data, the negative data, all of it is being generated in-house at Recursion. And that's allowing us to build AI models on top of something different, something proprietary. And you see the power of this not just in what we've delivered so far at Recursion with our internal pipeline, but even our partners.

We announced last month a $30 million option payment from Roche Genentech. This is our second such option payment on building data sets from which we can discover potential new medicines in neuroscience. And today, we brought in over $0.5 billion from our partners for a pre-commercial biotech company to bring in that kind of revenue, I think, is really, really impressive. And it speaks to this long-term investment of building the data set and the AI stack on top of it. But I know there's other points as well.

No, I think that Chris covered it. The only thing I'd add is like drug discovery and development, you were talking about the various versions. It's a long game, right? Having one point solution that works really well in one area is insufficient. So I think having that end-to-end integrated tech stack takes time, but it's crucial so that at the outcome, you can actually have something that's truly differentiated. And that's what we focus on. Chris mentioned the piece around data. We have about 65 petabytes of data, 40 petabytes of which is proprietary to us. Every day, you will see different models being developed, but they're usually trained on the same data set, public data sets.

Where does that differentiation come from? It really comes from that high-quality data sets. We run about a couple of million experiments a week, a 600-person company to have that much data that's been generated and supporting multiple programs wholly owned and with partnerships is difficult to do unless you also automate and have a wet and dry loop where your models get good enough that it actually predicts which experiments you should do. That's where the world for discovery and development is going.

So the 2.0 platform that we talk about right now really has that end-to-end tech stack from biology, chemistry, clinical development. chemistry from Exscientia and then clinical development, I just want to point on that a little bit is something that we built from the ground that no other tech bio company has the use of AI end-to-end across clinical development from design to execution, which is recruitment. Nobody else has that. And that is incredibly critical given 70% of the funds to make a drug actually reside in clinical development.

If I can ask a little bit about your interpretation of what AI is, right? Because right now, the industry -- there are so many different variations of what AI is from ChatGPT to other things like that. And when you think back to the last 50 years, AI technically isn't really new, right? Neural networks have been around for many, many years, many decades. So how do you -- like what is AI to you? And how do you use that in your platform?

Look, I think one of the things that's important to note, you mentioned this, right? AI has been around for years. In fact, Yoshua Bengio, who's one of the fathers of deep learning and neural networks is one of our close advisers in Montreal. But there's been this convergence of tools and technologies coming together that are really important. So neural nets have been around for years, but the compute scale to scale neural networks has only been around for a handful of years. Even the storage capacity, right, just storing 65 petabytes of data would have cost $1 billion 2 decades ago, right, per year.

So the convergence -- we're at this point in time where it's not just one technology that's coming together, it's neural nets, it's compute, it's storage, it's things like CRISPR that allow us to modulate biology in different ways. And it's at the interface and intersection of all these different technologies where I think we're really starting to see something exciting.

Now to your broader question of like what is AI. Unfortunately, that term has become very, very broadly used. And now it basically means anything anyone says that's maybe using a computer that's sophisticated. And obviously, that's not the true definition. But I think it's important for folks in our field to differentiate between companies that are truly leveraging AI as a core piece of what they're building and companies who are using LLMs to put their marketing materials together.

Yes. The only thing I would add to that is interpretability of your models and traceability, data provenance, whether you're doing something more with regulators or even in discovery, that is very different for companies that actually generate their own data and develop models. So we have fantastic AI teams in-house that are building the models and iterating on them time and time again.

We have a team called Frontier Labs, which is our 0 to 1 cutting edge, and that's the team that partnered with NVIDIA and MIT and Boltz-2 that we open source that many of you have probably heard about. And then also AI models end-to-end in our platforms. So I think it's also important to look at the talent that's actually working. This is why the point around bilingual team and talent that's very critical, and I'm going to continue doubling down on that commitment to not just have the team but also the culture to go after the hard stuff.

Yes. Okay. And what about the regulatory landscape in terms of the FDA and how receptive they are around some of these AI drug discovery platforms and just preclinical and clinical development, like how are you guys going to capitalize on that?

Yes. I mean, look, we are very, very closely engaged on all of the changes happening, both in the EU and in the U.S. We also have Namandje Bumpus, who was at the FDA on our Board. So fantastic to have her insights. She worked very, very closely with Janet Woodcock and others part of the senior leadership at FDA for years. So there are 3 areas, I think, from a regulatory perspective that's very relevant from Recursion. So number one, we have programs in the rare disease space. So as you've seen for RDEA, RDEP, there's many different evidence and endpoint frameworks and guidance that's been coming out recently. So we're deeply engaged in areas such as FAP, which is a rare disease, 50,000 patients and so forth in terms of some of the progress that's happening in the early engagement, the openness to early engagement with regulators in terms of trial design, endpoints, et cetera.

I think the other area is also in oncology. I mean, a lot of you have heard about Project Optimus, Project Roadrunner. There are so many of these programs. What it really means is the evidence bar is going up and also how do you generate that evidence in earlier development around some of the oncology programs. So there, we're leveraging a lot of our multimodal data and also causal AI prediction around which patients will respond to turn a lot of our early development programs from exploratory to more validation. So we're doing that with some of our current programs so forth.

And then the third is also around other areas such as the reduced reliance on animal testing. Recursion has invested early on in terms of predictive ADMET models, really critical, especially when you're doing small molecule discovery as well as other approaches such as organoids and of course, a lot of our multimodal data. All of the foresight and early investment is really helping us being able to match some of the guidance that's coming out and also be one of the test beds, frontrunners in terms of leveraging that in our programs.

I'll give you a very specific example, just to make it real. We have our FAP REC-4881 program where we should have more data next month. This came from our platform in terms of using an unbiased approach, taking cells that have the biology representation, which is the APC mutation and then going from disease to healthy. We screened a lot of molecules. The allosteric MEK 1/2 inhibitor was #1 on the list. We have shown good data in vivo and promising data in our clinical study that came out in May. Why am I mentioning this? It's a rare disease.

Contextualization of open-label studies with the real-world data is incredibly important. And some of the guidance has come both in the EU and FDA, data provenance. How do you actually account for the data that's used to train your models and the limitations of that data. So a lot of that, the team is already captured in, in terms of the evidence generation that we're doing and having the totality of the data for regulatory conversations. That's how you stay at the front edge of a lot of the guidance that's coming up, and it requires a lot of early planning.

Got it. And if we can double-click on the FAP program. Can you just remind us just what that disease is? I feel like not a lot of people are familiar with it. Just the unmet need there and what are you trying to solve?

So familial adenomatous polyposis is FAP. This is a disease that's driven by mutations in the gene APC. And patients with this disease start getting hundreds or thousands of polyps in their gut in their late teens, early 20s and 30s. And ultimately, today, 100% of those patients will get colorectal cancer if left untreated. The treatment today is removing the colon of the patient. So you can imagine if you're a 20-year-old and you have a colectomy, this has a pretty significant effect on your quality of life. And as Najat mentioned, we were able to identify this mechanism that was unexpected in the space using this unbiased approach, take that through animal models. And now in our clinical program with the first 6 patients, we were able to demonstrate that this molecule reduced polyps in these patients by between 30% and 80%. And this is in 5 out of 6 patients.

So the median polyp reduction was about 45%. That is about double the highest polyp reduction percentage that's ever been seen in any molecule that's been explored in this space. And what I think is also really important is 5 out of 6 patients is a higher proportion of patients that are responding than the other molecules we've seen in the space. Now the caveat is very small end. And in a few weeks, we'll be able to share more data next month about what the next set of patients in this program looks like. And obviously, Najat and the team will then be able to take that data and if it's promising and continues to look good, potentially go and talk with the agency about how we might find a path to a broader set of patients.

And just maybe a couple of points to note. I mean, this is a rare disease with 50,000 in the U.S. and EU. So a substantial patient population. As Chris was saying, nothing approved to date and surgery is the current standard of care. There's a huge amount of unmet need to your question, in terms of finding alternate approaches to delay surgery to delay the risk of cancer. And most of these patients start pretty early on in their 20s. So we're encouraged by the data that we see, the polyp burden reduction, which is above and beyond what others have seen.

We also shared some data around the Spiegelman score, which is a really important staging that physicians use in terms of the risk of colorectal cancer being able to bring that down. And then the other thing is in the trial design, we're looking at both on-treatment and off-treatment. So the data that we just discussed is 3 months of treatment, which is much faster than what others have studied, which is usually around 6 months of treatment, so faster onset. And getting some of the data, hopefully, that we will next month in terms of off-treatment also gives us some flexibility in terms of the scheduling regimen in terms of can you do pulse dosing and so forth. So a lot more to do, but next step is, of course, seeing the data next month.

Yes. Okay. So how many more patients do you think we'll get at the update? And it seems like efficacy would continue to look strong in terms of polyp reduction, maybe we'll get some off-treatment durability as well. But what -- like how are you guys framing that update in December?

Yes, great question. So in May at the DDW, we had NF6 efficacy evaluable patients. The goal would be to try to get to at least 10 by next month. So just in a matter of a few months. 4 milligram QD is the dose, so we should see additional data. We're seeing good activity at that dose. And we should also expect to see a bit more on durability and of course, continue to monitor safety and tolerability.

Okay. In terms of safety and tolerability, is there anything particular with the data set earlier this year? Like when we look at the data, there were some signals, obviously, small end, but of LV ejection fraction depression. So just talk a little bit about that, like talk a little bit about how that could impact the market opportunity for you guys and what you guys can do to navigate that.

Yes, great question. From a safety perspective, the 2 main areas that we see is rash and some LVEF, Grade 2 and grade 0 so far to date. Both of them are on target for MEK1/2 inhibitors. You've seen that with other MEK1/2 inhibitors as well. On the first one around rash, we've been leveraging prophylactic topical steroids, antibiotics and so forth, things that have been used by others, and we'll share a bit more data in terms of the safety profile next month, but it's become more manageable from that perspective versus the earlier management of the disease, which is where some of the data we shared in May.

And in terms of the LVEF, look, what we're seeing so far, grade 0, Grade 2, reversible upon stopping the treatment. And again, not different from what we have seen and not unexpected from what we have seen. Now in terms of the scheduling flexibility becomes important. This is a chronic disease. This is why in the study, we're measuring both on and off treatment. We'll learn more with that data and then also explore some of those options as we think about a potential pivotal if the data holds next month.

Yes. At the same time, it is like a risk-benefit equation, right, at the end of the day?

Yes, always. Yes.

How are you thinking about the future Phase III pivotal program for FAP? Like in an ideal scenario, what would that look like for you guys appreciating that maybe the FDA is not okay with just polyp reduction? And maybe they want some event-driven or some kind of clinical outcome endpoint in terms of time to the removal of the colon, et cetera, right? So how are you thinking about the different scenarios in terms of the design?

Yes. I mean, great question. So first of all, for FAP, unlike many rare diseases, there is precedent for pivotal study endpoint, and that's a composite endpoint like a PFS composite endpoint that has elements of polyp burden reduction, which is usually the primary for Phase II, which is also what we are measuring. In addition to that Spiegelman score that I mentioned before, which we're also measuring and other things such as progression to surgery or even death. So it's a large composite endpoint. In our conversations, we'll discuss, again, it all depends on the data and the effect size, et cetera, in terms of alternate versions of a pivotal endpoint that is much more conducive to a chronic disease.

The other thing that we're also doing, as I mentioned before, is we'll have a real-world study in natural history, which I think is really important to contextualize what the progression for these chronic disease patients looks like in terms of polyp burden. You mentioned surgery and so forth. So lots of conversations. This is where the early conversations and discussions with the FDA, leveraging and abiding by some of the new guidelines and frameworks that we have an internal team that's working on with Board members and others becomes very critical. So much more to come, but step one is the data next month.

Got it. Okay. And moving on to some of the other assets in the pipeline. CDK7, maybe talk a little bit about that and some of the updates that you guys will be sharing next year.

Yes, sure. Happy to. So CDK7, important target. The goal here is leveraging our platform to design a better therapeutic index and also leveraging our clinical development AI platform to hone in on the right patients and indications to go after. So we just completed our -- a few weeks ago, we mentioned at our earnings call, our monotherapy dose escalation. We have an MTD dose, et cetera, et cetera. From a safety perspective, safety is similar to what we expected for CDK7, but trending lower on some of the GI tox that we see from other published data, which is in line with what the design criteria is. And in addition to that, we did see some early activity, 1 PR, some stable disease, et cetera.

As with most CDKs, whether they're transcriptional or cell cycle, we -- our goal is really to look at combinations for potential efficacy. We have initiated our combination study in ovarian cancer, second-line platinum-resistant using a couple of standard of care, one more U.S., one more EU to be comprehensive in terms of what the standard of care is. And the insight to go into ovarian cancer versus breast cancer, which we're also exploring other indications, really came from a confluence of what we're seeing in cell line sensitivity, in vivo CDx models and then also using our multimodal data, our data from our partnerships with Tempus as an example, and our causal AI modeling.

So it's really those 3 areas that we harnessed in order to take ovarian cancer as our first indication. Next year, we should just have more rolling data in terms of the patients recruit. We haven't given specific guidance other than the fact that we should have some early combination data in 2027. When in 2027, we'll give more guidance next year as we enroll the patients and learn more about the study.

Okay. What about other indications beyond ovarian?

We are exploring other indications, but we haven't disclosed anything yet. We'll make sure we tell you that as we do.

And then I guess, lastly, just on RBM-319, that's something that was just completely internally developed. So there will be some updates in the first half of next year. So talk a little bit about that and what you are expecting.

Yes. Just -- the only thing I will say that all of our programs, especially CDK7 as well is internally developed. We are developing the molecules, et cetera. I would say for RBM39, again, coming from our platform, the biology part of the platform, we wanted to see what else could we target that's important for DDR modulation and other transcriptional stress sort of related mechanisms that's similar to CDK12, but not -- doesn't hit CDK13. RBM39, novel target, first-in-class. We have a degrader. We've designed the molecule as well. First half of next year, we expect to get some early safety as well as PK data. This is a degrader. So we also want to see engagement overall. And that's happening. We're exploring this in solid tumors.

And the last thing I just want to say that we also have multiple milestones coming from our partnerships, as Chris mentioned, we have achieved 4 out of 4 milestones with Sanofi, where we are actually designing the molecules for targets that we co-aligned on, designing both the wet and dry lab. And for Roche/Genentech, a couple of milestones on these maps, $60 million so far in the last year. And I just want to highlight that sometimes we get asked that question, how much wet lab does Recursion do 100 billion microglial cells that we developed and then created these maps of biology and then also for some of the other maps of 1 trillion iPSC-derived neuronal cells. So I just want to highlight that there's a lot of work ongoing with both partners around data, around models and insights that's now translating into programs.

Got it. Well, I think we are out of time, but it sounds like 2026 will be an exciting year, a lot of catalysts.

Thank you so much.

Thank you.

Thank you.

"

"

"

Hello, everybody, and welcome to Recursion's Third Quarter 2025 (L)earnings Call. My name is Chris Gibson. I'm the Co-Founder and current CEO of Recursion, and I'm so delighted to have you all joining us today.

I want to start off by talking about something that I'm really excited about, which is our executive leadership updates. And it's my pleasure to share with all of you that beginning January 1, the amazing Najat Khan is going to take over the role of CEO, President and Director of Recursion. I've been working with Najat for the past 18 months in an incredible partnership to build our platform to deliver on our pipeline and our partnerships. And everything that I have seen has convinced me that she is absolutely the right leader to take Recursion through its next chapter. And I'm so delighted that she has agreed to take on that role.

I'm also incredibly excited that I'm going to continue bringing my passionate unapologetic founder energy to Recursion as the Chairman of the Board and as an Executive Adviser. And finally, I want to say a huge thank you to our entire Board and especially to our Chairman, Rob Hershberg. He's been an amazing Chair and an incredible mentor to myself and Najat, and I am delighted that he's going to continue on with us, I hope, for a very long time as our Vice Chairman and Lead Independent Director.

These are really exciting changes that I believe are going to position Recursion to affect our mission to lead the Tech-Bio space, and I am so, so thrilled to get to share them with you today. Najat?

Thank you, Chris. It is such an honor to step into the big role to be CEO and President of Recursion. And I want to thank you. I want to thank the Board and just our incredible team for their trust and partnership. And Chris, first for a few minutes, I want to say, look, your vision and the courage with which you have taken this from 0 to what Recursion is today is unparalleled. And you've not just been instrumental in building the company, of course, but also creating a new sector that never existed before, really truly blending the best of technology and science to make medicines and radically improve patient lives. There's so much to do. So I'm really -- I'm deeply grateful, of course, for your leadership, but even more so for your friendship which I'm looking forward to doubling down on as you work together in all of your many roles.

When I -- I just want to say a few other words, too. Look, when I joined Recursion 18 months ago, I was drawn to the bold ambition, but an ambition in action. And what an 18 months it's been. Step 1 was the special combination with [ Excientia ] to really truly build that end-to-end tech stack, as we like to call it, engine and also the data generation. We've talked a lot about models, but proprietary high-quality data is critical and a critical moat to what we do.

Building out clin tech platform, we're in the clinic. This is going to be a critical part of what we do and also sharpening our portfolio, advancing multiple programs internally with our partners. Chris will share a little bit more of some of the latest updates there from Roche and doing so with a discipline and urgency that I think patients will be proud of. And this is a pivotal chapter for Recursion, one that will require bold focus. The boldness will never go away, but also navigating the complexity that, quite frankly, is drug discovery and development, having seen that for many years and the relentless sense of urgency and good capital stewardship that's going to be critical for us to fully realize our mission.

My focus, and I'll listen more and think more over the coming weeks, but really it's going to be translating these platform insights into repeatable clinical proof, whether it's through our wholly owned programs or with partners, scaling the platform that we have, where we have a clear, clear, clear advantage and building a company that delivers sustainable value.

The foundation is strong. The vision is clear. The opportunity ahead is extraordinary, and I couldn't be excited. The work will be hard. I am clear-eyed about that. And there will be bumps in the road, but it should be because it's deeply worth it because we're doing something that there's no blueprint for. It hasn't been done before. It's also deeply meaningful for the patients that we serve and for the future that we're building together. So I'm so excited. I couldn't be prouder of the next phase. Looking forward to partnering with Chris and the rest of the broader team, exceptional team, if I say so, to define that next phase and the next chapter for Recursion.

Thanks, Najat. And with that out of the way, I think we should get back to work. And one of the things that I'm extremely excited about is that Recursion has a tremendous amount of potential catalysts coming in the next 18 months. From our pipeline to our partnerships to advancements of our platform, we are going to continue to deliver exciting updates, I hope, on a really regular cadence to all of our shareholders, all of our stakeholders and ultimately, advancements and milestones that I think take us on a path towards really affecting patient lives.

And we're going to do all of that with a pro forma cash balance of almost $800 million as of a few weeks ago. And what I'm excited about is that gives us runway through the end of 2027 without any additional financing. This means we're in an incredibly robust position from a balance sheet perspective to deliver across all of these catalysts.

Now let me talk a little bit about something that was really exciting that we got to share last week, which is how our platform is fueling all of our partners, but in particular, our Roche and Genentech colleagues. We shared last week that we had earned a $30 million milestone payment. This is our second such milestone from Roche and Genentech for the delivery of a whole genome neuro map. And I want to point out that this brings our total cash inflows from partnerships to more than $0.5 billion. And I think that is a milestone that few pre-commercial biotech companies ever achieved and one that I think is a leading indicator of the kind of value that we are going to continue to deliver here at Recursion.

So let's talk a little bit about the Roche and Genentech collaboration. This is a collaboration that's a decade-long focus in neuroscience and GI oncology. We've already been able to deliver 4 whole genome phenom maps in our GI oncology space, generating over 100 billion GI oncology relevant cells. And we already have a program that's been optioned out of that particular set of maps and heading toward lead series, and I hope many more in the future as well. But what we announced last week was that in addition to the whole genome arrayed CRISPR knockout map of iPSC-derived neurons that we delivered last year, where we became, we believe, one of, if not the world's largest producers of these iPSC-derived neuronal cells. We have now delivered a second whole genome map this time in the incredibly challenging microglial cells, the immune cell of our brain. And what we are excited about is the number of incredibly novel potential targets that we have identified in both of these pheno map that we believe have real potential not only to lead to novel target options in the future with our colleagues at Roche and Genentech, but I hope really meaningful medicines for patients in the field of neuroscience where all of us agree, there's a lot, a lot of work to do.

So I want to talk a little bit more about this work. And I want to remind everybody what do I mean when I talk about a map of biology. We have knocked out nearly every gene in the genome in these microglial cells. And we have leveraged machine learning and AI techniques to turn images of these cells into functional maps of the relationships between every single gene. And these digital maps allow us to move from this empirical sort of one-in-a-time approach into really a search function where today, our colleagues at Roche and Genentech, our team can just type in any gene in the microglial map or the neuronal -- or the iPSC-derived neuronal map, and they can see the relationships across the rest of the genome.

This gives us incredible insight, novel pathways, novel targets. And it's extraordinarily significant, I think, for our teams to now be able to do this not just in the neuron, but in the immune cell of the brain. So really, really excited about this. I know our colleagues at Roche and Genentech are too. And to give you a sense of the road that we've taken to get here, this was something that I think many people did not think would be possible.

First, our team had to generate more than 100 billion microglial cells. Then we had to figure out how to get more than 100,000 different sgRNAs into these cells, so we could knock out 17,000-plus genes with multiple guides per gene and we were able to do this generating nearly 50 million microglial cell images. And we used our supercomputer, BioHive-2, still, we believe, the fastest supercomputer wholly owned by any biopharma company, at least for a couple of more weeks or months until [ Lilly ] potentially overtakes us to generate this first-of-its-kind microglial map. And from this map, we've already started mining for novel biological insights, and we have the team and the equipment and the expertise to validate those insights to deliver programs to our colleagues at Roche and Genentech. And our hope is that, that will lead to potential new therapeutic approaches.

So I'm so proud of the team for all the work they've done. I'm so proud of this particular map because I think it has extraordinary potential in the field of neuroscience. With that, I'm going to turn it over to Najat to talk a little bit about how our platform is fueling our pipeline.

Thank you, Chris. And just the example that you mentioned in terms of microglia is such a phenomenal example of how our platform and the combination of the wet lab and the depth of the wet lab approaches that we have at Recursion with, of course, our dry lab really creates something of unique value.

So if you go to the next slide, we shared this slide before. And I just wanted to spend a little bit of time on it today and double-click on a couple of areas. So first of all, this represents the heart of how Recursion operates. You hear a lot about the Recursion OS. And I love to kind of pull the hood and like really show what are the various components that we're focusing on here.

Step one, we're applying AI where it matters, where it can truly change either quality, speed or the impact of our decision-making. There are 3 specific modules here. The first is focused on deep biological understanding that's actually connected to patient outcomes. The second, how do we leverage AI to design better molecules that are more drug-like. And then the third is a ClinTech approach on picking the right patients as well as recruitment -- fast recruitment so we can get through our trials faster. All right. Having said that, I wanted to double-click on a couple of areas that we're really focusing on.

Next slide. Scientific agents. So we've heard a lot about agentic agents. One of the most exciting parts of the OS is how we are using scientific agents, AI systems that actively participate in the entirety of the scientific process. And these agents are helping us thinking about the data that Chris just mentioned, genomics, transcriptomics, real-world data with partners like Tempus and others public data sets like PubMed, [ JetMa ] and so much the list goes on and on. And we have early proof-of-concept agents that we're leveraging in order to really not just analyze and interpret the data real time, but to actually select the optimal tools, workflows, generate hypothesis and design new experiment.

This is going to supercharge our already extraordinary talent. The other reason I'd like to mention this is also around it captures the decision-making trail. That's really important. The way you get these agents to be highly effective is to actually understand the logic behind the recommendations and iterate in real time with our scientists and our clinicians. And that is something that we're doing. And having that inherent data, this platform helps us do it in action, not just theoretically.

I often get the question in terms of how do we drive economies of scale. We are a tech bio company. This is going to be one of the very important levers for us as to how do we do more given a lot of the insights that we're generating with what we have today. So just keep an eye on this, but I wanted to double-click on a really important area of progress for us that is truly applied to what we need to do to create programs that are differentiated.

The second area, if we go to the -- one more click, I will do it, is really around automated ADMET. Look, we talked a lot before around [ Bolts 2 ] and other programs that really help you understand binding affinity and so forth. But as we all know, in order to actually design programs, a critical element of it is to ensure that they are drug-like. So this is an area that is critical for us. And what we are doing, this is the automated platform that we have in Salt Lake City, combining high-throughput experimental automation with advanced machine learning. It's a fully automated closed-loop system that integrates ADMET property predictions directly into that middle module that we have, which is our AI-enabled precision design. So it does a couple of things.

Number one, we are generating proprietary data around ADMET, not just what's been published, but all of the successes and failures that we see as we are generating our own data. Failures are incredibly important to design better models. Second, the comprehensiveness of the ADMET properties over 50 or so is just a starting point, and that's only going to increase is important in order to ensure the algorithms are actually generalizable. And then the third, you've heard about models such as [ Mole GPS ], and there's new models that are coming up all the time. But these data feed directly into the model, which is actually iterating and helping our models to be retrained real time. So both examples of real-time iteration is critical for us to not just have a platform that's useful, but a platform that stays ahead of the curve and learns from both our successes and our mistakes.

So with that, I want to walk to the next part, which is how is the platform being leveraged to actually generate programs. And this actually slide came to my mind during the flight post popular demand from analyst questions and investor questions. Here's what you have. I'll just walk you through it. On rows, the 3 design components that I just mentioned, the biology, the chemistry and the clinical development that you saw on the last slide, everything from phenomics, transcriptomes, et cetera. And as I do the build for the columns, these are the various iterations generations of our platform. So just to give you a clear view on which programs using what component of our platform.

So it's important to note, as you'll see, the earliest programs built on our first-generation platform, which I will call V0.1, and we've shared that before, the programs that are now in the clinic, tangible proof that we are generating programs and also learning fast, the flywheel with every turn of the crank, we're learning really, really fast as to how to improve on what we're doing well and what we need to do better.

So the first one here, MEK1/2, more data coming in December. I'll share a little bit more of our plan there. This is a proof point around how are we leveraging genomics, an unbiased approach to drug discovery in order to really ascertain which compound, which mechanism, which we do not know going in, could actually help attenuate the hallmark of the disease, which is polyps, hundreds to thousands of polyps. More to come on that in a second.

Now often, I get the question that if this is part of platform 0.1 is that it? The story never ends there. As you will see with ClinTech, if you go to the third row, we are actually leveraging recruitment solutions as well as patient selection and stratification. So even programs that are in the early part of the platform, we're leveraging some of our recent components in CinTech to add more value creation.

The second iteration of our platform, I’ll call this is V1.0, we're beginning to combine genomics and biology as well as chemistry design as well. So examples such as RBM39, CDK7, et cetera, and of course, components of the CinTech platform. And then one more click. These are programs in discovery, which I hope to be able to -- we hope to be able to talk about soon that, as you can see, is incorporating all of the various components of the platform from discovery, biology, novel insights, chemistry and CinTech and not just for our wholly owned programs, but also for our partner programs. So more to come on that.

So where are we making progress? You've seen this slide before. I'd like to update it every time that we meet. So first of all, CDK7 combination cohorts have been initiated. I'll speak a little bit more in a moment in terms of some of the analysis and some of the data that we have from the monotherapy dose escalation. PI3K 1047R, the development candidate has been nominated, which was again one of our milestones for this year. And MEK1/2, we'll have a webinar in December in order to share some of the additional data from our 4 milligram cohort.

As you also heard from Chris, the $30 million option milestone received for the microglia map. This is in addition to the 6 Phenomaps and also some of the programs that we're generating and great traction across the other programs and partnerships, including Sanofi. I just want to pause and take a moment to say both of our internal and our partner programs are critical, critical to us delivering tangible proof as well as also learning fast to keep evolving our platform.

All right. So I'll take a moment to go through CDK7. So just a quick context, we've talked about this before. CDK7, really important master regulator, transcriptional kinase that has generated significant interest in oncology for some time, but has been plagued with historical challenges. And the main -- one of the main reasons for those challenges is the narrow therapeutic window and just the molecular properties that limited tolerability and efficacy. So what are we doing leveraging our platform that's different?

Number one, leveraging our platform, we set out to design a molecule that directly addresses one of these core limitations around therapeutic index, which is optimizing permeability and [ eflux ] so that we reduce the GI-related toxin variability that others have seen before. That's number one.

Number two, we are also leveraging preclinical data as well as multimodal real-world data, causal AI, all of those components of our platform in order to hone in on patients that might most benefit, how do we steer a product into the right patients, patient stratification. That's another area of differentiation that we're focused on. And of course, from preclinical models, we've seen tumor regression in both ovarian and breast cancer. We also shared some early data last year, early clinical data from last year that showed manageable safety profile as well as some partial response as well.

What is our next focus? So in terms of our next focus, next slide. We have completed our Phase 1 dose escalation. So MTD has been achieved in advanced solid tumors. I'll get to that in a second. In addition to that, concurrently, the team is also looking into alternate dosing schedule just to figure out ways to even further optimize the therapeutic index, especially for long-term dosing. The other areas of near-term focus for us, which is already well underway, is a Phase 2 dose expansion in the cohort that we had mentioned, the platinum-resistant ovarian cancer as well as combination, which is going to be key in this space with a couple of combination standard of care that you can see on the slide.

Recruitment ongoing across all of these cohorts. The other thing that's important to note is a lot of the trial design is focused on rapid and efficient go/no-go. And we'll share more in terms of the Phase 1 dose escalation data at a medical congress next year. And then in addition to that, we should have some of the ovarian combination data in 2027, safety, PK/PD, maybe early signs of efficacy, more to come in next year as we see the recruitment shaping up in terms of details on when in 2027. So stay tuned.

All right. So let's go into the monotherapy dose escalation. So a couple of things to note. As of September 29, which is the cutoff date, we have 29 heavily pretreated patients with advanced solid tumors that have received 617 across 6 dose levels. Just as context, these patients represent a rather challenging population, most with multiple prior lines of therapy and limited standard options. We have now established a 10-milligram once-daily MTD, maximum tolerated dose with a manageable safety profile, we'll talk about in a second and also preliminary antitumor activity that's consistent with what we shared in the 2024 update.

The most common dose-limiting toxicities were nausea, which is to be expected and some thrombocytopenia, which are both on-target effects for this target class. If you go into safety. Look, the safety data is consistent with what we saw last year. First of all, we have about 30% of patients that experienced Grade 3 treatment-related adverse events, and the majority were low grade 1 or 2. There were no grade 4 or 5 treatment-related events and only 2 patients, about 7% discontinued due to an AE. One thing I want to note here, and again, this is early, we're learning more, of course, the GI-related toxicities that we have seen, diarrhea, nausea, vomiting were relatively manageable and in line with class expectations. And to just put that in context a little bit more. Okay. To put that in context a little bit more, in terms of diarrhea, we saw about 69%, nausea 41%, vomiting 28%. Of course, looking at some of our peers also in the space, the numbers for diarrhea are about 82%, 77% for nausea and vomiting for 80%. So it's trending in the right direction, but of course, much more work to be done, but trending to be slightly lower than what we have seen in other prior published data.

All right. In terms of efficacy, first of all, on the left-hand side, this highlights the PK profile. highly selective inhibitor potent and also flexibility in terms of how we can dose it given the short half-life -- the relatively short half-life of around 5 hours. One thing that's important to note, so going back to the established MTD, which is 10-milligram once daily dose, the exposures exceed, as you can see, the CDK7 IC80 while remaining below CDK2 IC80, supporting the selective inhibition that we wanted to see.

In preclinical models -- what does that mean? In preclinical models, 10-milligram equivalent QD showed robust tumor regressions with about 2 to 4 hours to target coverage. And the early PD data that we have indicates that this is about 80% to 90% of transient [ POLR 2A ] engagement, which is one of the key PD markers that we are tracking in this space, again, consistent with that hypothesis.

So from all of the data that we've seen so far, 10-milligram QD is pharmacologically active dose. In addition to that, we're also looking into alternate intermittent schedules such as second day on off to further maximize the dose intensity while maintaining tolerability for long-term dosing. On the right-hand side, you're also seeing some of the early clinical translation. Look at 10 milligram is where we see stable dose and also the patient that had the PR. Of course, in this patient cohort, monotherapy is not an area that we were expecting outcomes and which is consistent with what we've seen with other CDK inhibitors, which is why our combination is going to be incredibly important. So stay tuned, more to come on that.

And last thing, speaking about the combination, I've shared this -- we've shared this data before, but ovarian cancer is the current area of focus, which is different from where others have gone, which has been much more primarily in breast cancer or a broad basket of solid tumors. pulling together everything that we have seen in cell lines on the left, in terms of ovarian cancer models, the sensitivity to CDK inhibition, combining that with what we have seen in vivo at 10-milligram dose, we saw complete tumor regression by day 27 as well as on the right, leveraging our patient level data from over 30,000 ovarian cancer samples, integrating DNA, RNA and clinical outcomes to really showcase that CDK7 is a likely driver of poor survival and some of the work we've done with our causal and friends and AI works.

So -- but again, the proof is always in the pudding. So much more to be done and expect more on the full data set I just mentioned for the monotherapy dose escalation next year at a medical congress as well as combination data in 2027. We'll give you narrow guidance or more specific guidance as we get full flow into our recruitment.

All right. And just wanted to heads up on REC-4881, which is our other program that has an important readout coming out next month. So just as context, high unmet need, 50,000 patients diagnosed across U.S., EU5, rare inherited disorder, APC loss of function. And standard of care is quite challenging. Surgery is a standard of care, [ colectomies ], et cetera, and no approved therapies to date. We also have orphan drug designation for this compound.

Just a recap of some of the earlier data that we've seen in May that was shared in DDW, 43% median reduction in total poly burden, that is the hallmark standard of care today, off-label use of celecoxib is usually 20% or so or 25%. But again, there is a range from 30% to almost to actually 83% in the small cohort that we had seen in May. We've seen about 6 patients. We expect that to be double or close to double by the end of this year. And what we're looking for, again, as I mentioned before, is to see if these trends will hold and a significant benefit over the 20% that has been seen so far.

In terms of what we're seeing from a treatment-related AEs, 19% grade 3, majority is rash and the prophylactic approach has really made it much more manageable and cardiac tox more grade 2 so far.

So again, December, we'll share more information in the coming weeks, exactly when in December. We'll have the Phase 1b/2 update. This is going to be an important update, as I mentioned, and then we'll also discuss some of the next steps for the program. If the trends hold, one of the core next steps will be to actually have discussions with regulators on a pivotal study. And I just want to say this is one of those, as I like to call green shoots in terms of leveraging our platform to see color burden reduction, both in vivo and then also starting to see in patients. But again, small data set, more to come. We're looking forward to the data cut in December. With that, I'm going to hand it over to Ben for our financial update.

Terrific. Thank you, Najat. Another thing that we are very excited about is going into that FAP data as well as the milestones in 2026 in a really strong financial position. So over the course of the year, you've seen us do a number of things. In May, we laid out a strategic plan that allowed us not only to hit on multiple high-value milestones, but also reduce our expense base by 35% from 2024. This was a really important step in us trying to put that discipline in place that Najat was talking about earlier. And then you started to see us hit those milestones. We've brought in almost $40 million from our partnership inflows over the course of this year so far, and we expect more of that to come in the future.

So with today, our announcement of having $785 million of cash in the bank as of October 9 is a really strong step in creating that foundation so we can look forward into the future milestones and say, we don't have that financing need to be able to achieve our near-term milestones and really deliver a lot of value to shareholders. And so when we look forward we've looked at how we can bring that financing together in a way that was going to minimize our dilution to all of the shareholders as well as really continue to allow us to focus on the business and move it forward.

We are also reaffirming our guidance for 2025 on an expense base of less than $450 million. That's excluding all of the partnership inflows. In 2026, we're also reaffirming less than $390 million over that time period. One note on that 35% expense reduction, that actually equates to over $200 million in expenses coming off of that 2024 base. And we've done that by really focusing in on what is going to be the aspects of the business that deliver the highest value and efficiently bringing that together. We will continue to look at our expense base. We are completed with all of the restructuring that was associated with the transaction, but we will continue to look at our operations and think about how can we do this better? How can we do this more efficiently? How can we get more out of every dollar that we spend and really focus in on the high-value partnerships -- or high-value projects.

From a partnership perspective, the $30 million in milestone from microglia that we achieved with Roche and Genentech was not included in the $785 million in cash. Importantly, we do not give guidance on revenue, but I know a lot of the services track it. So we just want to be really clear. First of all, we don't consider the lumpiness in our revenue to be any indication of our business. But what we do know is the timing of our milestones can impact how we recognize some of that revenue. For example, last year, we had a milestone associated with the neuronal map. And so we had a larger piece of revenue that was recognized in that quarter last year, in the third quarter of last year.

The Roche microglia milestone is going to be recognized partially in the fourth quarter, and it is important that's partial, not full. And so you would see some of that lumpiness come into our fourth quarter numbers there as well. Really importantly, the nearly $40 million that we've recognized this year from our partnership inflows actually brings us over $500 million in partnership inflows over the course of the company. That shows just what an important piece of nondilutive capital that part of the business and platform has become to the overall company.

We are also maintaining our guidance of over $100 million in partnership inflows by year-end 2026 that we laid out in May of this year. So we're making great progress along that and expect to continue to see a lot of that come through in the next year as well. And with that, I will turn it back over to Najat to talk about some of the upcoming milestones.

Great. Thank you so much, Ben. And we wanted to just capture both looking at this year as well as what's coming up next year. For this year, one big thing that as I look at the slide, which is missing is really the successful integration of the 2 companies coming together. It's been an incredible amount of work. And also the financial discipline that we've gone through internally to really extend our runway so that we can see through a lot of these catalysts.

So look, on the internal pipeline highlights, CDK7, I just mentioned the monotherapy update and the combo that's initiated. REC-4881, the Phase 2 initial update, potential first POC for our platform, [ MALS-1 ] monotherapy initiation and the PI3KDC nomination. I will also talk a little bit about our platform and then go to our partnership.

On the platform front, a ton of work, like 3 words on a slide, integrated design platform, but actually, I should say integrated our end-to-end platform, the amount of migration work, I have seen this across other companies before with the speed with which that was done and the utility, the fact that we had no slowdown in our productivity of our platform speaks to our fantastic engineering data science teams that exist.

And then also the work with MIT and NVIDIA on both 2, but much more that's going on in-house, which I would be a pleasure to share next time, especially with our Frontier team, which is our 0 to 1 cutting-edge AI team, really, really proud of the work that they're doing, Valence and Inception Labs. And then ClinTech, this has been a core focus for us in the last several months to really build out the tech stack end-to-end also into clinical development, crucial pillar as we execute on these programs.

And on the partnership highlight, Chris and Ben mentioned, of course, the Roche $30 million from microglia, but the real work that we're doubling down on is both of the maps in neuroscience and the additional maps in GI onc and really turning and Chris showed this really beautifully in his slides, turning those into insights with a deep functional validation with our partners to then become programs. That is the core focus for us translating that value. And of course, progress with Sanofi, 4 out of 4 milestones so far in immunology and oncology and much more work ongoing. We're so honored to be able to work with our partners, Roche and Genentech, Sanofi, Merck KGaA and Bayer. We learned so much from each and every one of them. So thank you for the partnership.

And then looking ahead, stay tuned for the REC-4881. We'll share more data in terms of our Phase 2 in December. In addition to that, for next year, RBM39, this is our first-in-class compound from the phenomics platform. It's also leveraging, of course, a lot of our clin tech approaches today. We should have some early safety and PK data from our monotherapy trial. And then in addition to that, ENPP1i, PI3K, both in GLP tox right now and pending that data, Phase 1 initiation, the team is all set up, pending what the data looks like. And from a partnership perspective, as I mentioned, deep focus on turning Maps into insights into programs. That's a huge focus for us in addition to some of the programs that we're already working on and additional Maps as well. So lots to do. There's never a dull moment at Recursion. I assure you, loving the pace, and we'll keep you all posted. With that, thank you so much. I'm going to hand it back to Chris for our Q&A.

Thanks, Najat. All right. Let's go. We've got Sean from Morgan Stanley. Ben, this one goes to you. Can you review expectations for cash burn through 2026 and how this works with runway expectations through '27 without additional financing? I know you just hit that, but maybe break it down for everybody. And then also, do you plan to use any additional ATM financing?

Yes, of course. So I think there's a couple of things that we looked at over the course of the quarter. So one, the most important thing for a growth company is delivering on high-value milestones. And so our role as a management team is to make sure that we have the resources to be able to hit those milestones. So we looked at all of the things that we've talked about before. So how do we do the right expense control, how do we prioritize the right programs? And then how do we make sure that we've got the right cash balance in place to be able to reach those milestones and really focus on them.

So what we decided to do was fully utilize the ATM over the course of the quarter, the remaining balance on the ATM. So that is now closed. We have not opened up a new ATM. And what that allows us to do is really go in and put in a cash balance that without any additional financing allows us to get to the year-end 2027 and achieve those milestones that Najat was just talking about as well as many others.

Just because I know the financing has been a critical question for a lot of shareholders. We really looked at 2 different aspects for that ATM utilization. One was if we look at the biotech financing market, there's a couple of things that we saw very clearly. One, there is increasing volatility. There are fewer open windows. There's a much shorter period of hold that we've seen from a lot of the investor base. And also the discounts have been increasing as well. And so we looked at the ATM as a very attractive cost of capital that would put us in a position to be able to execute on the plan going forward.

The other part was just there was so much focus that was becoming a part of the financing and the cash balance. It was actually starting to overshadow some aspects of the story. And so with such important data like FAP and some of the other milestones that are upcoming, we wanted people to really be able to focus in on the fundamentals of those events rather than needing to worry about are those events just going to lead to another financing or other aspects like that. And so we hope that this allows investors to focus. It also gives us a lot of ability to really focus on delivering those milestones over the coming months.

Thanks, Ben. So financing overhang has been struck. Alec from B of A, one question on platform utilization. It looks like older programs use parts of Recursion's capabilities in their development with Platform 2.0 assets, leveraging the full stack. Najat, this one is coming to you. How do you see this feeding into the quality or uniqueness of the newer assets? And anything to be read into for the current pipeline like 481, where it only benefited from phenomics in the version 0.1 of the platform.

Yes. No, thank you, Alec. And by the way, you were one of the voices that inspired us to create that slide. So thank you for always sharing your feedback. So look, 4881 phenomics, our platform today, even if we're making it multimodal, still leverages genomics a lot, right? And we are very, very excited in terms of some of the data we've seen to date and more to come later this year. But in terms of -- with every crank, like some of the clinical stage programs that we have come from the earlier stages of the platform in discovery, later stages of our platform. I think that's just the true iterative nature of drug discovery and development. And the improvement of our compounds doesn't just stop in discovery. It's also in development. So you see some of the innovative approaches that we're also taking in development for FAP, CDK7, et cetera.

So we look at it more holistically. But look, with every crank, the platform gets better, and that's just what it is for us, and we're learning fast, and we want to execute and iterate as quickly as possible to get them into the clinic as well.

Thanks, Najat. I'm going to bring this next one over to you as well. From Gil at Needham, Sean at MS and Manny from Leerink on the partnership side. Is Recursion looking to maintain current biopharma partnerships or expand to new partnerships in the near to midterm? And what are some of the milestones that we should be focused on?

Great question. Our partnerships that we have are deep, highly collaborative and transformational. We are very, very excited about the partnerships that we have. And I mentioned some of the milestones that are critical for us maps to programs, for instance, in our Roche Genentech partnerships. And for Sanofi, we're really making progress on the various programs that we have in immunology and oncology.

We are always having discussions in terms of potential new partnerships. We are being incredibly choiceful. So that is an area that we'll always be open to, but areas we can drive incredible value as well as our partner. It has to be a win-win. So the answer is yes, the door is always open, but we also -- we tend to curate a set of partners that we can really show tangible value with.

Thanks, Najat. Next one I'm going to take. This is from Guy and Alec. Would be interested to hear your thoughts on the evolving AI drug development landscape, especially with companies like Lilly throwing their hat in the ring and also partnering with NVIDIA.

So look, I think this is extraordinarily exciting. This is a sign that when we said a couple of years ago, we look like what the future of biopharma will look like that we were right. companies are starting to embrace massively scaled compute. They're starting to embrace AI. And so this tech bio sector is really, I think, just a harbinger of what the future of biopharma will become. And so this is super exciting to me. I'm so glad that Lilly is making this visionary investment. I'm so glad to see them partnered with NVIDIA. And I look forward to working alongside many companies in the future that come to the space. We want to move the entire field forward ultimately to bring medicines to patients. So really, really excited by that advance.

And then final question, I think, fittingly, over to Najat from Dennis at Jefferies. Congrats on the new role, I mean, effective January 1, we've still got a few weeks. Curious what philosophy you're bringing into the seat as CEO and if there are any near or medium-term priorities that are top of mind.

Thank you, Dennis. Look, my priority is going to be the core priority is going to be how do we translate the data, the compute, our amazing people, our platform, to tangible proof points that matter, whether it's our own pipeline, whether it's with our partners, that is the core element that matters the most. As Chris was saying, look, I've been in big pharma before. Now I'm in tech, bio, biotech, AI inspired, whatever term you want to use. At the end of the day, it's about making differentiated programs and then eventually medicines for patients that matter. That is the core focus for me. It starts with that, it ends with that.

We have -- this is a tough journey, 90% failure rate. I am aware, clear eyed of how tough it is across industry. And we're doing something in a way that's never been done before. That's going to be my core focus. The second is going to be really investing where we have a unique ability to win. That's our platform, that's our programs. We're going to make data-driven, and you've seen me do that before, go/no-go decision on our programs. That's why when you say all of the programs we're doing concurrent targeted, efficient approaches so that we can get to a rapid go no-go because unlike a lot of other companies, there are multiple other programs that we're bringing from discovery into the clinic. We need to be choiceful in terms of where we go.

And then the third is discipline and execution and good capital stewardship. You heard Ben talk through, we are grateful for the capital that we have. And my intent is to use every single dollar for what will truly create value for our shareholders and our patients. So those are some of my areas of focus. I'm sure I'll think more on it over the holidays. And with Chris' counsel, I am so excited to continue partnering with Chris. It's been such a great journey and so much more to come together.

Thanks, Najat. It's been an amazing first 12 years, and I'm looking forward to the next 12. Thank you, everybody, for joining us. I hope you have a great day.