Lattice Semiconductor Corporation Aktie

Great. Good afternoon, everybody. My name is John Vinh. I cover semis here at KeyBanc Capital Markets.

We're pleased to have Lattice Semi here with us. We have Lorenzo Flores, CFO.

Welcome, Lorenzo. Great to see you.

Thanks, John. Great to have you. Always a great event.

So let's start talking about kind of your data center growth, right? In the quarter you just reported, comms and compute grew 80% -- over 80%, which would obviously suggest that AI data center is growing faster than that. Can you talk about how much of this growth has been driven by general purpose server versus AI?

Yes. So I'm going to start with my best understanding, but I think I'll provide a little bit more color around the overall growth dynamic, which might actually be more helpful, right?

So at the beginning of the year, I think everybody in the industry was looking at industry expectations, from about 15 million units for total servers, something about a 20% AI mix, and then it grew up to 16 million. And now like the latest past 3 weeks, we've seen the estimate for overall servers about 20 million units and with a more -- higher AI mix to them.

And I would say that in parallel, as we've gone through the year, we've continued to see our demand going up and up and up. And at first, we were a little bit surprised and tried to -- went back and tried to understand it. And what's really going on is, sure, AI servers are growing, but also just the more standard foundational, CPU-based server infrastructure that is enabling this agentic workflow being broadly deployed.

And at the highest level, both phenomenon, growth in the AI servers and in the more general purpose servers is very beneficial to us. We've historically been really strong and had a pervasive footprint in standard servers. And as you know, and we've talked about many times, we have a really great opportunity in AI servers as you build out the racks with many, many more attach points for us, up to 100 attach points or even greater than that, in an AI server rack because of the complexity.

So when we looked at what's going on and, honestly, in real time, it's a challenge to parse those different drivers because our ship timing and the end customer ship timing don't necessarily align. But what we saw is a really strong unit volume growth in our products going to servers. That is accompanied by -- or that's driven by our attach rate is growing. As we've talked about, it was, historically, let's say, 5, 4 years ago, it was 1.x, then it went 2.x, 3.x and it's now even higher.

And we've been providing higher-value FPGAs into those opportunities. We've gone from our MachXO2, XO3, XO3D, XO5, and some Nexus parts as we are now seeing an expansion in our ASP from $3 to $4 and so on. And so you have this multiplier effect for us. Greater unit volume growth driven by higher attach rates as the complexity of the servers in both AI and general purpose applications grow, and increasing ASP.

So that's really setting -- we've seen great growth, it's setting the foundation for future growth. So I know I didn't only answer your question, but I thought you'd get there anyway.

No, that's super helpful. Do you have a sense of what your mix of general purpose versus AI is right now?

So what we've been looking at is our overall AI penetration, and some are on the data path somewhere in the server domain. We don't do very much direct AI compute. So that's grown into the high 20% of our overall business, which means a greater percent of the compute business.

But again, we know where our design wins are. We know where customers are shipping to, how they're actually using the devices. And not yet fully transparent to us, but it's, in a way, it's kind of like we'd like to know. More importantly, we want to ship all the units to them and drive the growth.

Great. Maybe just to double-click on just your attach rates, right? I think when you're at 1x to 2x, right, a lot of your core functionality, my understanding, was driven by Root of Trust and maybe I/O expansion as well. So as now you've gone to 3x and now you're saying it's even higher than 3x, what are some of the functions that are driving the attach rates to high levels right now beyond just Root of Trust?

Right. So I think -- actually, back in -- when it was a lower attach rate, it was primarily bridging an I/O expansion type of applications. And what we see then, we saw the security capabilities, including Root of Trust later on. And then now we are beginning to see an expansion into manageability applications like power and cooling. And then the basis for our acquisition of AMI builds on that growing functionality that we've seen in the FPGA business and the broader applications that's available to us, and their capabilities, which really augment us.

So there -- we used to say we're first-on/last-off on a machine, and they're first-on/last-off right with us. And then they layered on top of their boot capability this management capability, and that meshes really, really well with how we see our FPGA application set growing. So that's how the attach rate grows and then that's the growth path that AMI had intrinsically. And then for both of us together, that's this 1 plus 1 equals 3 phenomenon we're trying to drive with bringing more sophisticated solutions to the customers. And right now, we think it's very acutely needed in the data center.

Great. I do want to touch on AMI in a little bit, but just a couple more follow-ups on data center. Your levers on data center is obviously market TAM, ASPs and attach. So if you think about your data center business, I don't know, 3 to 5 years out, what's going to be a bigger driver of growth? Is it going to be ASPs or attach rates, do you think?

Right. So I'll answer that in 3 parts. One is FPGA part, one is AMI part, and I know you said you're going to get to that, but then I think the solutions together. So for the FPGA part, if we were thinking only about silicon devices, unit volume growth will be significant and will continue to be the primary driver. And we will be able to have some expansion. And what's going to be driving that is the number of attach points in the server for different functionalities.

And we'd already been working on things, and we've talked in the past before and I'm sure some of you have heard this, things like leak detection, which 18 months ago nobody thought about, 12 months ago, people were just kind of wondering about it, now is a really critical application. So an opportunity didn't exist before. We're really well positioned to drive that. The same thing with some of the more advanced management capabilities that companies run. So we would see continued attach rate growth on top of any industry unit volume growth, right? And our ASPs will continue to progress. But I think the unit volume will be the bigger driver in the FPGA business.

In the AMI business, their footprint will also expand, but also their capability -- their management capability is growing and is -- they have the opportunity to, again, expand the value. So that might be more of a driver than footprint. But there's opportunities they have, as they integrate open source capabilities with their inherent capabilities, to also grow unit volume. So there's that.

The third piece of it is, it's very important because it's core to the strategic value of the AMI acquisition, which is the solutions that we'll integrate together. And that is a unit volume play, but also a really high-value application we can bring to our end customers. So you'd see a much stronger revenue growth per unit from the solutions that we provide than from an FPGA or a manageability firmware kit by itself.

Great.

Is that greenfield or is that just a combination of the first 2?

Totally greenfield -- well, that -- there's a nuance in the answer to that question, which is the problem statement for the hyperscalers is the same. How they would solve it by using, say, an FPGA and some of AMI's firmware, could be the same. But integrating it and adding more features and functionality, that's greenfield.

So it's an upscale in value that we think we can capture. So yes. So the solution may look like the same, but I think the overall value will be much greater than the individual components.

I've got one more question for you on data center, is when you think about rack scale, I think the teardowns have shown that you've got an outsized number of FPGAs in a rack scale rack. I guess the question is, how are you thinking about this going forward? Is this content sustainable? I know at Xilinx, I know this is a very different company, there was this phenomenon, base stations where there was kind of an ASIC reversion. I know that you're more insulated from that for a number of reasons. But do you think it's sustainable? And then is there an opportunity for you to grow that content with newer generations of racks?

Right. And I agree with your premise, it is different than -- for those of you who haven't followed FPGAs forever, the history has been FPGAs get to a certain volume and the customer says "Okay, I'm going to design you out and put an ASIC in place." But I'll tell you, even in my prior company, where that did not happen was where the pace of change was such that you really couldn't afford to do that.

In our situation, given that we've taken -- deliberately taken this companionship strategy to the market where we are not going head-to-head for core processing capabilities in the data center, but we are surrounding that with capability that makes real economic sense to offload, right? You don't want to spend a 2-nanometer silicon, your silicon die budget, on I/O expansion. It makes way more sense to do it in a 65-nanometer FPGA.

And so by doing that, we, in a way, insulated ourselves, just as you said, from that phenomenon. And because you've also seen an acceleration in the design cycles for these racks, from 2 years to 18 months to 12 months, the need for flexibility is an inherent advantage for maintaining the FPGA footprint.

Then you layer on top of it just the performance characteristics of our particular product, which is very low power, right? If you add hundreds of devices to a rack and they're watts in consumption per, that's a problem. If you have milliwatts, right, it's actually pretty helpful to the overall solution, the power budget for the rack. So that helps us as well.

And then the last piece of the sustainability does again come back to being able to integrate functionality that we can bring from AMI into the overall solution and put it in the programmable. And then you do the updates and upgrades, taking advantage of the capability of the programmable FPGA, and you're in really pretty good shape to provide a flexible solution for the customer for longer periods of time.

Excuse me. You always bring me out to altitude.

Yes. Yes, we like altitude here. It's a thing. Okay, so maybe pivoting to AMI. I think you talked about that you've already engaged with 100 different types of customers, hyperscalers, OEMs, ODMs, other ecosystem partners. And it seems like feedback is pretty positive on the acquisition. Maybe what customer problems are generating the most amount of interest in the combined Lattice-AMI solution from these customers?

Yes. And it's the -- it will be very similar problems that the hyperscalers and data center operators would be complaining about to everyone, which is, how do I increase my uptime, that will refer back to inherent capabilities of the manageability that we have where you can look at what's going on in your system and take action to do it.

Security, a huge issue that we have to continuously advance our capabilities on and provide customers with solutions that are hardware and firmware-based, as threats keep expanding.

And then there is power, cooling, that you need to optimize how your rack consumes power. And can you keep it at the right temperature for optimal performance? And if something happens, how do you prevent catastrophic failure? You talk about leak detection often. If you can respond instantaneously and shut off the source of the leak and manage it, have a managed power down, your system is in much better shape.

And then back to the uptime. If you can bring up your racks all simultaneously versus a serial blade-by-blade boot, you're up and running much faster. So these are all capabilities we've been talking to the customers about in different forums.

And then finally, in the end, because of the demand as such, the hyperscalers are looking at trying to get their systems built, and they're going from -- they're going to a mode wanting to be able to use a heterogeneous source of supply, so different ODMs, different componentry, mixing old infrastructure with new infrastructure. And how do you manage that all together effectively? You need the manageability at the firmware layer to do that. And that's the very strong value proposition from AMI.

That's great. Physical AI seems to be kind of an emerging theme for you, but also many of your peers here at the conference. You seem pretty excited about it. What are the opportunities within physical AI where you're the most excited? And can you give us a sense of when we could see kind of meaningful revenues from this category?

Right. I think there's a picture we show sometimes, and some people get it, some people don't. But one of the really, really cool things about the way we approach the market is we play to our strengths that we're seeing manifest in the data center right now in the same way in the physical AI world, which is, what is our great strength?

It's like, well, we can bring in a bunch of different information, as I said earlier, about bridging, and aggregate data from multiple sensors, preprocess it, send it to the core processor, get some action back, and then drive a reaction through a physical system. That's the same thing like in a leak detection as it would be in a robot doing something.

And so what we see is the ability to apply this companionship approach there. We're very strong in sensor fusion and aggregating different forms of input, particular strength in vision. So multiple cameras coming into one stream with some preprocessing, and that's part of the Holoscan platform from NVIDIA that we're on. And then taking -- going out and taking action.

So we think we have a footprint in almost all of the robotics companies. And we are expanding that footprint, analogous to what we did in the data center, with continued access to different problems they're trying to solve. I will say that a lot of these companies are not as advanced as the more old-line, traditional industrial automation companies in terms of their use of FPGAs, right? As you know from a long time, FPGA designers are not all that common in the industry. So we're doing things to enable a more software-like programming environment for these guys using FPGAs. And so we're accelerating our momentum there.

And so then what you see then is, okay, so when does this volume take off? We're not sure. We'll probably start seeing it at the end of next year. But we're going to have multiple sockets per instance and we are going then to be able to capitalize similarly as we are on the data center-AI trend.

I would say that as a backdrop to that, our core industrial business, which is mostly traditional factory automation stuff, is actually doing well. We're starting to see that recover. I think you guys might have seen the PMI across the globe is actually doing pretty well, except maybe in Japan. But we're not exposed to auto that much anymore and we're still seeing year-on-year, let's call it, 20% growth in industrial from last year. And we think that that will grow again next year. So not sequential, mind you, but if you just look at similar periods last year, we're seeing consistent growth.

That's great. On the industrial front and maybe just more broadly, can you talk about what your lead times are doing right now, and talk about what your visibility is looking right now? It's probably getting much better than it was a couple of quarters ago.

It is. Do you want me to get into the supply side too?

Yes.

Because our visibility is increasing. We have a phenomenon where our book has continued to grow in total dollar value and then expand out in time. And we view that as very positive because we think we're getting a realistic view of when customers really want their parts. And we always get this double booking question. I think it's a valid concern. We keep watching out for it and we try to manage that.

But we have better visibility because our book is expanding, and we have really, really tightened down the channel inventory, so we can see directly what end customers are really after. So having done that, let me turn to the supply side, and earlier in this year, as we started to see this increased demand, our lead times did increase. In fact, in some cases, got over a year. But we've been working with the supply side and managing that back down to the -- in the 40s.

And so what we're working through right now is the pinch point is assembly, test or packaging. Our fab side is -- we're pretty comfortable with. We use older nodes, not as tight as some of the newer nodes. And then we've invested a lot in, you'd see this on our CapEx, in tester capacity for the final test and ship out to the customer. So yes, we're managing through the assembly test pinch point right now. And people, like in the factories, are people-managing this directly. And we're expanding our supply chain.

The way we think about it is this is not temporary. We had about a $500 million a year business last year. We will end the year approximately double that in the FPGA side, and we expect to continue to grow next year. So that means our supply chain has to be 2x plus more robust. And so that's what we're driving to.

And your visibility?

Our visibility is actually quite good, in terms of both the demand side and the supply side, right? So we've been able to do a much more rigorous mapping than was necessary in the past. We've had to.

I think Ford said that you have visibility until the end of '27 at this point?

Yes, absolutely. We've booked out, I'm going to be very careful in my words, because when I say booked out to, meaning we have bookings out through Q4 of next year. We are not booked out in the sense that we can't support more demand. We can support more demand.

Got it.

Yes. This is continue -- to support this growth trajectory.

Sounds good. Thank you so much, Lorenzo. We're out of time.

Anything else, John? Yes.

We're good.

Thanks for having me. Appreciate it. Thanks, everybody.

Greetings, and welcome to Lattice Semiconductor Second Quarter 2026 Earnings Call. As a reminder, this conference is being recorded.

It is now my pleasure to introduce your host, Rick Muscha, Vice President of Investor Relations. Thank you, Rick. You may begin.

Thank you, operator, and good afternoon, everyone. With me today are Fouad Tamer, Lattice's CEO; and Lorenzo Flores, Lattice's CFO. We will provide a financial and business review of the second quarter of 2026 and the outlook for the third quarter of 2026, followed by a brief overview of AMI and its business model. If you have not yet obtained a copy of our earnings press release, it can be found at our company website in the Investor Relations section at latticesemi.com.

I would like to remind everyone that during our conference call today, we may make projections or other forward-looking statements regarding future events or the future financial performance of the company. We wish to caution you that such statements are predictions based on information that is currently available and that actual results may differ materially. We refer you to the documents that the company files with the SEC, including our 10-Ks, 10-Qs and 8-Ks. These documents contain and identify important risk factors that could cause the actual results to differ materially from those contained in our projections or forward-looking statements. This call includes and constitutes the company's official guidance for the third quarter of 2026. If at any time after this call, we communicate any material changes to this guidance, we intend that such updates will be done using a public forum such as a press release or publicly announced conference call.

We refer primarily to non-GAAP financial measures during this call. By disclosing certain non-GAAP information, management intends to provide investors with additional information to permit further analysis of the company's performance and underlying trends. For historical periods, we provided reconciliations of these non-GAAP financial measures to GAAP financial measures that can be found on the Investor Relations section of our website at latticesemi.com.

With that, I'll turn the call over to our CEO, Fouad Tamer.

Thank you, Rick, and welcome, everyone, to our Second Quarter Earnings Call. Lattice delivered exceptional financial results this quarter, reflecting a healthy market environment, compelling catalysts and our own strong execution. We have a focused and consistent strategy to create durable value by growing faster than the markets we serve. We do this by expanding into new applications, serving leadership customers, delivering differentiated innovation and driving sustainable shareholder value as a result. We are executing against each of these strategic elements and the results are increasingly visible across the business.

Following last week's close of the AMI acquisition, Lattice is now positioned to deliver even greater value to customers and our shareholders. You will hear more about this acquisition shortly. First, let me update you on our second quarter results and outlook.

Revenue for the second quarter was a record $201 million, representing 18% sequential growth and 62% year-over-year growth with strength across all our end markets. Our Compute and Communications end market reached another record revenue level, growing 18% sequentially and 83% year-over-year, driven by continued momentum in data center AI applications. Demand for Lattice solutions continue to be fueled by increasing CapEx, increasing AI content per server, expanding FPGA attach rates, rising security requirements and the shift towards more complex disaggregated architectures.

We also saw a continued recovery in our Industrial and Embedded end market with revenue up 17% sequentially and 36% year-over-year. We see momentum building across a diverse set of end markets, including industrial automation, aerospace and defense, medical, robotics and other emerging physical AI applications. These applications increasingly require our Lattice differentiation in low power, small form factor, low latency and secure processing. With channel inventory at healthy levels and multiple new design wins beginning to ramp, we expect Industrial and Embedded to remain an important contributor to growth through the balance of 2026 and beyond.

As we had anticipated, profitability continued to grow significantly faster than revenue with second quarter non-GAAP EPS of $0.53, representing growth of more than 120% year-over-year. This outstanding performance underscores the operating leverage we have built into the model. We expect this momentum to continue based on demand trends building across our major end markets. This is evidenced by accelerated bookings and our backlog extending well into 2027. At the same time, design win momentum remains healthy across our FPGA portfolio and end markets. As demand continues to strengthen, we remain focused on working closely with our supply chain partners to ensure that we can support our customers' growth plans. Taken together, we expect these trends to support a sustained multiyear growth outlook.

Turning now to AMI. We are pleased to announce the successful close of the acquisition last week. It brings together Lattice's leadership in low-power programmable FPGA's with AMI's industry-leading firmware and infrastructure manageability portfolio, and it positions us to create the industry's most complete, secure management and control platform for AI data center infrastructure. We expect that this combination can double our addressable market.

In the 3 months since the announcement of the transaction, we have engaged with about 100 hyperscalers, OEMs, ODMs and ecosystem partners, including many at Computex in Taiwan. And uniformly, they have all given us unequivocal positive feedback about the time to market and value that our joint solutions can provide. This reinforces our confidence that the acquisition can provide long-term value to our shareholders.

AMI is a highly attractive business with strong profitability and durable recurring revenue characteristics. As we exit 2026, we expect the business to be operating at a revenue run rate of more than $200 million with gross margins above 75% and EBITDA margins above 40%. These metrics underscore the strength of AMI's market position, the depth of its customer relationships and the highly efficient operating model the company has built over many years.

The AMI business is well balanced with approximately 60% of revenue generated from its boot firmware franchise and 40% from its infrastructure manageability solutions. Revenue includes firmware licensing, royalties and platform enablement services. This model creates strong visibility, attractive lifetime economics and durable customer engagements across long product cycles. We see multiple avenues to grow AMI over time, winning more platforms, increasing content per platform and expanding further into AI infrastructure and embedded markets. This includes new trends such as rack scale architectures, secure boot, data center manageability and remote monitoring and control. AMI will continue to operate with the same open silicon-neutral approach that has earned the trust of customers and partners across the industry for decades. We are excited to welcome the AMI team to Lattice, and we have already hit the ground running together.

Looking forward to the third quarter, our revenue guidance for our FPGA business of $220 million at the midpoint represents approximately 65% year-over-year growth. When adding 2 months of AMI revenue contribution, our revenue guidance becomes $255 million at the midpoint, putting us at over $1 billion annualized revenue run rate. This strong outlook reflects our confidence in the accelerating momentum of the business and the breadth of demand across our end markets.

The midpoint of our EPS outlook is $0.56, which reflects roughly 100% year-over-year growth. This highlights the powerful operating leverage in our model, the differentiated value of our products and our disciplined approach to scaling Lattice. We expect that we'll be able to consistently drive earnings growth that significantly outpaces revenue growth.

Looking ahead, accelerating AI infrastructure demand, the emergence of physical AI and the addition of AMI create a powerful foundation for Lattice's next phase of growth. The secular trends fueling our business have never been stronger. We are confident that we are in the early innings of a multiyear growth cycle, and our focus is clear: deliver sustained above-market growth while converting that growth into strong earnings and cash flow to benefit you, our valued shareholders.

With that, I'll turn over the call over to Lorenzo for a comprehensive review of our second quarter results and outlook. Lorenzo?

Thank you, Fouad, and good afternoon, everyone. I will begin with a review of Lattice's Second Quarter 2026 Financial Performance, followed by our outlook for the third quarter. We'll then close with a brief introduction to AMI and its business model.

Q2 financial performance was exceptional, exceeding the high end of our guidance. Revenue reached a record $201 million, growing 62% year-over-year and 18% quarter-over-quarter. Earnings growth continued to outpace our revenue growth and exceeded the high end of our guidance. Q2 non-GAAP EPS at $0.53 a share demonstrated significant leverage, growing more than 120% year-over-year and 29% quarter-over-quarter.

Q2 non-GAAP gross margin was above expectation at 71.7%, up 170 basis points quarter-over-quarter and up 240 basis points year-over-year. Q2 gross margin benefited from favorable product and customer mix. Q2 non-GAAP operating expense was $67.1 million, up approximately 10% sequentially and 30% on a year-over-year basis. The sequential increase was primarily driven by continued R&D investment. performance-based bonuses and commissions associated with our stronger revenue and profitability also contributed.

Our Q2 non-GAAP operating margin expanded 390 basis points sequentially to 38.3%, while our EBITDA margin increased 340 basis points to 43%. GAAP net cash flow from operating activities for the second quarter of 2026 was $88.3 million compared to $50.3 million in Q1. Free cash flow in Q2 was $81.3 million with a 40.4% free cash flow margin, up from $39.7 million and 23.2% in Q1. Strong financial performance and the fact that we paid out our 2025 annual bonus in Q1 were factors in the sequential improvement of cash flow. In summary, Q2 demonstrated the strength and leverage of our financial model with non-GAAP EPS growth significantly outpacing revenue growth.

Now for our guidance, which will include our FPGA business and approximately 2 months of the AMI business given the closing at the end of July. Our FPGA business continues its accelerated growth trend. Revenue is expected to grow into the range of $210 million to $230 million. AMI revenue for the partial quarter is expected to be between $33 million and $37 million. In total, Lattice revenue is expected to be in the range of $245 million to $265 million. Gross margin for the FPGA business is expected to be 70%, plus or minus 1% as we continue to manage our supply chain and costs in the face of increasing pressure. Combined, Lattice Q3 corporate gross margin is expected to be 69.5%, plus or minus 1% on a non-GAAP basis. This guidance reflects transitory issues in the AMI business, and I'll discuss those in more detail shortly.

We expect non-GAAP operating expense to be between $83 million and $90 million on a combined basis. Most of the growth in OpEx will be in R&D and reflects our continued disciplined investments to drive long-term sustained revenue growth. We expect income tax rate for Q3 to be between 4% and 6% on a non-GAAP basis. We expect non-GAAP EPS to be in the range of $0.54 and $0.58 per share.

In summary, our Q3 outlook continues to reflect strong revenue and earnings momentum with EPS growth expected to once again significantly outpace revenue growth. This underscores the leverage in our model and our ability to scale profitably while continuing to invest in long-term growth. Earlier, Fouad provided a strategic overview of AMI in his prepared remarks. I'll provide additional color on the business model and the near-term factors that will affect comparability as we integrate AMI. I will also cover the acquisition financing.

AMI brings a highly attractive business model that is closely aligned with Lattice's long-term financial framework. We expect the base revenue of greater than $200 million in 2026 will achieve significant growth in 2027. AMI has built a very profitable business with gross margins in the mid- to high 70% range and EBITDA over 40%, which should improve our already strong business model. As AMI is integrated with Lattice, we expect to see meaningful accretion to EPS starting in Q4.

AMI revenue is primarily driven by firmware licensing, recurring maintenance and subscription revenue and per unit royalties that scale with customer platforms over time. AMI also provides platform enablement services that support customer adoption and help establish durable long-term royalty stream. One of the transitory issues referenced above is a low-margin hardware pass-through business that is not core to AMI's strategic value. AMI began proactively divesting this business before the acquisition. While we expect this noncore business to be fully exited by the end of 2026, Q3 and Q4 will include this revenue. Completing the exit in Q4 should structurally expand AMI's margin profile in line with our expected go-forward model. We anticipate that any other adjustments will be normalized by the end of 2026 as well.

To reiterate, as we integrate AMI, we will show accretion across our business model with meaningful accretion to EPS starting in Q4 while we enable additional strategic growth opportunities.

Regarding the acquisition structure, we purchased AMI for $1 billion in cash and 5.2 million shares of our stock. We funded the cash portion of the acquisition with $925 million of financing and $75 million of cash from our balance sheet. We put in place a financing structure with a $1.15 billion credit facility consisting of a $950 million term loan, drawing down only $925 million of it and a $200 million revolving credit facility. We were pleased with the strong participation from a high-quality syndicate of financial institutions, reflecting confidence in the Lattice AMI combination. Given the strong free cash flow profile of the combined company, we currently plan to reduce leverage to below 2x EBITDA by the end of 2027.

In closing, this has been an incredible few months for Lattice with our record Q2 results, our closing of the transformational AMI transaction and our record Q3 guide. We are very well positioned for strong near-term growth as well as the next level of strategic growth with accelerating revenue, earnings and cash flow generation.

Operator, that concludes our formal remarks. We can now open the call for questions.

[Operator Instructions] Our first question is from Quinn Bolton with Needham & Co.

Congratulations on the continued strong results and outlook. I guess, Fouad, just wanted to start with the Comms and Compute business. Obviously, very strong growth in AI data centers and general purpose as well. But wondering if you could talk about trends you're seeing in terms of FPGA attach rate per server dollar content or ASP per FPGA. Have you seen those trends continue to increase sort of on a quarter-to-quarter basis here in 2026? And then I've got a follow-up.

Thank you, Quinn. A few things that are worth noting this quarter. Number one, the latest Digitimes report shows the server TAM is now up to 20 million units forecast for 2026. which is a really strong growth, much stronger than prior year. And so we're seeing the Agentic revolution still continuing, and we're seeing that drive not just the AI server, but more traditional server, networking, storage, all the cloud infrastructure that goes along with this inference and Agentic revolution. So that has helped our business because we participate in both. We participate strongly in the AI attached server and times. We also participate in the supporting infrastructure, the cloud infrastructure. So we're really happy about that.

The attach rate continues to grow. The CapEx continues to grow. The new applications continue to grow. The ASP of some of the new products continue to grow because we're coming in now with further and more complex security requirements as an example. And we continue to be very positive on the characteristic of our FPGA such as low latency determinism, parallel processing connectivity, wide range I/O. These are the 1.2 to 3.3 volt I/O in the data center, connecting up to 1,200 sensor and some of these servers together. And so our FPGA continues to find use cases in numerous new applications such as, for example, power and cooling.

Excellent. And then I guess maybe for Lorenzo, as you bring AMI on board, it looks like it has gross margins in maybe the mid- to high 70s. I think you said op margins above 40% in the core FPGA business. You're already approaching a 40% op margin, and you've kind of highlighted that earnings growth will grow much faster than revenue. And so now that you've closed the AMI acquisition, can you give us some thoughts on where you think a longer-term model might be for gross margin and operating margin? Could it be in the low 70s and low 40s? Is there another range we should be thinking maybe a year or 2 out on a combined basis?

Yes. So I'll try to answer your question in near term and long but we are benefiting right now in our business model from very, very strong revenue growth. But if you look underneath, we continue to invest. AMI is also a R&D heavy organization because they continue to invest for the future. So the model that you see right now is probably approximately what we'll see for the next little bit.

And as we get into 2027 and our longer-term growth aspirations manifest themselves, we see a little bit better performance in our business model than we may have been expecting before. So 70-ish percent on the gross margin and a little bit sooner to 40% on the operating margin level than we had seen in the past is probably where we're taking it. But that's also -- keep in mind that's in the face of with the accelerated revenue growth we're expecting. So in that model, you'd start to see a very significant acceleration in EPS as well.

Our next question is from Christopher Rolland with Susquehanna International Group.

Maybe just following up on the last question about gross margin. Lorenzo, I think you said there was a hardware business associated with AMI. How much revenue is that hardware business? Was that on top of the $200 million? Or does that take from the $200 million annual revenue? And is that the reason for margins being down in September?

Yes. So this -- look, this is a really transitory issue. As we pick them up, we have a partial quarter. They have already begun to, as I said in my comments, divest the hardware part of the business while maintaining the royalty stream. So it's not -- from the Lattice perspective, overall, it's not a meaningful amount of revenue. And by Q4, it won't have an impact on our overall financials or actually even the AMI-specific financials. So it's something that we wanted to point out because it does cause a Q3 step down from what we're expecting.

Okay. And was that revenue on top of the $200 million or...

The $200 million a year run rate for AMI would exclude that in the end.

Okay. Excellent. And then perhaps a follow-up. I'll talk about I&E. I think the situation with I&E is maybe they didn't understand we were in a new semiconductor cycle, and we're kind of holding back spending. Any update on bookings there? Has it accelerated? Have they finally got the message? And I think we -- back in the day, we're talking about maybe a $75 million normalized run rate for that business, but it now appears maybe to be higher, particularly with the guide for the fourth quarter, I would imagine it's higher. Can you talk about I&E and what's happening there with the channel, what a new normalized level is, and that would be great.

Yes. Thank you, Chris. Our Industrial and Embedded segment is doing great. I&E is doing good. And we have talked about the really strong sequential and year-over-year growth, and we see this continue to grow throughout the rest of the year. The PMI now is at levels -- they are very positive levels worldwide. There's a bit of a temporary slowdown in China, but continue to be very excited about the design wins, the physical AI momentum. And the recovery of that business. The channel is -- inventory is now where we need it to be. We're not focused as much on the channel inventory anymore as we focused on supply being the main focus and very positive on all the different segments, including some of the new robotics and humanoids, we're doing quite well. Some of the new autonomous vehicles, new medical application, aerospace and defense. So the list goes on, on penetrating few new market segments and accelerating the growth into 2027.

Let me just add on one of the things we are not -- at Lattice, we are not impacted by automotive. This is a relatively small business for us. So that end market weakness is not having a drag on us.

The only other one -- Please go ahead.

Our next question is from Melissa Weathers with Deutsche Bank.

I wanted to touch on something you just talked about the supply side. With everything seeming like it's coming back pretty hard. Can you just talk about any constraints that you're seeing on the supply side, how you're managing it? And could this actually gate your growth going into next year?

Yes. Thank you, Melissa. As I said, we are doing good on sort of the fab side at the front end and the testing side on the back end. In the middle, on the assembly side, there are constraints across the industry right now. And so us and the rest of the industry are experiencing these constraints. We are putting capacity agreement in place that -- and then qualifying new capacity that we believe will get us in line -- supply and demand in line by September, and we should be in good shape in Q4 and definitely for 2027.

Great. And then on AMI, congrats on getting the deal closed. I know you've talked about your SAM, I think, doubling with the inclusion of AMI. So can you just talk about, is there any like new kind of long-term growth rate framework that we should be thinking about for FPGAs? Is there like an adder to whatever percent growth you could have grown in the past? Just any help on how we can think about AMI layering like actually accelerating your FPGA sales, I think that would be helpful.

Yes. So Melissa, a couple of things. I'd like. Number one, from our long-term aspirational goal, we'd like to hit $3 billion by 2030. So that's the goal. We are already ahead by about a quarter. So if you look at -- we had 3 months ago said that we'll hit $1 billion run rate by Q4 and the combination of Lattice plus AMI, we just hit this in our guide in Q3, a quarter ahead. So we're seeing the benefit of this. We should exit this year at a $1.2 billion run rate. And so you could see where our growth is accelerating.

Our next question is from Kevin Garrigan with Jefferies.

Congrats on the great results. Fouad, just on the supply again. With these new negotiations that you're doing, are any of your manufacturing partners trying to negotiate higher prices? And can you pass those along to your customers?

Yes, the costs are increasing across the industry, not just cost from a supply chain point of view, but a whole bunch of expedite fees because the customers are all under pressure to get supply ASAP. So we're seeing the cost increase across the industry. And we're doing the best we can to absorb some of these costs, and we're going to have to pass some of these costs. So it's going to be a mixture of us absorbing some and passing some.

Okay. Great. And then you continue to see strong bookings, strong backlog kind of into -- all the way into 2027. I mean has that visibility extended over the last 3 months? And how much of that backlog is noncancelable?

Very good question. The visibility is increasing daily. I mean it's really unprecedented. We've got visibility all the way to the end of 2027. 2027 is pretty much booked. And so we're seeing that to be very strong. We are putting capacity agreements in place with our supplier where we're going to have to take the capacity. And so we are, in turn, putting capacity agreements with our customers and partners to make sure that our customers provide us with the same commitments. So yes, I mean, I think you're seeing this across the whole supply chain now as with our supplier and as with our customers, putting all these agreements in place.

Our next question is from Ethan Potasnick with TD Cowen.

Congrats on the results. Just a near-term question. You guys guided the stand-alone FPGA business. But I was wondering, could you guys help us think about the sort of the relative growth rates across the various segments as we move through the second half?

Yes. So we do break down our FPGA business. We do break down our Comms and Compute and our Industrial and Embedded so those have been broken down. And you could see that the Comms and Compute has been growing very strong. So this Q2 was 83% year-on-year growth, 18% sequential. Industrial and Embedded 36% year-on-year growth, 16% sequential. We haven't broken this up for Q3. On the AMI side, we expect AMI to grow about 25% year-on-year. So that would give you all the 3 major segments.

We're seeing strong demand across our end markets, though. And I'd just say the Industrial and Embedded business has been lumpy in the past and probably going to behave that way in the future, but the general trend is up.

Okay. Okay. Great. Very helpful. And then last quarter, the team suggested AI-related revenue would sort of approach 25% of company revenue in '26. I was wondering if there was an update there and sort of how AI demand tracked during the quarter and where within that opportunity that growth is sort of coming from?

No, we're on track to meet this 25% coming from revenue and exceeded. We had a discussion, actually multiple discussions around this metric. And what's interesting is the AI and the Atech revolution is driving actually more of the traditional infrastructure. So we're seeing tremendous growth in the traditional supporting cloud infrastructure from servers to networking to storage to memory. that drives, in turn, our FPGA demand.

And so it's a bit harder to just say AI, which we -- if you want to categorize AI as sort of AI with GPU or CPU or XPUs inside, this is the 25% or where, in some cases, we assist that. But the other 75% of the business is also growing very rapidly. As you could see, inside our Comms and Compute, our server business is growing even faster than Comms and Compute, which grew at 83% year-on-year. So you could see the rest of the traditional infrastructure is growing actually at the same rate, if not faster.

Our next question is from Ruben Roy with Stifel.

Fouad, maybe just to follow up on that last point, thinking about the server growth. We've heard a lot recently about CPU attach and CPU. You talked about agentic a little bit in the prepared remarks. I'm just wondering if you could maybe talk about where you are on the CPU side with the core processors. And is that starting to drive some of the growth that you're seeing? Or is that still on the come? How do you think about that as you think about 2027?

No, CPU has definitely been a very strong driver of growth with traditional servers. And we have not broken up our server growth, but our server growth is higher than the Comms and Compute, which was at 83%. So you could see the server has been growing very nicely year-on-year. We also feel that the AMI acquisition is going to help both the Compute and Comms and have the Embedded-- Industrial and Embedded, both segments of our FPGA business grow faster. There are some synergy at customers, and we're working together on joint solution. And so that AMI acquisition should be able to help drive a higher growth rate in both the Compute and Comms and the Industrial and Embedded. And in turn, we should be able to help them drive faster growth in AMI itself. So go ahead, Lorenzo.

And just refresh what Fouad said earlier that in the traditional servers, our attach rate is growing as well. So one CPU is multiple FPGAs going with it in the infrastructure that's supporting AI. So a very healthy ecosystem for us.

Yes. Got it. Maybe just a follow-up. Fouad mentioned 25% year-over-year growth for AMI. Was that for Q3 or second half? And I guess kind of the bigger question around that is I think you framed previously that AMI growth was maybe in the high teens, accelerating into '27 off of that rate. And thinking through firmware attach rates on servers, could we assume at some point that AMI grows something closer to your own server growth rate?

So right now, I'll just clarify that when we said 25%, that's our expectations for 2027. I thought that would be helpful for you guys to start building your models. And we are still in the very initial period of integrating, right? We closed on the 27th of July, and we are beginning to put meat around the bones of the strategy we talked about when we were talking about the acquisition, which is developing these solutions that we'll bring to market in the future that would further accelerate that growth rate. And that we've yet to really quantify, but that is a strategic driver of this acquisition.

[Operator Instructions] Our next question is from Srini Pajjuri with RBC Capital Markets.

Fouad, on the new products, I think you gave us a number -- target number for this year, roughly 20% to 25%, I believe. Just wondering how that's tracking? And also, given how strong of a growth you are seeing, I just looked at one of your competitors reporting, and it looks like you grew 50% over competition. So I'm just trying to understand, is this driven by share gains for you, new products? Or is it pricing related? I know you talked about demand being very strong, but just curious to understand how the new product ramp is going.

Excellent. Good question, Srini. Thank you. Yes, we had guided to that range, and now we're going to exceed the high end of that range. So we expect new products this year to exceed the 25% of total revenue. So as you could see, our new products are doing fine and definitely contributing to that growth rate. And then the growth rate is across all our markets. So we're seeing it in Compute and Comms. We're seeing it in Industrial and Embedded, and we're going to start seeing it with AI. And so we're quite excited.

The one thing, Srini, I wanted to point out is, recently, I've been reading these reports about people talking about the Rule of 60. It used to be the Rule of 40, now the Rule of 60, and the rule was like adding CAGR plus EBITDA 1% margin. And I just want to point out, we're a Rule of 105 right now. So we're above 100. We'll enjoy it for both Q2 and Q3. We're not promising to do this in the future, but 62% growth and 40% plus EBITDA gets us to about 105%. So you could see not just the revenue growth, but also EPS and profitability growing faster than revenue growth. So we're excited about that.

Got it. That's very helpful. And then on the AMI, the 25%, I just want to clarify that it's 25% over $200 million because you did talk about some hardware pass-through revenue.

That's correct. It's 25% over the $200 million, yes.

Got it. So my question on that Fouad or Lorenzo, is that it's a pretty solid growth. It's a very healthy environment out there. I'm just curious, I mean, I know it's early days, but does it include any of the revenue synergies that you talked about? Or is it still kind of early days? I just -- is it more of an organic?

It is early days on the revenue synergies. I do believe we're going to have revenue synergies on top of that, but we -- this does not include the revenue synergies.

And it'll take a little bit of time to bring to market, Srini.

Our next question is from Quinn Bolton with Needham & Co.

Just had a couple of quick clarifications or follow-ups. On AMI, will you report that as a third segment? Or going forward, do you plan to put it into Comms and Compute and Industrial and Embedded?

That's a great question. What we are thinking right now is as we close the quarter and we report, we're going to talk about the FPGA business as a segment and the AMI business as a segment, and we'll provide revenue and gross margin by those. And we will -- we are working right now on how to best articulate the operating margins given that we're in the process of developing shared infrastructure and support.

Will you break out Comms and Computing and Industrial Embedded within FPGA?

Sorry. Yes, I said that. We are within FPGA, we'll give you the same revenue looks we're giving you today.

Perfect. And then I don't know if I missed it, I apologize if I did, but did you say where channel inventory ended the June quarter? Was it below your 2 months target that you discussed last quarter?

Yes. Here's the way we're thinking about it now because we've gone from an environment of having to manage that down to get to the right level. We got there. And what we're doing right now with the channel is using it to help us ensure supply to our customers. So I think I've said before in different forums, once we got to the 2-ish, it would probably fluctuate up and down. And that's what we're seeing, but we're staying in that range.

[Operator Instructions] Our next question is from Melissa Fairbanks with Raymond James.

I can't argue with the Rule of 105. That's not a bad target to hit. I had...

So we can achieved Melissa that's not a target.

I know, I know. I know. I'm teasing, I'm teasing. But I had a follow-up question. I know that AMI helps expand your addressable market and how much you can actually address within those markets. In Industrial and Embedded, I'm assuming, especially as we get into more of like robotics or some defense applications, automotive, some of these more highly regulated applications. Is the go-to-market a little bit different even with AMI than it is in the compute segment?

Yes. I think there's going to be tremendous opportunities for AMI and Lattice to work together in Industrial and Embedded because these systems are very big on the need for platform firmware and for infrastructure manageability. And the early meetings, we've had many meetings at Computex in Taiwan. We've had many follow-on meetings with other partners on physical AI and very excited about how the integration of AMI and FPGA can offer new solutions to customers. So stay tuned. We'll have a lot more to say on that in the future calls.

We have reached the end of the question-and-answer session. I would like to turn the floor back over to Rick Muscha for closing comments.

Thanks, everyone, for joining us on the call today. We'll be attending the following investor events this quarter: the KeyBanc Technology Leadership Forum on August 11; the Jefferies Semiconductor, IT Hardware & Communications Technology Conference on August 26; and lastly, the Benchmark TMT 1-on-1 Conference on September 10. Thank you very much for your participation, and have a good evening.

This concludes today's teleconference. You may disconnect your lines at this time. Thank you for your participation.

All right. Welcome back, everyone. My name is Mark Garcia with JPMorgan, and very happy to be hosting the management team here of Lattice Semiconductor. We have Fouad Tamer, CEO; Lorenzo Flores, CFO; and special guests, Sanjoy Maity, CEO of AMI. Great. So for those who have questions on AMI, this is great opportunity.

So Fouad, it's been over 18 months for you now. Maybe you can just sort of catch us up and what did you inherit there at Lattice? And then -- but really importantly, what's changed the most operationally today?

Yes. Thank you. So when I joined, the vision of the company was to do a full portfolio of FPGA all the way from low end to high end, and there was a road map to do chips with millions of gate and -- including SoC, which is integrating the processor inside the FPGA. And when I joined, we refocused the company on the small and midrange, no SoC, meaning we have refocused the company on what's called -- the small FPGA is called our Nexus product line, N-E-X-U-S, and our midrange is Avant. And we have 3 product lines, what we call the pre-Nexus, Nexus and Avant. So Nexus, it goes up to about 200,000 logic cell, Avant to about 600, and that's the focus of the company. And we did this because our belief is that this is our sweet spot for FPGA. This is where FPGA make a lot more sense being deterministic, being low power, being low latency, being fast boot time as a companion chip to our silicon partner on the CPU, on the GPU, on the board management controller, on the NPU and MCU, on the networking switch and NIC, on the variety of different sensors. So on the sensor, we sit near an image sensor, let's say, we've got partners like Sony on the radar, for example, we announced a partnership with TI. We've got a variety of sensor work with ADI, so image, LiDAR, radar, infrared, ultraviolet as well as all kind of industrial sensor like temperature and then a variety of motor control, so different actuators. And so we feel this is the sweet spot for FPGA.

We're a high-volume supplier. So this year, we'll do over 200 million units, on our way to over 250 million. And this is our focus. So what's different is we don't try to go after these large FPGA or the millions of gate where, in my background, when we were -- I was at Broadcom and eventually Inphi, onboard at Marvell, I saw these large FPGA going in for 18 months and being replaced with ASIC because they are very high power, very expensive. So what we do is the small FPGA. We're literally on a rack of server today. We went from 10 FPGAs per rack to 100 FPGAs per rack, very low cost, very cost effective, very fast boot time.

On the SoC, we go partner with folks like NXP or ST or Renesas, where we, instead of trying to put the processor inside of the FPGA and then you're stuck with designs, our competitors like Altera and Xilinx have these SoC, they are 10-, 12-, 15-year old with really old interfaces and et cetera, we go partner with the greatest -- latest and greatest from NXP and ST and Renesas where we go jointly to market. And what we show, like, for example, we go partner on industrial control and humanoids where, together, we're going to have the best of both worlds. So you end up with the best deterministic low latency from an FPGA and the best latest tool set, the latest DDR, the latest PCI from NXP. And together, we're actually going to bring a solution that's lower latency, lower power. It's counterintuitive. If you put 2 of us together, we're together lower power because the deterministic task get done on FPGA versus the high-level stuff in NFP.

And then we refocus on data center. Data center has been growing very fast for us. The server market for us had been growing 85% year-on-year. Demand is strong for the foreseeable future, well into second half '27 and up in to the right. So that's a quick summary.

Great. Actually, well, let's stick with data center then talk about your segments. Compute and Communications, as you said, has been growing more rapidly than the very high growth CapEx market. So can you kind of take us through that? What are the drivers there in terms of attach rates and ASPs for you, especially on the AI side within the Compute segment?

Yes. So if you look at these, we -- it's a multiplication of various factors, right? So 5 factors, if you wish. Number 1, we grow along with -- obviously, we grow along with CapEx, and CapEx, you could think of CapEx as a number of servers. So we grow along with the number of server and that's been the CPU, and the CPU has -- recently has been growing faster because of agentic. So that's number one. Number 2 factor that's multiplicative is AI as a percent of the total server because we do have higher content on the AI server than traditional servers. So Number 1, number of server; Number 2, AI content. Number 3 is attach rate. So attach rate, we've got a number of applications that are continuing to grow. When I joined Lattice, we had 10s of FPGAs per rack, today we have 100s of FPGAs per rack. So the number of application we're finding and quite a few we can talk about are growing. Number 4, the needs continue to grow. So security has been, for example, a big application for us. Over time, the security needs have been growing higher and higher, driving a bigger FPGA per board and hence, a higher ASP.

And finally, Number 5, recently with agentic, it's not only CPU, but also storage has been growing. And so that's Number 5. The second part of Number 5 is the architecture have been changing. So if you look at the -- disaggregated architecture have taken over, disaggregated drives more FPGA because like, for example, right now, you go from 1 board being the main processor board to 2 boards, HPM where you've got the processor and where we have I/O expansion, and the second board, SEM, which is a satellite board where we do the security with root-of-trust. But also the unit of compute has migrated to where before it was 1 rack with everything in it to now 4 racks, a compute track, a communication rack, a power rack, a cooling rack. And so with that, you could see we end up having a variety more application, variety more needs for FPGA.

So you really take maybe it's really 6 factors together, and you can see that's why we're growing faster than CapEx. The latest numbers that we've seen for CapEx were just the hyperscaler for next year in '27 is going to be over $1 trillion, over $1 trillion. This number continues to grow, and this is just the hyperscalers. This doesn't include the neo cloud, does not include the sovereign, it doesn't include enterprise, so this number continue to grow very fast and will grow faster than the CapEx.

It's interesting, [ Harlan ] was just up here before and talking about the growth in custom in the data center space. And what's really different this time because it seems like custom and FPGAs are really going to coexist here now going forward. So even with the growth of custom, the FPGA business is going to continue to grow at a more rapid rate.

Yes. This is a very good comment. I mean, this is probably Number 7, the growth of custom. So that's another factor. We are actually totally agnostic to whether it's custom, whether it's standard product, we do want to stay Switzerland, and that's why this AMI acquisition was so powerful is we are Switzerland for companion chip for silicon and whether it's standard or custom, AMI comes in as Switzerland agnostic for firmware. And so together, we're bringing these solutions to customer and continue to support the wide ecosystem. But you're right. It's very interesting today to see the custom going against the standard product, but we support all of them. So we don't take sides. We support all of them, and we're friends with everyone.

Great. Okay. So moving on to the Industrial and Embedded segment, and this is potentially the next big wave of growth from physical AI, big opportunity. Maybe you could take us through that opportunity.

That's an excellent question. So yes, you're going to see these factors superimposed when you get to '27. So not only do we have growth in data center and see demand for the foreseeable future, on the Industrial and Embedded, which is what we call now our segment for industrial, for medical, aerospace, defense, consumer, we put everything in this Industrial and Embedded segment, automotive, that segment had been dragged down by inventory situation. When I joined Lattice about 18 months ago, we had 6 months inventory in the channel. We said we were going to bring it down to 2.x, and we are 2.2 as of last quarter. Our goal is to go under 2 this quarter, and we're on our way to go under 2 this quarter.

The last time we're at 1.x, we had 10 quarters of strong demand ahead. And we do believe we're entering this phase now where Industrial and Embedded should see a really strong -- a number of years ahead of strong Industrial and Embedded. So Number 1, we are growing now to natural demand as opposed to undershipping demand because inventory in the channel has been normalized. But even more importantly, we've got some phenomenal design win across all these application, industrial, humanoid, drones, autonomous vehicle, aerospace, defense, medical, consumer that are all going to market this year and will inflect into '27. And content of some of these like humanoid is very significant. So we do feel like Industrial and Embedded is going to be another growth vector for us.

If you go back to the Lattice in the '22, '23 time frame, the growth was all from Industrial and Embedded. And now you put these 2 on top of each other and should be very strong. So 3 growth vectors if you think of us in the future, data center being one, Industrial and Embedded being 2 and AMI coming on top, being 3.

That's great. Well, let's go to AMI as the next catalyst of growth here. So I think a lot of folks in the audience maybe weren't familiar with AMI, which has this incredible history and is very close to the server, CPU, the GPU customers. Can you take us through what does AMI add for Lattice? And how does it change what you can offer your customers?

Yes. So I'll give a high level, and I'm going to turn to Sanjoy because he's the expert. AMI started 43 years ago and a very rich heritage of -- in this space. And so when we started like last week, we did a couple of calls with the sell side because on our earnings call, we -- it was -- it's not clear what it is. So -- and then we're going to -- today, is one of the many calls hopefully we'll have on the buy side and we'd be organizing some of these calls as move forward. When -- the first question when I ask folks on the call, hey, what do you think after we explained the acquisition, people said, "Oh, let me go get my software analyst." I said, no, no, wait a minute, this is not software, okay? So that's why I think we want to explain this from the basic level.

So there's 2 businesses in AMI. It's a simple business, 2 businesses. Number 1, 55% of the business come from what's called boot firmware. And the second one, 45% of the business come from manageability, okay? So that's it. Very simple.

Number 1, boot firmware, what's boot firmware? Without AMI, there is no server. There's no data center. So AMI basically run on a flash near the CPU. it sits near Intel, AMD, ARM, NVIDIA, custom processors. And it's -- fundamentally, that firmware brings up the CPU. It boots the CPU. And then it looks on the server, it looks what components are in the server, recognizes the components on the server and then it boots the operating system. And then the software run on top of the operating system. So think of this AMI as a fundamental layer that's even underneath the hardware. You should think of AMI as hardware. That's how we think about it, okay? So this firmware is like hardware. And -- so that's about 55% of the business. And we had some great calls with all the different partners on the CPU side.

On the manageability, which is about -- and by the way, that AMI business is 75% server when it comes to boot, okay? On the manageability side, think about it, this is firmware that runs with the board management, the BMC, the board management controller. And these are board management controller from it could be ASPEED, could be Nuvoton, could be an NXP, could be new guys that are coming to market and not announced yet. So a variety of different BMC. And what -- think about the extreme example of having data center into space. And I think whether you think is going to happen or not 10 years from now, it will definitely drive the right behavior, which is, if this data center, in this space, you can't really sell a human over there to go change some stuff. So you want this data center to be totally self-managed, self-healing.

We've got -- Number 1, you want to collect a whole bunch of telemetry. We actually have customers on the telecom side of the world. They are coming to us and saying, "Look, we want to collect the same telemetry on voltage, temperature, et cetera, to be able to do, instead of doing predictive maintenance, we want to do predictive failures." The failure rate of some of these components is actually extremely high, higher than you would think. And so collecting all of this telemetry and be able to do predictive failure on AI model based on this data, which by the way, does not exist in the [ common ]. So this data on 43 years of predicting all of these failures attached to voltage, temperature resides within the 4 walls at AMI, does not exist in an open domain.

So you don't -- even if you wanted to train Claude, you don't have the data, the data is AMI. We are training the models to be able to do this kind of stuff. So Number 1 is phenomenal thing on collecting all this thing. Number 2, you want to be able to operate even if the OS is down, and that's what they do. So AMI operates even if the OS is down and you want to recover from a failure, they would be able to do it. And then they have a view of all the different components in the data center. And the high level you could think about the main cockpit for the data center where you control the whole data center, data center management could be enabled by the AMI firmware. And so very, very fundamental, very strategic.

Finally, the value prop is that they do it across all of these vendors. So you have different people that in turn may have their own boot, AMD may have their own boot, ARM may have their own boot, NVIDIA may have their own boot, but AMI works across all this. Again, different BMC folks may have their own BMC firmware, but AMI works across all this. So being this agnostic Switzerland firmware across all of these different components is a humongous value proposition that we bring. If you want, for example, to do plug and play. So today, you're working on certain components, tomorrow you want to bring a different component. That's what -- and by the way, that FPGA companion chip is tremendous because we have a low latency. So let's say you want to do a leak detection for cooling, with the FPGA sitting down with a very low latency being able to shut down that leak faster than any component you ever see because our latency is very low, okay?

Think about if you want to optimize the performance of the GPU. You want to be able to direct cooling to the GPU that's actually more -- that's being more tasked and doing more performance than the one that's idle, well, guess what, AMI allow you to do this kind of stuff. So very powerful together. Longer term, we think we'd will bring solutions to market. There is a presentation on the website that talks about some of this, but that's a high-level view. And Sanjoy, I'll turn it over to you for more detail.

Yes. Thank you. So yes, as Fouad mentioned, right, I will just give another analogy why it is foundational for every server or every data center. So without AMI firmware, I would say that the servers are basically metal and electronics. It's not operational in that state. So its analogy goes like a car, right? You see the ignition system. When you turn on, there is an engine check, there is a transmission check, there is airbag check, all these checks happen and then it starts the car. That's our boot firmware. Then if you think of the continuous, when you were running and driving the car, there is a system continuously checking the temperature, checking the -- everything is -- engine, transmission, everything is running optimally. If it is electric car, you have a battery management, everything is going, that's our manageability. Without these 2, car will be not functional. And same way, our boot firmware and the manageability firmware, the servers will not be functional.

And as we talk about software, operating system and software, that is the software like infotainment system in a car or a navigation system in a car or something other than playing some apps or CarPlay, those are the software. So we are foundational in this data center. And as he mentioned, I will get into a little detail that there are inflection points in the industry today based on 3 things: scalability, security and sustainability. And nobody can ignore these 3 today. And that is driving and changing the technology landscape tremendously. And we see a totally different way of making servers and the racks and powering up and the data centers are built that way.

So there are a tremendous amount of opportunity to accommodate that because it needs a runtime programmable hardware at the fleet. It cannot be determined at the beginning. And FPGAs are very critical to do that. It's the same analogy I will give. If you connect to a device to your USB port, it could be camera, it could be storage device, it could be printer, it could be anything else, but the software and operating system recognize and do this automatically. In case of hardware, there is no programmability. Only programmability you have is FPGA. So these technologies are coming because of the scalability reasons. And that is why the growth of this kind of programmable hardware is growing tremendously. And AMI is working with the large hyperscalers as well as the neo clouds, as well as the enterprise level customers and ODMs and try to make sure that this kind of technology shift, which is happening in the industry are well pre-PoCs level it is done, and it is delivered by these companies together ahead of time.

Great. So Fouad, so then take us through the -- how you will expand Lattice solutions through or across that AMI footprint, which, as you just explained, is a very, very deep relationships?

Yes. So the one thing I'd tell you is, as I said on the earnings call and I've done this twice before, I've done at Broadcom. When I joined Broadcom, we were Number 2 on the switch, Marvell was Number 1. And we acquired a company called LVL7 that brought us protocol firmware that we used to come up with this reference design. And we had reference design, hardware and software that were turnkey tested that we provided to our ODM partners in Taiwan. And very early on, partnered with ODMs in Taiwan to create the white box server and the white box switch. And pre-Broadcom, these were all OEMs. Broadcom invented this white box switch, white box server and LVL7 was an important piece of the puzzle to bring these reference design to market.

The second time I did it at Inphi. And at Inphi, we were selling PAM DSP inside the data center, and we had a very difficult decision to decide to go into the module space to do a DCI for Microsoft between data center 80 kilometer, very different business because we went from a component vendor to be a subsystem vendor, we're selling these modules. And the modules ended up being 20% of the Inphi business at the time, but that knowledge that we had of having silicon photonics, having these system being designed and tested were tremendous for the component side of the business because when the component side of the business had a question on how to test, on how to take the stuff to market, the knowledge was in-house.

And so these are 2 examples in my career where adding this capability was tremendous. And when we made this acquisition, Inphi, they were not -- people did not understand them. So we acquired a company called Cortina and people thought we were acquiring Cortina to go into telecom framer and mapper business, little they know that we acquired them because they had high-speed analog people that we wanted. When we acquired ClariPhy, which was a coherent DSP at Inphi, again, people thought we're getting in telecom DSP business, well, guess what, nobody knew we're doing this for Microsoft. When Microsoft announced this, it was a shock in the industry going to this DCI business. And so these are 2 examples in my career where I've done this before.

We believe AMI would open up a similar opportunity for us. We've got a slide on our website that shows some of these solutions that we are working on together with customers. And Sonjoy talked about PoC, things like rack boot. So people don't want us to boot a server at a time, they want to be able to boot the whole rack. We're working on solutions together. Things like power and cooling, things like retrofit, so people want to go back to the old fleet and be able to retrofit the old fleet with solutions for security. Things like plug and play, where you want to be able, for example, to plug and play new components inside of your existing server. So all of these are going to be enabled by a combination of AMI and Lattice.

Now where does this take us? Well, Number 1, it opens up a short-term opportunity to build a multibillion-dollar company. We said we'll exit this year at $1 billion run rate in Q4, and we're confident this will happen at very good gross margin, very strong cash flow and very strong EBITDA. So we're saying it would be at least 40% free cash flow. Next year, we grow even much faster. And taking this through the longer-term vision, this could be the first one of our many acquisitions that move the company from just a component supplier to a solution supplier. And today, we're into all these markets where we're seeing the various market across from -- today, this is just compute, but there's COMs, there's power and cooling, there's a whole bunch of industrial embedded type of market along vision and motion control, et cetera, that we can get into. So that's where we take the company.

Great. And how about on timing, timing of closing and then thinking ahead maybe to the integration and how to think about success going forward?

Yes. So look, integration is very important. I've done it like 25 times. At Broadcom, I had 18 different companies. Inphi, we had like half a dozen. Even Groq, we acquired a very key piece of Groq that enabled the $20 billion acquisition by NVIDIA. So integration is very key. Sonjoy and I are going down to Atlanta together, it'd be my second trip there. We're going to spend a few days with the team. Then we're going to go to China -- not to China, to Taipei for Computex and visit. They've got 600 people in Taiwan. Then we're going to go to Chennai and Pune where they've got 700 engineers. Here is a very interesting data point. There's only 5,000 -- I'm not sure if this data is correct, but if that's correct, it's like crazy. There's only 5,000 firmware engineers in the whole world, 5,000. That's it, 5,000. Over 1,000 of them are at AMI. Over 20% of the firmware engineers in the whole world are at AMI. And if this is correct, this is like totally crazy. Is that correct?

Yes.

So integration is going to be important. We're on top of it. As far as delever, we think we delever very fast. The regulatory just U.S., we think we get this done in 45 days. And we've got an inside date of July 6. We think we'll close by then. Delever, let me at least ask Lorenzo to -- he has at least 1 question, so you talk about delever, Lorenzo.

He's up next. Yes.

Well, so while Fouad brought it up, Fouad said, we expect to have combined, very strong free cash flow and the debt we're taking on will be a little bit less than $1 billion from this acquisition. We are going to delever rapidly and be at or under 2 by the end of next calendar year. So we'll be well positioned. We've always had a strong balance sheet, and we intend to maintain that.

Great. There is 1 question here to see in the back first.

How is it that your return on invested capital over the past few years has been low? Or am I wrong?

It's fundamentally the depressed profitability we've had given the decline in revenue from -- in the mid-700s 2 years ago down to the low 500s where we were. So that's what's driving the lower ROIC. We get a lot of leverage out of our business model as we grow revenue, both the -- the OpEx will not grow nearly as rapidly as revenue. This year, we're investing. That rate of growth in investment will slow down. And then our fixed capital investments have been really high, relatively high in the past couple of years because we've been doing some catch-up investments in infrastructure, that should tail off also. So the ROIC should improve significantly.

Think about Q1. In Q1, we delivered 40% year-on-year revenue growth, over 80% EPS growth. So EPS is growing at twice the rate of revenue, and we should be able to get much better as revenue scale.

Thank you for the excellence. That's really appreciated. So 2 questions, 1 for Fouad, 1 for Sanjoy. So the question about FPGA is, obviously, you guys have placed a lot of investment into Ultralight. Can you talk a little bit about, obviously, the AI Edge use case? Obviously, there's a lot of questions around new devices going online, localized AI processing. And then for Sanjoy, in terms of AMI, are you -- do you have any plans to open source some of the AMI frameworks for OS development or custom OS development, obviously, with a lot of AI going up online, looking to run the orchestration layer beyond what AMI is doing?

Two great questions. Let me start with the first one and turn it over to Sanjoy for the second one. On the first one, we distinguished between near Edge AI, which is really the heavy processing. And there, we do believe that's the domain of NVIDIA, Qualcomm, our partners. We don't want to go into this new Edge AI. The other FPGA competitors, they do, and that's fine, more power to them. But we are very focused on called far Edge AI, which is near sensor. We call it contextual intelligence, near sensor intelligence. There, our Number 1 role is to be this companion chip where we ingest all of the whether vision, motion control and then do the preprocessing. So our role, we think, is adding AI as a preprocessor to make the near Edge AI being NVIDIA, Qualcomm more efficient and more productive. Sanjoy?

Yes. So for the open source and open architecture perspective, it's not a plan, we already did that. And we are the only -- we are the platinum member of OCP, and we are the only firmware company in the world who has contributed the boot and the manageability format the open source. As a matter of fact, today, we are actually doing a major event in the -- jointly with Meta at Meta's campus. And across the industry, including Meta, Microsoft, Google, all the speakers are there, and it is happening today about our open source strategy and what we are doing. We are doing it for last 3, 4 years. We are a big contributor to Linux Foundation OpenBMC. We had, as I mentioned, the most secured, OCP S.A.F.E. is the security compliance stack. At OCP, we're the only one who contributed that and maintaining it for the industry.

And Lattice is also very committed to open source. We've also been a big contributor on OCP. And we do believe that together, we want to be the Red Hat for the space. So with open source, Red Hat grew a great company by contributing to Linux, and we'll do the same on manageability and control.

Okay. Clearly, demand is very strong. So maybe, Lorenzo, we can talk a little bit about the supply side. What actions have you had to take, or are you taking in order to ensure that you have the supply, which is very tight out there in order to meet the demand?

Sure. I'd first start -- should give some credit to our Head of Operations, who started to sense the tightness in the fourth quarter of last year and laid the groundwork for getting supplier commitments through the year, starting on the wafer side, which we are fortunate we use older nodes. They tend to be less constrained so our wafer supply got locked in. And we did actually buy in advance of the demand we saw so that we have a wafer supply. So we have a fundamental inventory strategy of building inventory in-house, and we've begun to execute that, like I said, as soon as the Q4 last year.

Then the other side of the assembly -- of the overall process is the assembly and test process. And that's where we've seen and the industry has seen a little bit more constraint. When we started to identify that, we went out and we were in a program of diversifying our supplier base, multiple size from the same vendor and more vendors. But we're still seeing what everybody else is seeing in the industry and some pressure on costs, price increases from our suppliers, but we've been able to manage that both with negotiations as well as obviously passing it through where we can to our customers.

So we feel we're in a good position from ability to support the demand. We're in the -- we have the fortunate problem -- good news problem of demand keeps increasing. So we have to keep working at it. And so far, we've been able to manage the cost side and sustain our gross margins, and we expect to be able to do that for the rest of the year.

Great. And we're out of time.

Ladies and gentlemen, greetings and welcome to the Lattice Semiconductor First Quarter 2026 Earnings Conference Call. [Operator Instructions]

As a reminder, this conference is being recorded. It is now my pleasure to introduce your host for today, Mr. Rick Muscha, Vice President of Investor Relations. Please go ahead.

Thank you, operator, and good afternoon, everyone. With me today are Ford Tamer, Lattice's CEO; Lorenzo Flores, Lattice's CFO, and and Sanjoy Mike, AMI's CEO, will provide a financial and business review of the first quarter of 2026, an overview of the AMI acquisition and the business outlook for the second quarter of 2026.

Both a copy of our earnings press release and the press release announcing our planned acquisition of AMI can be found at our company website in the Investor Relations section at latticemi.com.

I would like to remind everyone that during our conference call today, we may make projections or other forward-looking statements regarding future events or the future financial performance of the company. We wish to caution you that such statements are predictions based on information that is currently available and that actual results may differ materially. We refer you to the documents that the company files with the SEC, including our 10-Ks, 10-Qs and 8-Ks. These documents contain and identify important risk factors that could cause the actual results to differ materially from those contained in our projections or forward-looking statements. This call includes and constitutes the company's official guidance for the second quarter of 2026.

If at any time after this call, we communicate any material changes to this guidance, we intend that such updates will be done using a public forum, such as a press release or publicly announced conference call. We will refer primarily to non-GAAP financial measures during this call. By disclosing certain non-GAAP information, management intends to provide investors with additional information to permit further analysis of the company's performance and underlying trends. For historical periods, we provided reconciliations of these non-GAAP financial measures to GAAP financial measures that can be found on the Investor Relations section of our website at latticesemi.com.

Lastly, we've streamlined our financial reporting to better align with our strategic focus. Beginning this quarter, we'll break out revenue across 2 primary end markets: compute and communications and industrial and embedded. Our consumer business is now included within the industrial and embedded market. For comparability, we've recast all prior period results so you can make a direct apples-to-apples comparison. With that, I'll turn the call over to our CEO, Ford Tamer.

Thank you, Rick, and welcome, everyone, to our first quarter earnings call. Lattice delivered an excellent start to with results that underscore both strong market tailwinds and our disciplined execution against a clear strategy. Our first quarter performance exceeded expectations and our second quarter outlook reflects our expected continued momentum across the business. This is the seventh earnings call since I joined Lattice, and I hope we have now demonstrated that we consistently say what we do and do what we say.

And these positive factors in aggregate provide the foundation for our proposed acquisition of AMI. This acquisition positions Lattice to create the industry's most comprehensive secure management and control platform and enables us to deepen our customer relationships and expand our long-term growth opportunity.

Now turning to our results and outlook. Revenue for the first quarter was $170.9 million, representing 42% year-over-year growth, with strength across all end markets. Our compute and communications end market achieved record revenue, driven by continued momentum in data center AI application. In Q1, 62% of our revenue came from compute and communications products with expanding opportunities ahead.

As Rick highlighted in the safe harbor, we have now merged our industrial and automotive end market with our consumer end market into what we now term industrial and embedded. The revenue from our industrial and embedded end market grew more than 20% sequentially, reflecting improving market conditions and expanding adoption of Lattie solutions.

As importantly, along with increased consumption, China inventory reduced from 3 months last quarter to close to 2 months of inventory on hand, and we expect this trend to continue to under 2 months in Q2. As we anticipated, profitability grew faster than revenue with EPS up 86% year-over-year. These results demonstrate the operating leverage in our model and our ability to scale efficiently as revenue accelerate.

The main trends continue to build across AI servers, networking, industrial automation and emerging physical AI applications. We are seeing accelerated bookings which now support a strong backlog that extends well into 2027. We're also witnessing improved customer visibility and healthy design win momentum across our FPGA portfolio.

Taken together, we're confident that we're in the early innings of a multiyear growth cycle. And our ability to deliver sustained above-market growth for the foreseeable future. Our results also highlight the progress we've made in evolving Lattice into a system-level solutions company. Customers increasingly value Lattice, not just for low power program and hardware, but for complete solutions, spanning connectivity, security, management and control.

As system complexity increases, particularly in AI-driven and advanced computing architectures our customers are giving the highest priority to platforms that reduce integration risk, shortened development cycles and enable faster deployment at scale. These trends continue to expand Lattice's role within customer systems, increased attach rates and drive higher value per design.

We also continue to benefit from our everywhere companionship strategy, positioning Lattice broadly across the ecosystem. Rather than competing with CPUs, GPUs or other processors, our low-power FPGAs enable and enhance them, providing secure boot, power sequencing, platform management, I/O aggregation, sensor bridging and control. This approach allows Lattice to participate across hyperscale data centers, communication infrastructure, industrial automation, aerospace and defense, automotive, medical and emerging physical AI applications, while remaining silicon-agnostic and ecosystem neutral.

Looking to the second quarter, our revenue guidance of $185 million at the midpoint represents nearly 50% year-over-year growth. This underscores our confidence in the accelerating momentum of the business. Our midpoint EPS outlook of $0.44 reflects roughly 80% year-over-year growth. It highlights the powerful operating leverage in our model and differentiated products we bring to market. We maintain a disciplined capital strategy and believe we'll be able to consistently drive earnings growth that significantly outpaces revenue growth. And we are committed to continue to do so.

Turning now to the planned acquisition of AMI we announced earlier today. We are excited to have signed a definitive agreement to acquire AMI, a leader in firmware , orchestration and system-level manageability. The combination of Lattice's low-power programmable hardware with AMI's industry-leading solutions, including [indiscernible], BMC and platform security create the industry's most complete secure management and control platform.

Together, we'll enable customers to accelerate development, simplify system integration and bring increasingly complex platforms to market faster across AI servers, advanced compute, communication infrastructure and industrial applications. Strategically, this acquisition represents a pivotal milestone in advancing Lattice's long-term growth strategy.

AMI firmware is expected to remain processor and silicon agnostic, preserving open ecosystems and customer choice, while Lattice FPGAs provide a complementary hardware foundation reinforcing our everywhere companion chip strategy. We expect this transaction to be accretive to gross margin, free cash flow and EPS on a non-GAAP basis. It also supports our trajectory toward exceeding a $1 billion annual revenue run rate by the end of 2026. We look forward to welcoming the talented AMI team to Lattice and expect this combination to strengthen our system-level road map and long-term growth profile significantly.

Looking forward, we're encouraged by the continued durability of demand across our end markets, the best of customer engagement and the expanding role of Lattice next-generation systems. With a differentiated strategy, a scalable financial model and an increasingly complete platform, spanning hardware, firmware, security, manageability and control, we are confident that Lattice is exceptionally well positioned for the future.

With that, I'll turn over the call to Lorenzo for a comprehensive review of our first quarter results. Lorenzo?

Thank you, Ford, and good afternoon, everyone. We will begin with an overview of our first quarter 2026 financial performance and our second quarter outlook, followed by an overview of our planned AMI acquisition. With a quarter this good and guidance is strong, it is worth repeating some of what Ford said. Revenue reached $170.9 million, growing 42% year-over-year and 17% quarter-over-quarter. Earnings performance was even stronger as Q1 non-GAAP EPS demonstrated the leverage in our model. EPS grew more than 80% year-over-year to $0.41, a 30% increase quarter-over-quarter and above the high end of our guidance.

We expect Q2 to continue this growth trend, and I'll detail our guidance in a few moments. Back to Q1. Revenue growth was driven by a record performance in compute and communications up 86% year-over-year and 15% sequentially. We continue to benefit from strong data center growth as Ford told you.

Additionally, our industrial and embedded end market grew 21% quarter-over-quarter, primarily driven by increased demand in factory automation, robotics and medical applications. Q1 non-GAAP gross margin was a little better than expected at 70%, up 60 basis points quarter-over-quarter and 100 basis points year-over-year.

Our gross margin continues to reflect the value and differentiation our products provide for our customers. Non-GAAP operating expense was $60.8 million, up roughly 8% sequentially and 18% on a year-over-year basis. Much of the sequential increase is from performance-based bonuses and commissions as our revenue and profitability are exceeding expectations. We also continue to invest in order to capitalize on our near- and long-term opportunity.

Our Q1 non-GAAP operating margin expanded 370 basis points to 34.4%, and our EBITDA margin increased 310 basis points to 39.6% both were a little better than expected. Q1 cash flow was impacted by last year's annual bonus payout as well as revenue linearity in the quarter associated with our rapid growth. GAAP net cash flow from operating activities for the first quarter of 2026 was $50.3 million compared to $57.6 million in Q4. Free cash flow trended with operating cash flow. In Q1, free cash flow was $39.7 million, down from $44 million in Q4. We expect a strong recovery of cash flow as we continue to grow.

During Q1, we repurchased $15 million of stock. We ended the quarter with $140 million in cash and no debt.

Now for our guidance. We are targeting closing the AMI acquisition in Q3, so this guidance reflects expectations for Lattice stand-alone. In Q2 2026, we expect revenues to grow in the range of $175 million to $195 million. At the midpoint of this range, this is almost 50% growth from Q2 '25 and 8% over Q1. We expect gross margin to be 70% plus or minus 1% on a non-GAAP basis. We expect non-GAAP operating expense to be between $64 million and $67 million. Most of the growth in OpEx will be in R&D and reflects disciplined investments to drive long-term sustained revenue growth.

We expect income tax rate for Q2 to be between 4% and 6% on a non-GAAP basis. We anticipate non-GAAP EPS to be in the range of $0.42 per share and $0.46 per share. At the midpoint of this guidance, we expect that we would again exceed 80% year-over-year earnings growth as we continue to demonstrate the leverage in our model.

Turning now to the AMI transaction. I am just as excited as for our Board of Directors and our leadership team that we have entered into a definitive agreement to acquire AMI. AMI is a leader in platform firmware, secure boot, device management and system control software. This acquisition represents a strategic expansion of Lattice's capabilities to deliver system-level solutions further accelerating our growth. The total consideration of the deal is expected to be $1.65 billion with $1 billion of cash and $650 million of equity. This is approximately 5.4 million shares based on the closing price on May 1.

We expect the acquisition to be equally compelling from a financial perspective. With AMI, we expect our revenue to exceed an annual run rate of $1 billion by the end of this year. We anticipate AMI's software-centric asset-light model will further enhance Lattice's already strong business model. We expect that the transaction will be immediately accretive to gross margin, free cash flow and EPS on a non-GAAP basis. We will cover our pro forma expectations in more details after we close the transaction.

In closing, we are truly excited about our organic growth and financial performance. We are all very enthusiastic about the opportunity to combine Lattice's strengths with those of AMI.

Finally, PAUSE -- the last team remains focused on execution and taking advantage of the expanding growth opportunities ahead. We are well positioned to drive continued short- and long-term revenue growth, expand our operating margin increased free cash flow and grow earnings faster than revenue.

Operator, that concludes our formal remarks. We can now open the call for questions.

Operator, before we jump into questions, can we introduce with us today, AMI CEO; Sanjay Mike, who has a few remarks. Sanj?

Hello. Thank you Ford. At AMI, our management team, our employees, Board, investors, and we are equally excited to be joining with you and the Lattice Semiconductor team. The strategic combinations with Lattice semiconductor pairs the low-power programmable leader. We're the leader in the platform format and infrastructure manageability for cloud and AI data centers.

Lattice and AMI, we shared a long history, a collaboration and a common vision for secure management and control platform. Now together, we can build on that foundation, extending the reach of Lattice's low-power FPGAs and AMI's trusted platform. While we will maintain the open silicon-agnostic multi-vendor support, our customers value. We also share the same commitment to disciplined execution, strong margins and focus on building values for our investors.

Thank you again. I'm very excited and looking forward to build a great future together.

Sanjoy, great to have you here. Welcome to Lattice. And operator, we can now take questions.

[Operator Instructions]

Our first question is from Ruben Roy from Stifel

Yes. Congratulations on the strong results and outlook and the deal announcement. I guess for it to start. In the press release, you talked about doubling the SAM opportunity here. Can you talk about how we should think about that expansion? How much is incremental addressable opportunity from and their existing firmware installed base versus you talked about combined solution categories, et cetera, that perhaps neither company could attack independently.

Can you help us with that? And I guess as part of that question, the core business is inflecting, particularly on the compute side. If you think about 2026, are you thinking about mix of revenue from servers specifically and maybe AI overall?

Thank you, Ruben. Good question. So yes, we expect our total service available market to us to double from about what we think was about EUR 6 billion currently to about EUR 12 billion jointly with MI over the next 3 to 4 years. And that main increase would come from this compute and communication subsegment. And as you pointed out, the 2 major indicators in that segment are the percent server and the percent. And so these are a good follow-on question on the percent server that has been growing steadily for us from sort of the teens a couple of years ago to this year expected to be and about 38% of our revenue -- total revenue coming from server.

And the second facet to the question in this compute and communications market or subsegment is the percent of our revenue coming from AI. And again, it grew from the mid-teens in '24 to the high teens last year, to where we expect it to be about 25% of revenue in 2026 and AMI plus that is going to be uniquely positioned to be able to provide solutions to customers in this compute and comms market.

That's great. If I could ask you a quick follow-up for Lorenzo. It's great to see that your flagging the deal is immediately accretive. On gross margin and free cash flow, EPS. Can you give us a framework for the gross margin profile? I know it's early, Lorenzo, but any thoughts on the software and firmware business relative to your 70% non-GAAP gross margin run rate at this point, how much of the accretion as you look ahead would be structural versus maybe dependent on synergies revenue or otherwise?

Yes. So I'm going to actually -- that's a great question, Ruben, and we'll get this again, in more detail once we close. The way to think about this acquisition, as Ford said, is it's very strategic. And in the midterm, opens up really significant growth opportunities for us. But the really nice thing about it immediately is thinking about this deal strategically. And looking at the very complementary P&L structure and operating model that AMI has is we are not dependent upon synergies to make the deal accretive.

In fact, AMI's business is very high gross margin. It's higher than ours, and we'll share some more detail on that later. And then they have a different structure, but at the operating margin level is pretty close to ours. And so they generate a very significant EBITDA percent of revenue, and that's close to ours or maybe slightly above ours right now. So if you think about it that way, you can see that there is not a dependency on cost cutting, we'll look at efficiencies through time for sure. But on a go-forward basis, we're able to fund the debt and cover the interest and still show accretion immediately.

Our next question comes from Christopher Rolland with Susquehanna International Group.

So yes, I just wanted to dig in a little bit more on the strategic value of AMI. So I just looked over the website and it seems like they offer firmware, but also like infra management software, would love to know just kind of the cross synergies here between Lattice FPGAs and what you're going to do with this software PAUSE -- and perhaps if you could also talk about the growth rate for AMI. I know you talked about $200 million revs in '26, but that growth rate expectation moving forward would be very helpful for us as well.

Yes. Thank you, Chris. So let me start in trying to explain this. Let me start going to my background where I've done this twice already. So this is the third time that I do something very similar. So it was at Broadcom we were actually second on the switch market share to Marvell. And we -- by the time I left our team have done a great job becoming #1. And one of the key acquisitions we did along the way was a company called Level 7 that provided us with protocol software.

And what we're able to do is to use that to come up with reference design for the ODMs in Taiwan that made it much faster for customers to go to production with their switch system. And so this is number 1 that was successful and very relevant to the discussion we're having. And again, this product called software. Don't think about it as software. Think about it as hardware. So this is very low level stuff.

And then the second one was at Inphi, where we were -- actually, when I joined, we were selling a TIA by the time I left, we were the #1 leader in optical interconnect. And that journey, we decided along the way to have a partnership with Microsoft to deliver a DCI [indiscernible] for Microsoft for between data center, 80 kilometers at the time was colors and some of you on the line know it very well. And that sort of complementary addition of a system or a subsystem skill, if you wish, is module was very critical.

We debated at length whether this was going to be a departure from selling silicon and components. But it made -- and it was about almost 80-20. 80% of the business ends up being these components that we're selling as a platform and 20% of the business was this module that we're saying to -- started with Microsoft, they eventually became the whole market.

And today, it's very standardized across the whole market. But the addition of that module system -- subsystem skills we're very critical in the success because these folks had the folks on the silicon side go to market faster and do a better job. So the history of AMI has 40 years of developing these test cases and very deep knowledge of the whole industry from server to switch to next to, they're the first 1 to be brought up when a CPU gas product. They're the first one to go up when a GPU gas product. The first one to be brought up along with a lot of systems. They're a very key complementary partner to the BMC Board management and controller today.

And we intend to continue to be a very strong partner to all of these Board management controller, the HP, the [indiscernible], the NXP, the others in the market. And so it's an extremely strategic move for us that complements our FPGA low-power FPGA business. And the growth rates are going to be in the high teens, and we expect next year to be actually accelerating. And we do expect to come to solutions to market together that we are growing our revenue faster than you can see we're growing at 40% on the revenue.

And by the way, as you noticed, we grew 80% on EPS and we should be able to grow faster on both revenue and EPS together in the '28 time frame as these solutions go to market. solution or being feeded today in discussion today, we had many discussions with many customers about this solution. They're very excited about it. And it's going to be a very exciting growth. Now we've got a presentation, investor deck on our website -- and if you could turn to it, please, that details the AMI acquisition. And on Slide 5, it shows you the challenges that the data center faced today as you go from managing servers to racks to post the 1 data center this modularity becomes extremely important. AI is adding a lot of complexity.

Uptime these components, GPUs, switch are very expensive. Optum is very key, and there is a huge pressure on time to market and shift left as shown on Slide 5. And then if you jump to Slide 11, Slide 11 shows -- I'm sorry, Slide 11 PAUSE it shows the solutions that we are providing together through these challenges. And so things like rack boot, power and cooling, retrofit and plug and play are going to be solutions that we provide together along with our low-power FPGA and the platform firmware and manageability infrastructure that AMI provides. So very exciting future ahead.

Excellent. And congrats on this deal. I guess, maybe as a follow-up, if you could talk about -- I think you said you guys said inventory maybe was even under 2 months at this point in time. I mean we should have an uplift here. I think we can see it in the guide. But if you want to talk about it more broadly, just as you are no longer burning. And could there potentially even be an opportunity to refill. How are you guys thinking about all this into the future? And next quarter, will we will we be balanced?

Yes, good question. Good question. So look, I mean, we're very excited about it. And what I would say is, this is part of telling you what we're going to do and do what we say. So I told you when we joined -- when I joined now about 1.5 years ago, this is my seventh quarterly part that we're going to bring this under control. When I joined the numbers were closer to 6. And we thought you would be by 3 by end of last year, and we did plan a '25, we got to 3. I told you we're going to be in the tools. We are in the tools. And I told you actually we'll bring it on the 2 and we're on our way to under.

The last time the company was under 2 -- we have 10 good quarters ahead than 1x. And we -- so we may be entering a very strong period here. And you can see our Industrial & Embedded business grew 22% sequentially, which is amazing. PAUSE -- and hopefully more to come. Lorenzo, do you want to add to this?

Yes. No, I think the way to think about channel inventory right now is it's no longer business imperative to bring it down, what we're really working on is keeping the right balance of inventory at distributors across the globe and the right type of inventory so that they can service their customer needs. So I think this is what I would characterize as a nonissue for Lattice going forward. And we're really happy that we were able to manage through this. And now what we get is much greater visibility, much more direct visibility into what end customers are demanding.

So our build is much more efficient.

Our next question comes from Melissa ethers with Deutsche Bank.

Congrats on nice results. It's an interesting deal and Sandoy looking forward to working with you in the future. I guess for my first question, I wanted to touch on the data center side of things. In the past, you guys have given like an FPGA attach rate per server, and it seems like those applications PAUSE that you can use an FPGA for -- in the data center is growing massively and those conversations with engineers are happening live.

We heard Jensen talk at GTC about using more FPGAs in those racks. So can you help us PAUSE I don't know if there's an updated framework that we can think about in terms of FPGA attached in the data center. I'm also really curious on the wireline side in addition to the service side. So any help on, I guess, content in the data center would be helpful.

Yes. Thank you, Melissa. So the couple of trends that I'll highlight and the recent customer visits, I had to a server and they showed me how the unit of RAC has now gone from where they had all in 1 rack to now 4 racks together, a compute track, a networking rack, power rack and cooling rack. And so that's one change that is pretty profound and is going to allow us to increase our content as -- in the comps as well as power and cooling.

The second one, when you go visit the data centers, as you realize now these cooling racks are actually attached by these big pipes coming from the ceiling. It's going to be much harder to change this cooling rack and so this cooling rack may have longevity need that's actually closer to our industrial embedded business that may last for many years as opposed to the faster cadence of moving on the compute side. PAUSE In the presentation that we have about AMI, we are highlighting rack boot on Slide 11, where now some of these -- again, a new application where the club would like to not just for up a server at a time, they'd like to bar up the whole rack at one point.

And we can have some very interesting application RPG and that new applications. And so that's pretty exciting. The third bullet there on Slide 11, you can see that of that AMI application, you can see that people want to go and refit some of their older system for better uptime and better security, better form detection. And again, that's another new opportunity. And so we're finding opportunities all the time. And from a modeling point of view, we still have this forecast of about 16.5 billion server in 2026. We're roughly saying about 3 FPGA per server on overall, okay, that gives you a number of total FPGA. And we're roughly saying our you can calculate the percent -- you could calculate the business, right?

We just gave you today what we don't typically do, but we gave you a breakdown of that server business in Q1. So you could take that 38% I just gave you, and you've got our number for Q1, you can see how much revenue we have and gives you a bit of an ASP. So that gives you all you need, I think, Melisa.

By the way, our ASP is continuing to increase on a per unit basis through this progression. We keep finding more value-added opportunities for our customers.

Yes. So as we said in the past, Melissa, what's helping us number of server increasing AI increasing actually even shorter term, we've seen a big demand increase, not just only from AI, but traditional CPU and storage because of things like cloud another agent type of coding have driven not just AI, but also the position of CPU and storage. The attach rate to continue to find new applications new ASP with the new product, the ASP continues to increase, hence, increasing that server dollar amount.

Perfect. And then maybe just a quick follow-up. I mean, these growth rates seem to be a lot faster than maybe what we were expecting coming into the year. So from a supply perspective, can you just talk about your ability to secure supply? It seems like your customer visibility is increasing, but what about your supplier visibility? Do you have the front end? Do you have the back end is there anything there that we should review that?

Our Senior VP of Operations, Divia Shah has been in the industry for a long time. He lives on a plane and lives at the supplier, and we've had many cause our supplier definitely have our strength, our -- this definitely is straining us and working hard on it, though, and we have been able to secure supply. It comes at a cost. So we are -- the COGS are are challenging, but we're working with our customers and our suppliers to deal with this, and we're in good shape.

Yes. So I think, some other industry players, our wafers are more legacy node wafers and to our supply there is less challenged. The back end is where we see pressure, and we keep expanding our supply chain in that area to provide a diversity of suppliers and additional capacity we're actually beginning to bring our lead times down as we get that expanded supply.

Our next question comes from Tristan Gerra with Robert W. Baird.

Just as a follow-up to an earlier question, is there any step function increase in the content for root trust security with the upcoming Cyber and payment platform? And also is there any potential for event content in data center? Or is that going to be other end markets?

Yes. We're not commenting on specific platform, but our security continues to be a major factor in allowing us to grow our business here. So -- and the second question...

As regarding events and whether there is any data center potential opportunities for the higher density PGA that's coming out.

Yes, absolutely. Absolutely. Can you -- I'm not sure can you be a bit more specific? We -- what -- can you -- I'm not sure we understand the question. Go ahead.

Yes. I was just wondering what type of states you see for event and whether there's any data center applications potentially data path or anything else outside of sort of trust security or even for a root-of-trust security.

Sorry, I'm get an answer. You're asking whether our midterm mid-range FPGA Avon platform has application and data center. And so far, has been mostly our Pre Nexus and Nexus product is what is applicable for data center. As time goes by, I won't may find its way there, but Avancis really focused on our industrial and embedded platform segment.

Okay. Okay. And then just a quick quarter up. Your gross margin installing to to increase again and your lead times have been expanding, which typically is good for I know you only guide a quarter at a time, but what's the potential for gross margin to go higher given the supply constraint and the state of demand versus supply.

Yes. So we've talked about this a few times, especially leading into the year where we thought that the -- we should be prudent about our outlook and gross margin because we saw the supply chain increases coming. And what we have been able to do is continue to work with our customers on ways to offset the cost increases we're seeing. We do also, though, expect that the cost pressure will continue and increase relatively in the second half of this year versus the first half. So we haven't gone -- we just -- given where we are in the year, we are not going to give very precise guidance about the second half of the year. But we are in the range, I said before, 69.5%, plus or minus 1%. This quarter, we happen to be 70%, a little bit higher than that.

But this is going to be the approximate range we see going forward. PAUSE and we will provide you more specific guidance as we get into the second half of the year on how the cost increases are playing out.

Our next question comes from Joshua Buchalter with TD Cowen.

This is Lanny on for Josh. Congratulations on the quarter and I'll extend my congratulations for the deal as well. focusing on the core business really quickly. You've mentioned Lattice is still on track for hitting that over $1 billion run rate in the fourth quarter of this year. Can you clarify if that specifically for the core business? Or is that inclusive of the AMI acquisition as well since I know you've given kind of an estimated revenue for the year.

It is excuse the AMI acquisition.

Got it. And then specifically, kind of was there a follow-up to that? I had a follow-up question afterwards, but if there are additional thoughts.

Go ahead. Go ahead.

Yes. So the real PAUSE to AMI. I'm curious to know what kind of capabilities that gives Lattice that you didn't have before? And if you could talk about AMI's current go-to-market monetization strategy and how that fits in with Vadis' business model currently and then going forward as you integrate together.

Yes. Thank you. So first, as you said, we together expecting to exit the year at this $1 billion run rate at this roughly 40% free cash flow. So you could see we'd be able to delever pretty fast from there. So that's what we think is a beautiful business. The combination is going to be very strong financially. And then from a capability point of view, it gives us much stronger system skills jointly that -- and allow us to bring these solutions to customer much faster.

So it's a capability that we'll be adding. We'll be able to discuss further at the next quarterly call. So when we get in front of you to discuss and guide to Q3, we'll be able to give you a lot more detail on some of the AMI business and go into a bit more of the detail on their financials.

But today, we want to focus on our business, introduce the acquisition, and it's going to take us a couple of months to close expected in early Q3. And then at that time, we'll be able to fully give you all the details on their business.

Our next question comes from Quinn Bolton with Needham & Company. PAUSE -- last PAUSE.

For squeezing PAUSE I'll also add my congratulations. Forward, maybe just a high-level question on AMI acquisition. AMI talks about security, board management PAUSE you've historically talked about similar things for your FPGAs. And so can you just spend a minute, is there any place where the 2 businesses compete Or is it truly complementary -- does the Lattice FPGA root of trust protect the AMI firmware bios as it resides in the servers. Just is there any sort of direct overlap between the 2 businesses?.

Yes. Great question, Quinn. We have been working together since 2019. So it's been Seven years of close collaboration, there is no place where we compete. This is totally complementary. And it's actually very complementary to our customer and should really benefit our customers and partners. And we are committed to remain agnostic. We -- they support all the other silicon partners. And we are partnered to the same silicon partners. So we see it very complementary and beneficial to our -- not only the customers but also the partners. And so it should be very strong, 1 plus 1 equal 3.

And Sanjoy is sitting with me today. He both thanks for 4, 5 and 6, hopefully, and he's laughing years.

Excellent. Excellent. And then, obviously, you guys a great start to the year in terms of revenue growth. I think we came into the year thinking the server business could be up something like 20% to 40% in industrial and now, I guess, the embedded business up 5% to 15%. It looks like you're tracking well above that. I don't know if you're prepared to perhaps talk about growth rates for those businesses given the strong start to the year, but I figured I'd ask.

Well, for the year, it's still early. I would say that the trend that started late last year and has led to our results in the first quarter continues, and that trend is our business continues to book in very, very strong. The customers are continuing to increase their demand. And we are booking out even longer in time. So at this point in time, we have high confidence that our growth this year will be strong, will be stronger than we originally thought at the beginning of the year. Comms and compute as an end market will be the key driver of that for the reasons we all know.

The industrial and embedded though, is recovering, and we saw signs of that in the first quarter. I would say the extremely high year-on-year growth rate might not hold for the rest of the year for the compute and the server, but we'll be pretty pretty strong. And the comms business will be aligned it kind of goes up and down relative to the compute growth rate, but it's still going to be pretty high. And like I said, in industrial and Embedded, we'll continue to grow, but probably not at the rate that we saw in Q1 versus Q4 of last year. They did grow 10% year-over-year, industrial embedded. And that's a pretty good range to think about the year.

And Quinn, right now, as Lorenzo was saying demand is strong for the foreseeable future and bookings well into '24.

Got it. And can I ask a quick just clarification on the deal. Will THL partners be locked up for a period of time post close on that $650 million of equity issued -- or are they free to sell once the deal closes?

Yes, they have a lockup that extends through the year. It's 25% per quarter.

12 months from close.

Yes, 12 months from close, 4 a year, not through the year.

Our next question comes Duksan Jang with Bank of America.

Congrats on the strong results as well. Just following up on some of the earlier gross margin questions. You've been talking about bookings well a strong '27. Backlog is building up. The lead times are expanding. -- that the AMI margins are stronger. So is it possible that the margin structure now is more structurally just higher than before. I think you're touching into untouched territories. I think you haven't really sustained 70-plus before.

But now should we think this is something more achievable at this point?

No, Duksan this opens up opportunities for us that we may have been shied away from before and across the various markets. So we don't intend to go much above that. And can be sure we can. But I think right now, there are opportunities for us that we haven't gone after that we could go after. So it could potentially open up the higher top line.

Got it. And then one follow-up is just on broader competitiveness of the supply chain. And I asked this because clearly, Intel has now divested Altera PAUSE and it's now a stand-alone company. We haven't heard much sense from them, but they clearly have a much different supply chain with the internal manufacturing team. So do you think that gives them a lot more advantage in the supply-constrained environment. And if not, could you explain why?

Yes. Look, we -- our suppliers have been fantastic. We are with UMC, Samsung and TSMC on the fab side and extremely supportive and we have strong assembly and debt partners. We're adding more because this is where the shortages are. And so, so far, we feel very good about our supply chain and our ability to supply. So that's not an issue for us.

Ladies and gentlemen, that concludes the time we had for the Q&A session. I will now turn the call back to the company's Rick Muscha for any closing comments.

Great. Thanks, everyone, for joining us on the call today. We'll be attending the following investor events this quarter. The JPMorgan 2026 Global TMT Conference on May 19 in Boston and the TP Cowen 54th Annual TMT Conference in New York City on May 28.

This completes our call. Thank you very much for your participation, and have a good evening.

Goodbye. Ladies and gentlemen, the conference call of Lattice Semiconductor has concluded. Thank you for your participation. You may now disconnect your lines.

Great. Welcome back, everybody. I'm Joe Moore, Morgan Stanley Semiconductor Research. Very happy to have today the management team of Lattice Semiconductor, Ford Tamer, Lorenzo Flores. I don't cover the company, so if I ask questions that I shouldn't ask, just answer the question I should have asked. It's fine. No harm done. But we've spent a lot of time in the FPGA space, and I like the Lattice story a lot. So Ford, I think you wanted to start with some opening kind of commentary? Or you want to share...

Sure, if you'd like me to. So first, thank you for having us, Joe. And for those of you who don't know, it's Joe's birthday. So happy birthday.

That's getting a lot of advertising, but thank you.

So excited to be here and see some familiar faces. I'll take you through -- I've been at Lattice now for 17 months. So let me take you through the past 17 months and then fast forward to today. I joined 17 months when the inventory situation at the time was 6 months of inventory in the channel. The revenue had decreased from $730 million to $500 million in 2024. We had to cut costs. We took a 14% restructuring.

And at the time, we said what we were going to do. And at the time, we said we were going to grow 25% in '26 compared to 2025. We're going to drive inventory in the channel down to 3 months by end of '25. and we'll keep bringing this down. We said we're going to be growing in 2026 big time in data center AI and Nexus, which is our small range FPGA product line. And then in 2027 is a year of physical AI and Avant, our midrange.

And we basically did exactly what we said. End of '25, our inventory in the channel was down to 3 months. We're back to growth and consensus is now above that 25% number for '26 growth. Margins still very strong for both gross margin, operating profit, EBITDA and free cash flow, and very excited about where the future can take us. Future is bright.

Great. Thank you. So I guess you described Lattice as the everywhere companion chip. Can you describe what that means? Can you describe the philosophy underlying that?

Yes. So the realization when I came in is, look, we don't want to be the chip. The FPGA in the past has been wanting to be the chip, the NIC, the wireless processing, the near-edge AI. We want to be partners to all the major chips that are driving the growth in both data center AI and physical AI. So the MVP, the most valuable player is the GPU or the AI accelerator or the network switch or the NIC or the baseboard management controller, the microprocessor, microcontroller or the different sensors, and we're a companion chip to all of these chips.

We call the GPU, the MVP. MVP is a very important player, the big player on the field, but they cannot win a game, let alone a championship without a team. And we're that team. We're that team providing all these functions to the tunes of -- it used to be tens of FPGA per data center rack to now hundreds of FPGA per data center rack all the way from booting, power sequencing, control management, I/O expansion, security, power and cooling, et cetera, to the tunes of tens of FPGA in a humanoid providing the major control and connectivity to the various sensor, LiDAR, RADAR, and the tens of motor controller there are in these humanoids, robotaxi and various physical AI applications. So we do this.

Being a companion chip doesn't mean you're -- when we first started calling ourselves companion chip, our team was worried, oh, that means we're weak. Well, we're not weak if we're a companion chip. We're actually -- we provide a very powerful function, because we are Switzerland. We provide it across all the various vendors. We provide it across all these functions that I just took you through. We're everywhere. We provide this companion chip function anywhere from the comms and compute market that are growing the fastest.

This year in 2026, it will be over 60% of our revenue. In industrial, automotive and all these new physical AI emerging spaces, humanoid, robotaxi, medical, aerospace, defense, autonomous vehicle and even AR/VR in the consumer side. And we do it across everywhere as well in every application. So if you look at application, we're not selling FPGAs anymore. We migrated the company to be a solution provider. So we sell you security, and it innovates the discussion to a different level. So we go in there and we talk about security. We talk about rack management. We talk about power and cooling.

We talk about PQC, post-quantum cryptography, which is protecting the assets against quantum decrypting in the future. We do it near every image sensor, every LiDAR, every RADAR, every ultraviolet, every infrared image sensing. We do it across a variety of other industrial sensors like temperature, pressure, et cetera. We've started recently to have these partnerships that are driving us into these high-volume applications. So a partnership we discussed is a partnership we have with NVIDIA, where we have a solution that we support NVIDIA with called Holoscan, where we can feed, let's say, 15 different camera feeds into our FPGA, some preprocessing source in there that preprocess data before feeding it to Orin and Thor, making them more efficient.

We've announced a partnership with NXP, where we had a joint meeting of our customers in Europe in January and the VP Strategy from NXP flew from Austin to Munich just to be with us, be the keynote speaker before flying back. And we're putting a partnership where you can have big FPGA, small micro; big micro, small FPGA; medium, medium, it doesn't matter; providing the best solution to the customer. We're announcing many partnerships with image sensor and LiDAR, RADAR, where we make these solutions easier to adopt and easier to install for customers. So that's the companion chip story across every function, across every market, across every application, everywhere to benefit the customer and eventually the investors.

That's great. NVIDIA and NXP probably doing pretty good these days. I hear about this NVIDIA company. The attach rates for FPGAs per server, I think you are expected to move above $3 with ASPs above $4. Can you talk about the longer-term trajectory there? And what ultimately caps that number?

Yes. So what Joe is referring to is the fact that our revenue is growing faster than CapEx. So last year and this year, CapEx in cloud, CapEx for data center is growing 50% to 60%. Last year, we grew our server business at 85%. We grew our communication business at 60%. So growing revenue faster than CapEx. And we're going to continue to do so this year.

And what's driving this is the following: Number one, the number of servers is growing pretty fast. Number two, AI as a percent of revenue is growing fast for us. So we went from, in 2024, AI being 10% of unit, to last year being 12% of unit, and this year 15% of unit being AI servers. And just to put things in perspective, AI drives a larger percent of revenue. So last year on 10% of unit for revenue, we drove 20% of the revenue dollar. So hence, providing higher dollar per unit.

Number three, attach rate, we keep finding the new applications. So security has been a very strong application for us, where we started with Root of Trust. We added attestation. We're adding post-quantum cryptography. We keep growing in security. We found this new application on cooling, for example, that we recently adopted and that's, again, growing very fast. So the attach rates across applications are growing. The needs are continuing to grow, hence, bigger FPGA and hence, bigger ASP. So overall, you could see, if you put all these together, the multiplicative and revenue grows faster than CapEx, which is already growing at a pretty fast rate. So that's what's driving us.

Okay. Yes, I guess how do you think about that longer-term trajectory then with hyperscaler CapEx? This year, we're sort of looking at the hyperscalers cash neutral. I'm pretty bullish spending will still be up a lot next year, but how do you think about that dynamic affecting you guys?

Yes. Look, we're booked for the year and through mid next year. So I mean, bookings are extremely strong, and we see them to be strong for the foreseeable future. So we don't see a slowdown right now. We're up and to the right. And this year has been the year of data center AI for us. Next year is going to be the year of physical AI, where you've got a whole bunch of these new applications I just described that go to market. And it's interesting, some of the discussion we've had at this conference, Joe, is can humanoid be bigger than data center, because people are saying, look, we've got as much content in a humanoid as we have in a rack. And so 15 million server, 3 million racks, the numbers for humanoid could be anywhere above that.

I was listening to the Figure AI CEO on the Moonshot podcast. He's predicting more than 1 humanoid per person. So that would be 8 billion humanoid. So whatever the number, even if we're wrong by 2 orders of magnitude, 80 million is still a pretty big number. So all of the new physical AI applications coming to market next year could be super interesting. And we're seeing real productivity gains. I mean, us at Lattice are using Claude for our own development. We're seeing tremendous productivity gains. We have all the tools, Cursor or Claude. But Claude over the past few months has been amazing. And so you're seeing this accelerating. These new applications are coming to market and really the rate of change accelerating in a pretty big way.

I think the irony of all of this is that your 2 biggest competitors were bought by compute companies with the idea of moving FPGAs more into the compute space and that you're having all the success. I mean, is that in a way because of that, because they're focused on that central processor and trying to use FPGAs for that?

Look, I mean, our success is directly related to where the sweet spot for FPGA is. The sweet spot is in the low to mid-range FPGA non-SoC. That's exactly what we do. We're focused on this single mission. And we're going to take the company in the solution space. So you're seeing us invest in solutions. We've got people that understand these solutions across all these markets, and solutions where we're going to take the company and grow it in a big way.

And when you talk about solutions rather than FPGAs, I mean, it's still, at its core, a programmable FPGA, right? That you're orienting around those solutions.

Absolutely. At the core, we're still providing future-proof. So let me give you why FPGA. So security, the algorithms are changing very fast. You want to be able to change -- like the old algorithm like AES-256, when quantum shows up, it is sticking like a gum to paper. It's like no protection whatsoever. So you've got a huge imperative to move to this post-quantum cryptography like today. And these algorithms for post-quantum cryptography, PQC, are changing still very fast. So we've got these Lattice algorithms, and no pun intended, these are the name of the algorithms. These are 3D algorithms, but they're changing.

So having the ability to change these algorithms in the field for future proof is very important. When we talk about humanoid application and physical AI, FPGA is great, because you need low latency. So if the roaming humanoid needs to stop, it can stop very fast. You need to have parallelism for fast performance. You need to have accuracy for applications such as screw driving, bin picking, welding, you need high accuracy. You need connectivity to be -- we have EtherCAT that can connect all these motors and vision sensors together for perfect synchronization. You need futureproofing again. So the FPGA provides all these functions. For all these different various applications, FPGAs are critical, but you also want a low-power FPGA that's going to be small size and cost-effective and fast boot. And by the way, you want to go talk solution. You don't want to talk FPGA. You want to talk security. You want to talk vision. You want to talk motor control.

Okay. Very helpful. Lead times are stretching out. Can you talk about what you're seeing there and how that's impacting your business? And any opportunities around that to see higher pricing, anything like that?

Yes. So the lead times are indeed stretching out for us and for everybody else, obviously, reflecting the strength of demand. And what we've been doing with our customers is working with them directly and with our channel partners to say, when do you really need product, right? So what we've done with the extended lead times is try to work with the customers to get the bookings through the year. And as Ford said, we're seeing them in through the middle of next year, past in some cases, so that we can stage our supply chain and meet the customers' end demand.

So the way it's helping our business is in planning, so we can actually deliver the parts. The visibility that we now get from a tighter channel, right? We dropped the supply in the channel down to where we have much more direct visibility and are able to target supply to end customers that really need it. That's really helping our customer relationships. And then on the pricing side, we have a different approach than some other people in the industry. We have taken a very long-term view of our strategic customer relationships. We intend to be with our customers for a long time.

And what we're working with them on is we have had a consistent mature product price increase type of program. And we're keeping a balance between potential cost upside in the supply chain with how we're approaching pricing. And so we're comfortable that lands us in the gross margin we've guided through the year. So we're definitely not doing a short-term pricing move to optimize margin in any quarter. We're looking all the way through this demand cycle and then through the product ramps that Ford described in the opportunities we see in front of us.

Can you talk about other puts and takes around gross margin? It seems like product mix should be favorable for you. You said the pricing isn't changing, but it seems like other factors may.

So it's a very dynamic situation. So I didn't say pricing wasn't changing. I just said we're taking a balanced approach and long-term approach to it. Obviously, we're on the lookout and constantly in communication with our supply chain on where we -- so that we are able to sense where price increases from them will come from. So we're trying to manage our cost. We took steps starting about 6 months ago actually on both the cost side and on the supply availability side, because our Head of Operations saw some of this coming. And so we're trying to get ahead of that. But we still have to be realistic that it's a factor in our industry, right, the supply chain constraints and possible cost increases.

So we started to manage that. We do have the new products. We do intrinsically have a high-value product, right? And our customers do want to work with us and use our product because of the low power, high-performance characteristics of it and the suitability in all the applications for its data. So we are trying to manage that side of the equation on the pricing side to keep it in balance. But we do have great product, and we do have strong demand, but we're all very aware of the possible supply chain issues.

Great. And on that new product revenue, I think you're talking about a mid-20% of total revenue this year. Can you talk about the progress of that and then the Nexus versus Avant timing?

So we think that 20% will get to the mid-20% this year. So good progress and that growth could continue to grow. Last year, we grew 70% year-on-year. This year, we think we grow 60% or more year-on-year on the new product. Nexus is really 2026 with Avant really taking off in '27.

Okay.

There's a question there.

Yes. I'm sorry. I'm just sort of... If you were to explain somebody not very bright like myself, how does an FPGA fit into the broader hardware world? And I know you keep making the point that there's a broader role for it in this new AI, GPU, et cetera, universe. But how does it slot in? And why is the opportunity so much larger relative to where we were in the past?

Yes. No, it's a very good question. So at a high level, if I were to draw a continuum, right, from a -- the most flexibility is a processor, right? So you'd have microprocessor, microcontroller can give you the most flexibility programming in higher-level languages, C, TensorFlow, PyTorch, all the way to the other end of the scale in ASIC that is going to be fit for that purpose. So it's going to be a GPU, a switch, a NIC that is going to require hundreds of millions of dollars to produce, but the best power, the best performance for that application, right? So that's the continuum, right? We're in the middle.

And think of an FPGA being hardware-like performance, right? But the issue we have is you have to program it at hardware-like level. So we don't have the same appeal today just because the microprocessor, microcontroller, you can program and see, you've got many more people that can program in this micro versus FPGA.

But if you look at these applications such as the application of data center we're providing to the humanoids, I talked about the robotaxi, aerospace and defense, medical, all the applications require hardware-like performance. So they require the performance that we can provide. And you don't want to go spend hundreds of millions of dollars to develop an ASIC, which if you've got an application like that's really high volume, such as an AI accelerator, switch, NIC, board management controller, may as well go spend tens of millions, hundreds of million dollars to do an ASIC.

We can provide you not quite the same performance as an ASIC, but quite a good performance, much better than a microcontroller. We provide you performance. We provide you latency. We provide you deterministic, we provide you accuracy. We provide you much lower power. We provide you many more benefits than you'd have in a microprocessor, right?

So we're in this in between. Why is today FPGA becoming more important, okay? So the cost of developing these ASIC continues to grow. The time it takes and the size of the team it takes to develop this ASIC continues to grow. And you've got applications such as the one I just described that require that hardware performance, that require this unprecedented, you want to have this -- you need this performance, you need this power. And yet you don't have tens of millions, hundreds of millions. FPGA is a great solution, right?

So as the cost of ASIC continues to grow, the cost of putting these functions on a 2-nanometer, 3-nanometer advanced node continues to grow. So I look at this AI chip and this big chip that takes over the whole reticle as much -- I've got as much logic I could put on this chip. And would it make sense to offload this power sequencing, this control, this boot to it? We're in 65-nanometer, 28-nanometer, we're in these back processes that make this function much more cost effective. So it makes more sense for me to offload these functions to this FPGA.

There's more and more pressure on system time to market. So NVIDIA, AMD are driving to 1 year between different generation of GPUs, okay? Elon Musk comes in and says, I want to go 9 months between my xAI type of GPUs, right? So there's more and more pressure on going to market fast. And so a lot of these functions that typically you'd have the time to go put in an ASIC, you don't have the time anymore. You're going to have to say, okay, FPGA is good. I mean FPGA gives me the performance, gives me the cost, gives me time to market. And by the way, it gives me flexibility. So the security algorithm, as I said, keeps changing.

What FPGA allows you to do, FPGA allows you to program in the field in the future. So today, you program for a certain algorithm for security. The bad actors are going to change those algorithms on you. They're going to come back tomorrow and say, okay, hey, I just changed this encryption. Now I can't know how to decrypt it. Okay, well, guess what, I'm going to be a step ahead of you. I can go change this in the field in my FPGA. So many, many factors.

The ease of use AI will allow us to close the gap. So look at Cursor, look at Claude being able to program at a higher level. AI will allow us to close the gap and make FPGA broader and more available and more interesting. So we're seeing adoption across the board. But yet, you want adoption with low power, low cost, fast boot time, small area, and that's what we provide. So all of a sudden, we're seeing this thing mushroom. I mean people in the meetings today saying, "Hey, we've never seen FPGA grow at these rates. So this is all the factors that are helping us grow the company at this rate. Makes sense?

Yeah. We've another question there.

Could you elaborate on the humanoid opportunity in terms of how you provide value to that new market? And what you see as the time frame over which it will evolve into a very big market?

Yes, yes. So look, first, we start with vision. So humanoid obviously has to see where it's going. It needs vision, and you've got many cameras. Sometimes you've got a LiDAR or RADAR to guide that humanoid, right? And near vision, you need to have an image processor that allows you to do this image processing, to do the control, to link up all of these different vision pieces together, and we provide that in FPGA.

Next, you've got motors. So there are many podcasts that talk about each of the hands having upward of 20 motor per hand. You've got motors all over the torso and the shoulder and the legs. So it could be anywhere from 45 to 70 motor controllers, and we could provide the motor controlling support control. You think about the humanoid having 70 motors. Well, the 70 motors better be synchronized and work together at the same rate. We provide connectivity to provide this EtherCAT that allows you to synchronize all these motors together. So a lot of value we provide and security, by the way.

Time frame. He asked about the time frame.

Time frame. Look, I mean, it depends who you hear. Some people have -- there's probably a 2x difference in projection of how this thing grows, but it starts next year. It starts in 2027.

Can you talk about -- when you see this kind of visibility out to mid next year in some of these markets, can you talk about the risk of double ordering, anything like that going on?

Yes. So I talked earlier about how we're utilizing the extended lead times to get more visibility and expand a little bit more on what I said about tightening the channel inventory. We are really applying rigor to overall demand we are seeing to make sure that customers are aligning their orders from us with what they see as demand and, in fact, trying to do our best to look at the rest of the kit that they have in their BOM, so that they don't have our parts sitting on the shelves and no memory, for instance.

And we're then working with our overall sales force in different areas to see what they are hearing on the ground in terms of the end demand at that. And I think there's a lot more rigor that we're applying as we are managing through this with the priority being supplying our end customers with the parts they need and doing additional checks on it. But is there some double ordering? Maybe. But because we're allowing customers to book out in time and giving them confidence we will give them the supply, we think that tendency is diminished. And by tightening the channel, we try to decrease any buffer that's held there as well. So it's not going to be perfect, but we're doing our best to not be surprised by it.

Great. And then the other question we get, when FPGAs start growing into high-volume spaces like AI, is ASIC replacement risk, ASSP replacement risk, can you talk about that?

Yes. I mean, look, if we were a really expensive part that has a lot of power and very visible on these boards, we'd be at risk of being replaced. Today, we're into very reasonable prices, and we have so many applications and across so many vendors. If a hyperscaler wants to come in and design their own to replace us, they're going to miss -- I mean, we learn across all of the hyperscalers, the Neocloud, the enterprise, the server vendors, the ODMs and come up with these FPGAs that are optimized across all these use cases, across all these vendors, across all these functions. And so many use cases, it doesn't make sense to replace us. It's not like we've got one really expensive high-power, high runner that you can easily identify and say, I'm going to replace. I mean this is a widespread, many tens of use cases per rack, per industrial application. So...

Right. In the BOM of our customers, we are at the small tail of their cost. And as Ford said, it's a multitude of functions, so they would have to contemplate doing a multitude of ASICs, which doesn't seem to make economic sense.

Yes. Okay. Helpful. Obviously, a lot of growth on your plate here. But can you talk about M&A strategy? And is there anything that you would think about where you need to add capabilities?

Yes. So look, we've done actually a few small tuck-ins that we haven't discussed, because they're small, like a few either acqui-hire or IP licensing or small tuck-ins to mostly around IP software tools and solutions. And over time, I think we would like -- I think we feel very strongly about the organic growth that we have, it's very strong organic growth. So we don't have to do an M&A. Now if the right one shows up and aligns with what I just described on the vision that we just outlined, it would make sense for us to move forward. But we're going to keep a high bar and be disciplined on how we do it.

I mean do you need scale in this business given the size of your 2 bigger competitors?

Look, scale always helps. I mean, but we can get scale via acquisition or we can get scale via organically growing it. The cost per engineer from the time I joined to now, we drove our cost per engineer from $150,000 per head to $100,000 per head. And so we're growing very fast in India and Penang and Philippines. And so we're looking to grow in Taiwan as well. So we're growing in all these geographies and adding a lot of people organically. So ideally, yes, we can do an M&A, but that's not required.

Yes. The way we have thought about scale on the organic side is exactly as Ford described, Joe, which is we actually lowered OpEx between '24 and '25. We have more engineers. And we're helping the productivity of those engineers, not with just the IP acquisitions, like Ford said, but also we deliberately invested in infrastructure. It's an unnoticed unglamorous-type of investment, but it's really helped the productivity of the engineers. So we're getting to scale that way.

Very helpful. We have 2 minutes left. Anybody has any questions?

Yes. So the question is, can we give a rough breakdown of our Industrial segment.

What I would tell you, our Industrial segment could go back to where it used to be. So I think if you think about it, we do break down the industrial and comms and compute. So last year, industrial was about $195 million of the $523 million. So we did $523 million, $195 million was industrial, $290 million was comms and compute and the rest was consumer, okay? And this year, we expect the comms and compute to become over 60% of our revenue from sort of 56% to over 60%, and industrial consumer to be roughly flat and industrial to be the remainder. So you could see what that growth is.

And next year, we expect industrial to go back and increase faster. And what happened to us in industrial, we built this inventory, and we've been bleeding inventory at the rate of about 3 weeks per quarter for the past 5 quarters. So we're now through that, and we're down to inventory level that makes sense. So we're now going to be shipping our revenue at natural demand, which will help this year. And next year, we do expect an inflection in industrial because of all these new applications and products that I discussed.

So at a high level, we are committed to be a 2-legged stool to start. So leg #1 would be this comms and compute that's very strong today. Leg #2, industrial that's going to get stronger and continue to get stronger next year. And we want to develop a third leg over time, and that could be around the solution angle. So eventually, we're going to have a more stable 3-legged stool to continue to grow the company.

Yes. I'm not sure underneath your question might be what's in industrial. So we actually call it industrial and auto, and you probably derived from the commentary, auto is not a very big part of our business. We don't have a lot of exposure there. But within auto, we have classic industrial factory automation type of applications. We have medical applications. We have defense as an emerging application set for us. We have industrial power management, a bunch of things like that. So it's a very broad category for us. We don't break it out any further. But despite the fact that's named industrial and auto, auto is not a significant part of our overall business.

Okay. We'll wrap it up there. Thank you so much.

Thank you. Good to see you.

Greetings and welcome to the Lattice Semiconductor Fourth Quarter and Full Year 2025 Earnings Call. [Operator Instructions] As a reminder, this conference is being recorded. It is now my pleasure to introduce your host, Rick Muscha, Vice President of Investor Relations. Thank you, Rick. You may begin.

Thank you, operator, and good afternoon, everyone. With me are Fouad Tamer, Lattice's CEO; and Lorenzo Flores, Lattice's CFO. We'll provide a financial and disreview of the fourth quarter of 2025 and the business outlook for the first quarter of 2026. If you have not obtained a copy of our earnings press release that can be found on our company website in the Investor Relations section at latticesemi.com. I would like to remind everyone that during our conference call today, we may make projections or other forward-looking statements regarding future events or the future financial performance of the company.

We wish to caution you that such statements are predictions based on information that is currently available and that actual results may differ materially. We refer you to the documents that the company files with the SEC, including our 10-Ks, 10-Qs and 8-Ks. These documents contain and identify important risk factors that could cause the actual results to differ materially from those contained in our projections or forward-looking statements. This call includes and constitutes the company's official guidance for the first quarter of 2026. If at any time after this call, we communicate any material changes to this guidance, we intend that such updates will be done using a public forum, such as a press release or publicly announced conference call.

We will refer primarily to non-GAAP financial measures during this call by disclosing certain non-GAAP information, management intends to provide investors with additional information to permit further analysis of the company's performance and underlying trends. For historical periods, we provided reconciliations of these non-GAAP financial measures to GAAP financial measures that can be found on the Investor Relations section of our website at latticesemi.com. Let me now turn the call over to our CEO, Fouad Tamar.

Thank you, Rick, and welcome, everyone, to our fourth quarter and full year 2025 earnings call. At the end of 2024, we told you what we're going to do. And over the last year, we did what we said. And we accomplished that by delivering on our commitments, stabilizing the business, normalizing channel inventory, improving execution and positioning Lattice to capitalize on 2 of the most powerful secular trends shaping our industry, data center, AI and physical AI. As we begin 2026, I'm most excited to see that low-power FPGAs are being widely adopted at an accelerating rate, becoming the everywhere companion chips with the Super Bowl still fresh on everyone's mind, I will use a sports analogy.

The primary processors, GPUs, customer accelerators, CPUs and MPUs are the system's most valuable players or MVPs. These MVPs are powerful, but cannot win a game that had on a championship without a team and Lattice is that team. We provide the FPGAs those everywhere companion chips that perform many of the critical system functions. These include boot, power sequencing, security, control, IO expansion Board and rack management, League detection, power and cooling, bridging sensor aggregation, sensor fusion, preprocessing and many other valuable system functions.

We do this across both data center AI and physical AI. We do this across all major markets. communications, compute, industrial, automotive, aerospace and defense, medical and consumer. And we do this across some of the world's fastest-growing applications. security, track management, communication, quantum cytography, humanoid, industrial automation, logistics, robotaxis, space and AR/VR wearables. Those powerful companion chips provide pervasive interoperability across all these vital functions, markets, applications and diverse suppliers.

You can consider Lattice as Switzerland for data center and physical AI applications. The other big benefit of our companion chip strategy is that it drives sustainable diversified growth across all markets and applications, which is helping us to deliver on our financial targets. And this companion role becomes foundational as AI workloads push system-level complexity higher and higher and drive to faster time to market. As AI servers disaggregate into processor boards, networking cards, security cards, power and coining modules, storage and other specialized blades, FPGAs show up everywhere in the data center. We are also seeing the same trend accelerating in physical AI. Intelligence is moving closer to the sensors where data is created. A single robot can have multiple vision sensors, including image, LIDAR, radar and infrared requiring fusion aggregation and preprocessing. In addition, humanoid robots could have dozens of motors requiring high precision and low latency control. Companionship FPGAs sit besides those sensors and actuators to synchronize high-bandwidth vision streams with real-time motion and delivered deterministic responses. And instead of software running on a microcontroller, Lattice FPGAs do this in hardware, making it easier to guarantee the same cycle accurate responses every time.

Kindly in some cases, FPGAs can also serve as the primary compute such as signal processing, real-time networking and what has become to be known as far edge AI.

We, at Lattice , defined far edge AI as near-sensor contextual AI with tiny or small self-contained bonds, and our momentum is building. For example, we recently won a design in a human machine interface, or HMI, Industrial robotics, and we are seeing the pull for applications under 1 tops and under 1 watt. We came out of our recent global sales conferences with strong momentum, reinforcing our optimistic outlook for 2026. We are seeing accelerated design win momentum in both data center AI and physical AI, and we are excited about the future ahead with our Tier 1 customer deployment.

Now let me turn to the numbers. In Q4, we delivered $145.8 million in revenue, up 9.3% sequentially, our strongest sequential performance in 7 years and up 24.2% year-over-year. Full year revenue of $523.3 million was in line with expectations. Looking ahead, our Q1 revenue guidance of $165 million at the midpoint, representing over 37% year-over-year growth reflects our confidence in a strong recovery and accelerated momentum.

Our Q1 EPS guidance of $0.36 at the midpoint represents nearly 65% year-over-year growth as we expect to continue to deliver earnings growth that is faster than revenue growth. As you can see, we've got a lot of positive operating leverage. We are operating on mature loans, which means our capital spending is more reasonable than competitors on advanced nodes, and this allows us to drive significantly faster EPS growth than revenue growth. New products remain a key driver for our long-term growth. In 2025, new product revenue grew approximately 70% we remain on track for new product revenue to reach the mid-20% range as a percent of total revenue in 2026 entering the year with strong momentum as NEXUS and Avant adoption continues to broaden.

Given the scale of the opportunities ahead of us, Lattice is investing for the future accordingly. Our 2026 slogan. Go Big Be Great reflects our ambition and our commitment. We are making investments across silicon, software, systems, operations and infrastructure to support growth at scale and to extend our leadership in small- and mid-range FPGAs.

In summary, 2025 was a year of disciplined execution and meaningful progress. We stabilized the business, built tremendous momentum in data center applications, advanced our product and software road map and significantly improved operational performance. including the normalization of channel inventory. We entered 2026 with high confidence. That confidence is supported by a strong backlog, durable data center and demand, industrial market returning to growth, expanding companion use cases and continued new product ramps. Our focus remains on differentiated innovation deeper customer engagement and delivering long-term shareholder value. With that, I'll turn the call over to Lorenzo for a comprehensive review of our fourth quarter and full year results. Lorenzo?

Thank you, Fouad, and good afternoon, everyone. We will begin with a brief overview of our 2025 fiscal year performance, including our fourth quarter, followed by our first quarter 2026 outlook and our current framing of fiscal '26. For the full year 2025, we are pleased to report that Lattice delivered on expectations with revenue, gross margin, operating profit and EPS all in line with our outlook for the full year, including the fourth quarter. We did this while completing our transformation and investing in areas that we believe support long-term profitable growth. Our full year 2025 revenue increased 2.7% to $523.3 million, in line with expectations. Growth was driven by our platform successes across communications and computing with revenue up 28%.

This growth was partially offset by an 18% decline in revenue in industrial and automotive, which was expected as we successfully normalized channel inventory throughout the year. Our non-GAAP gross margin continues to reflect our value proposition to our customers and expanded 190 basis points to 69.3% in 2025. We improved revenue and gross margin in 2025, even as we reduced non-GAAP operating expense approximately 1% to $213.5 million.

As we completed our restructuring, we made targeted investments in talent, infrastructure and technology to give us a more robust and efficient platform to drive long-term growth. As a result, in 2025, our non-GAAP operating margin expanded 340 basis points, and our EBITDA margin increased 320 basis points to 35%. And we delivered non-GAAP EPS growth of 17% to $1.05, demonstrating leverage in our business model by growing earnings faster than revenue.

Other metrics improved accordingly. In 2025, GAAP net cash flow from operating activities increased to $175.1 million, up from $140.9 million in 2024 with GAAP operating cash flow margin improving to 33.5%, up from 27.7% in 2024. Free cash flow in the full year 2025 was $133 million with a 25.3% free cash flow margin up from $120 million and 23.5% in 2024. Focusing specifically on Q4, our performance started to ramp, we will continue into '26. In Q4, revenue hit $145.8 million, an increase of 9.3% quarter-on-quarter and 24.2% on a year-over-year basis. Growth was driven by a record performance in Communications and Computing up 25% sequentially and 60% on a year-over-year basis. We are clearly benefiting from the exceptionally strong data center growth as Fouad discussed. Q4 gross margin was 69.4%, slightly down from Q3 on a non-GAAP basis. Our gross margin, even as our mix shift continues to reflect the value and differentiation our products provide for our customers.

Non-GAAP operating expense was up to $56.4 million, up roughly 5% sequentially and 7% on a year-over-year basis. As Fouad and I have stated previously, we see significant near and long-term opportunities and are investing to expand our future leadership in the small and midrange FPGA markets and our companionship program. Our Q4 non-GAAP operating margin expanded 170 basis points to 30.7%, and our EBITDA margin increased 90 basis points to 36.5%. Q4 non-GAAP EPS grew 14% quarter-on-quarter to $0.32, in line with our guidance. GAAP net cash flow from operating activities for the fourth quarter of 2025 increased to $57.6 million, up from $47 million in Q3 and with a GAAP operating cash flow margin of 39.5%, up from 35.3% in Q3. Free cash flow in Q4 was $44 million with a 30.2% free cash flow margin, up from $34 million and 25.2% in Q3.

This remains a focus area for Lattice, and we expect the trend of increased free cash flow margin to continue. We are achieving these improvements while strategically investing in CapEx in support of R&D and operational improvement projects. One last point on results. We achieved our overall target level of channel inventory and are well positioned to benefit from growth in all the end markets we serve.

Now let me turn to capital allocation. Our balance sheet remains strong as our operations have improved our cash flow generation and we remain debt-free. We have ready access to capital if we need it. This leaves us well positioned to navigate macro uncertainties and invest for future growth. Given our balance sheet strength and our business model, returning capital to shareholders remains a key component of our capital allocation strategy. For the full year 2025, we repurchased approximately 1.8 million shares for $100 million of the company's common stock. As we completed that repurchase program in Q4, our Board of Directors authorized the company to repurchase an additional $250 million of its outstanding common stock.

Now for our guidance, which has our typical detail for Q1 2026 and additional commentary to help frame expectations for the full year 2026. For both Q1 and the full year 2026, our outlook reflects the exceptional strength we are currently seeing in demand. In Q1 2026, we expect revenue to grow into the range of $158 million to $172 million. At the midpoint of this range, this is 37% growth from Q1 2025 and 13% over the prior quarter. Gross margin is expected to be 69.5%, plus or minus 1% on a non-GAAP basis. Non-GAAP operating expense is expected to be between $59 million and $61 million. Most of the growth in OpEx will be in R&D and reflects disciplined investments to drive future growth.

Income tax rate for Q1 is expected to be between 4% and 6% on a non-GAAP basis. Non-GAAP EPS is expected to be in the range of $0.34 and $0.38 per share. Regarding the full year, we begin 2026 excited about the catalysts driving multiple growth opportunities ahead of us. I would like to share with you our financial framework for the year. We have stated previously that we are highly confident that we will grow at least 20% year-over-year and the start of the year indicates that we have improved visibility to growth above that. We plan to update you with more specific guidance as we move through the year.

Our gross margin for the full year 2026 is expected to be in the same range as we are providing for Q1 and with perhaps some fluctuations during the year as customer mix varies. On OpEx, we continue to see significant opportunities to expand our leadership in small- and mid-range FPGAs. These investments primarily in R&D, drive the increase in OpEx in Q1. We expect another increase in Q2, but then slower growth in the second half. We have a tight grip on the nose levers that control OpEx and will adjust as necessary to hit our profitability objectives. Our income tax rate is expected to be between 4% and 7% on a non-GAAP basis. With channel inventories now normalized, revenue in 2026 should more closely track consumption setting up a strong environment for growth in 2026 and beyond. We will continue partnering with our channel to support end customer demand while reducing channel inventory where appropriate.

For internal inventory, we are deliberately building inventory to support this growth environment because we see such strong growth potential and our products have long life cycles, we see a low risk of obsolescence. In closing, we remain focused on executing our strategy and making investments to strengthen our leadership in small- and mid-range FPGAs. We are highly confident in driving revenue growth while growing EPS at a faster rate. Operator, that concludes our formal remarks. We can now open the call for questions.

[Operator Instructions] One moment, please, while you poll for questions. Thank you. Our first question comes from the line of Ruben Roy with Stifel.

Fouad and Lorenzo, congrats on a strong finish to the year. I guess for the first question, Ford, if you could walk us through, obviously, really strong server growth and you've talked about sort of the dynamics in the data center and I guess, the MVPs out there, thinking about how you're looking at that business going forward. We've heard a lot about server CPU strength from some of the CPU vendors. There have been some supply constraints in the mix. So maybe you could frame the -- those dynamics as well as your growth relative to unit volumes versus dollars of content per server.

Yes. We do see demand being strong for the foreseeable future is driven by the CapEx growth is driven by our attach rate is driven by the increased ASP is driven by the increased number of applications that we're finding in these servers. And so at a high level, as we look at server units growing from, call it, the 15.3 million units total to, call it, 16.5 million units in 2026. Our attach rate has been steadily going up every generation. So in the 2024 time frame, we were in the mid-1s. And as far as the tax rate -- last year, we're in the mid-2s and this year, we'll pass 3 units per server, FPGA units per server. So you multiply this by the number of server units or attach rate growing and our ASP is growing as well from the $3, call it, to above 4%.

You put all this together, and you could see that this is driving an overall very nice increase in that business.

Great. Lorenzo, maybe just a follow-up on the guidance for Q1, very strong guidance. Can you walk us through sort of assumptions by segment? It sounds like we're finally seeing some recovery in industrial and automotive. And obviously, the communications and compute continues to grow, but maybe you could parse that out for us, please.

Yes. I think in aggregate across the business, we're just very, very positive about the trends we're seeing underneath it. Of course, the Holman compute areas of our business are the leaders in our strength. But as we said, we've gotten our channel inventory down toward our target, and we expect to see growth there as well.

Our next question comes from the line of Melissa Weathers with Deutsche Bank.

Congrats on a nice outlook. I know it's taken a lot of work to get here. I guess I'm going to try on the revenue side. It doesn't sound like you guys are super willing to guide to give full year guidance for 2026, but last quarter, you indicated comps computing could be up 20% to 40%, I think, in industrial up 5% to 15%, if I remember right. You're starting up the year a lot stronger than that. So just in terms of magnitude, like is there any other color you can help us with in how to think about which of those grows faster and maybe where total revenue growth could shake out for '26.

Yes. We're very strong. Our conviction has increased on the demand in both segments, Melissa. So first, thank you for the question. Yes. The demand environment is much stronger now than it used to be 3 months ago across both the Comms and Compute and industrial markets. So in comes a compute, as we said, we are very well booked for the year and now booking to 27, so great visibility. And the industrial automotive, where our shipping to the true demand, the Inventory in the channel is under 3 and we'll be in the 2s in the Q1. So we are confident in the growth in both segments.

Got it. Maybe as my follow-up, you spent a lot of time in your prepared remarks on the physical AI opportunity. I think the companion piece and the the attach rate in data centers is pretty easy for us to follow, but it's a bit harder to -- for us to kind of figure out how big the physical AI opportunity could be. So I don't know if you want to frame it in terms of the TAM or some kind of unit framework, but any way for us to help -- to help us size how big this fiscal AI piece could be? Because it seems like you're seeing a lot of momentum there.

Yes. We're seeing tremendous momentum there along with this companionship strategy. So we are very strong with companies like NVIDIA that have -- we've publicly announced together the [ Honasan ] design reference design, where we feed different video streams into the FPGA and into the NVIDIA processors. We've made some great progress in our companion chip strategy with NXP and along with the microprocessors. We had an event in Europe in January. We're the top industrial and automotive and telecom and aerospace defense companies in Europe joined us an NXP and a joint event where we launched the year -- we're making progress with quite a few other partners on the sensor side, on the analog side and others in the microprocessor side. So stay tuned on those.

The potential expansion is vast because we are gaining share in markets where we were not big players. The best example is aerospace defense that has gone from very little content in. So we expect a very big year coming in, in '27 across all geographies, U.S., Europe, Asia, all geographies are observing some very strong design wins in aerospace defense with our new rad hard and red tolerant designs. In industrial robotics, we continue to grow in industrial robotics and AMR as well as humanoid, the humanoid opportunity is vast. We are winning across the board in vision. So we've done very well in vision and humanoids.

And we have gotten 2 marquee design and motor control, and we see this as a beachhead as a teaching customers on how we're going to make progress because the latency determinism, parted performance, accuracy of an FPGA hardware-based solution is vastly superior to the alternative. And so we're very excited about what humanoid can mean for us as a market going into '27. We've also made great progress in robotaxis. So we have a win in robotaxi that we're, again, very excited about going into '27. We're making progress in medical and telecom across the board in physical. So I think it's hard to give you as clear of a metric as we can on the server because the applications are so much more diverse and a much more broad-based across our 11,000 customer and really broad-based channel reach.

But we're very excited about the opportunity in industrial and with a bond coming in, in 2027. early this year in '26.

Our next question comes from the line of Quinn Bolton with Needham & Company.

I'll take a shot at 2026. And just looking at the segment guidance you guys talked about last quarter of 20% to 40% in common compute in the 5% to 15% in industrial and automotive looks like just based on the strength of the first quarter guide, you're probably tracking to at least the high end of those ranges, if not above wondering if that's the right way to think about growth in the respective segments? Or do you think that 1 of the segments is significantly stronger than what you were thinking 90 days ago?

Yes. Quinn, thanks for pointing out that we talked about this 90 days ago, and I think you should have picked up -- you all should have picked up during our commentary, our confidence in the basis of our outlook as strengthen considerably over the last 90 days. So I think the overall view that the Comms and Compute will be the fastest-growing of our end market sets maintains, but we're also seeing improvement across the board in our industrial end markets as well and in all geographies.

Yes, Quinn. The 1 thing I'd point out is that if you use the high end of the guidance we gave you last quarter, you still don't get to the revenue we need to get to this. So you can imagine we're on the higher end or above at the high end on both markets.

Yes, I would think you're probably at the higher end since you beat March by almost $20 million relative to consensus. Okay. and then I guess maybe Fouad, you mentioned in response to 1 of the questions that you're booking, you're pretty much all the way booked through 2026 in Comms and Compute starting to book into 2027. Can you just talk about, one, kind of lead times? Are they stretching out. Are you comfortable with lead times? And given that you're already booking out to 2027, is there any indication that you may be starting to pick up double ordering given that you're booking that far out at this point?

Yes. Quinn, as you can imagine, this is a very important question. We spent a lot of time as a management team as our channel partners, our customers across the globe, really digging into this question. And we at this point, do not believe there's double ordering. I mean, we believe -- what's giving us confidence is these orders are being scheduled throughout the year. So it's not like these are bunching up in Q1. This is scheduled throughout '26 and into '27. So you could see 5, 6 quarters out as far as orders.

Obviously, the lead times have increased for us and everybody else in the industry. We have been very forthcoming and proactive with our customers and pointing out these increased lead times and I think we're doing a service to our customer by pointing this out. We're seeing competitors that may not be as force coming. And we're being very force coming and making sure that we work hand-in-hand with our suppliers and our customers to deal with the situation. And we've been ordering substrate and putting incremental orders in place since the summer. So we've been like for the past 6 months now, increasing our orders to our suppliers, and it is putting us in a good shape to weather this next few months.

Our next question comes from the line of Christopher Rolland with Susquehanna.

Thank you for the question. Maybe just piggybacking Quinns around inventory as well. I think INA was perhaps a little bit disappointing last quarter. Do you think there was some inventory burn there -- and then looking forward, like is this a trough growth rate and we build from here potentially even with some inventory build on top of that? Like what do you guys view as kind of the normalized no inventory burn quarterly rate for INA that we can build off of.

Yes. So I'll give you the construct that we look at it. One is the inventory -- overall inventory number that we talked about was Obviously, our aggregated number at the end of the year across all of our end markets our distributors. And if you go back to where we started, it worked down first in Comms and Compute and then as you pointed out, in industrial and auto. And what we have seen now is drive, in some cases, across all end markets from our channel partners to build inventory. So we're going to continue to manage it down with the end objective not really being channel inventory, but getting to the right levels of customer support and responsiveness, so there will be times when it might have to go up to support demand that's expected and there will be times where we're going to continue to manage it down tactically.

So I don't think it would be actually that useful to give you another number to baseline off of.

Okay. And then, Fouad, maybe a 2-parter for you. I think you gave some data that was server CPU, but I was wondering if you could maybe talk about FPGA to XPU or AI processor attach rates, like where they've been and where you think they're going.

And lastly, you've talked about Lattice being a larger company, perhaps via acquiring new IP and capabilities I was wondering if maybe you could give us an update there, particularly in light of TI buying Silicon Labs.

Yes. Chris, good question. I talked about the number of server units, not the number of server CPUs. So the 15.3 million units in 2025 going 16.5% is our estimate based on market research of the number of server units TAM. Out of that total TAM, we have said in the past that our attach rate in 2025 was about 12% AI server is what we've seen. And it seems to be about correct. The 12% AI servers corresponded to a roughly high teens, call it, 20% of revenue. So that gives you a feel that the AI server as percent of revenue is higher than the traditional servers.

Recently, we've seen a nice pickup in traditional servers. So traditional servers even are growing faster than expected. So across the board, we're seeing some nice growth. And on the organic side, we have actually completed 4 little tuck-ins that we have not publicly discussed. They're all small IP and software type of tuck-ins. And we're still working on a few more of these tuck-ins, software tools, IP solutions enablement is what we're investing right now. And we'll continue to do the small tuck-ins.

The larger acquisition is something we want to do, but we want to take our time to do the right one. All the stars have to align. It has to make sense strategically accretion. And we need to want to all work together. So there's a lot of different factors, but we're not in a hurry. As you can see, the business is very strong. So we'll do it at the right time.

[Operator Instructions] Our next question comes from the line of Kevin Garrigan with Jefferies. Please proceed.

Let me echo my congrats on the strong results. As you continue to gain share in companion chips in the data center, are you seeing any increased competitive pressure or pricing aggression from some of your competitors trying to reclaim these sockets?

Kevin, we based our strategy on the customer. So we always go customer first. We make sure that we several our customers understand how we're going to provide a solution that will provide the maximum benefit enable them to grow their revenue, increase their success in the marketplace. And we are very, very focused on providing them low power, small size, low cost, fast boot time, longevity a whole bunch of attributes that we believe differentiate us in the small and midrange FPGA compared to competitors. So we have seen our market share grow, and we expect it to continue to grow. And so in that small to mid-range of PGA, we feel very strong in our market share and convicted about where we are.

Got it. Okay. Yes, that makes a ton of sense. Then Fouad, as you look out to the rest of 2026, I mean you guys are firing on all cylinders. What keeps you up at night? What could potentially kind of derail the Lattice story?

I think at this point, what keeps us up at night is our own self. So we just need with some very ambitious plans on what we're investing in and what we need to execute and deliver on and extremely focused on delivering to those plans. We want to make sure that we have the supply for our customer. We want to make sure we continue to be a good partner in what we call the 3 Ss, the solutions support and supply. And so these 3 Ss are very important to us, especially in an environment like today where the supply is short. So we're spending a lot of time ensuring that we provide the correct supply and we want to continue to be good partners. So I think the biggest thing to worry about is on self executing to our plans that are pretty ambitious.

Our next question comes from the line of David Williams with Benchmark Company. Apologies. Our next question comes from the line of Gary Mobley with Loop Capital.

Last quarter, I think you mentioned a record level of design wins measured in lifetime value. I'm just curious how those tiles in the fourth quarter. And then as you sort of look back on fiscal year '25 and tally the lifetime value of all those design wins? What kind of revenue growth rate, excuse me, is that support above?

Yes. Thank you, Gary. We have continued to get to record design wins. So every quarter, we continue to get the record design win. We had another record design win quarter in Q4. And so we continue to tally a larger number of design wins that are going to production this year and next year in 28. So that sets us up for a multiyear growth supporting at least a 25% growth. We set that forward.

I don't know how important this is or not, but I did notice that the revenue bracket in the guide this quarter is $14 million, normally $10 million. Is there anything to call out there? Is there sort of a -- maybe if you can just sort of look at the puts and takes or get the puts and takes for the lower end versus the high end of the guidance range?

It really -- there's a couple of aspects of it. One is just the quantitative view as the revenue grows higher, you're looking at a similar band around the center point and you get to a higher number. The other is there are -- as we look at what the revenue potential is when we were preparing it, you had different potential scenarios into what we would be delivering to customers and what timing it would go through the quarter. So I would say we're really, really comfortable with the midpoint. And so I think the issue is how the linearity looks during the quarter, but we're -- we've just looked at that, and we're very comfortable with that as well.

Our next question comes from the line of Joshua Buchalter with TD Cowen.

Congrats on the results and guidance. I think my -- a lot of my peers have tried to get some more color forward-looking in sales. So maybe I'll try to go backwards. If we look back at 2025 in the comms and computing segment, can you maybe walk through some of the growth drivers that took you from a sub-60 growth rate in 1Q to above 90% in 4Q. Can you give us maybe rank ordering some key sockets or whether we could think of GPU versus CPU versus ASIC and -- or order of magnitude of how much the PC headwind was in the beginning of the year? Just does any help as we think about how you guys are aligned moving forward I think would be very helpful.

Yes, for sure. So at a high level, if you look at how we help to try and frame the growth and he comes and compute, we did break up the growth in server from the growth in wired comms, right? So the growth in server year-on-year was 85%. And so you could see this is our fastest-growing segment was the server segment in the Comms & Compute. The second one, the overall comp growth was 61%. But within that, this includes wireless. So the wired comms is ahead of that. So you could see the #2 growth vector is the wired comms.

And then within those 2, so if server is #1 and the wired comms is number two, within the server, the fastest-growing piece is the AI server piece of that business. And on the AI server piece of that business, we discussed the various factors. So we discussed number one, the CapEx. So CapEx going into the year and CapEx at the end of the year was significantly higher than entering the year. And we've seen the same phenomenon happening in the past 3 months, it seems like, Josh, this is going even faster. So we got into December, thinking that 2026 CapEx for the top 5 hyperscaler would be $500 billion in January, we thought it would be closer to $580 million and post earnings now, it's at $740 million. So just within a space of December, January and early February, we've seen now the number increased significantly.

We had the same phenomenon entering 25 and ending 25%. So that CapEx number kept growing through the year. Number two, we have design wins with all of the GPUs as well as the AI accelerators. And every generation our content increases. So as you go from generation of NVIDIA to the next one as you grow a generation of TPU, generation of Meta and Amazon and Microsoft, all of our attach rates are growing from one generation to the next. So as you go through the year and the new generation get introduced, that's a #2. Number 3 is the ASP, as you go, we have new parts that are being adopted for security. So we have a bunch of new applications. We entered the year, for example, with strong application and boot power sequencing control, IO connectivity, which is the bread and butter FPGA.

As we progress through the year, we start being adopted for more at the station and more post-quantum cryptography and leakage, leakage detection is an application that we've discovered in the second half of the year and is being rolled out because we are such a good latency in being able to detect the leak and stop the leak that were being rolled out as soon as they find out because it's a very critical piece of this AI infrastructure. So these attach rates continue to grow. And as these attach all continue to grow and we go to these larger type of security type of application is driving the attach rate for larger FPGAs, which then means it's a larger ASP.

So you could see the compounding effect of all these factors together, driving to a higher and higher donor factor. And you're going to see this continue in 2027, '26 and '27, all right? So we're excited about where this takes us in the future.

Thank you for the color there. Follow-up, maybe I'm screwing up the arithmetic, but if I apply that 70% new product growth rate to, I think the mid-teens level you hit 2 I'm getting over 20% -- comfortably over 20% exposure from new products in 2025 already. I guess any help you can give us on where you expect new products to land in 2026. I think you said more than 20%. Like is -- our new products is part of this just that pre-Nexus parts are coming back now that they are corrected? Or should we still expect new products to outgrow pre-Nexus parts?

Yes. So your math is correct. You've done the right math. So good job. We did pass the -- we're in the low 20s as a total percent. So you're right, that 70% year-on-year growth puts us in the low 20s as a percent of total revenue in 2025. And we expect that to go in the mid- to high end of the 20% and the 26.

Our next question comes from the line of Srini Pajjuri with RBC Capital Markets.

From Res guys, on a solid quarter. I joined a little late, so I apologize if these questions have been asked and answered. So Ford, my question is on the supply side. I know you talked about demand being much better than expected. So just wondering if you're seeing any supply constraints on your side, or any other component shortages such as memory, et cetera, if those shortages are sort of having any impact in your view on your business?

Yes. So I'll answer the first part of the answer was already answered. I answered it again. And the second part is new, Srini. So thank you for the question. So first, on the supply side, as I said, we are committed to our customer in 3 assets, the support supply and solutions. And we've been ordering more materials since the summer or I've seen the lead time increase consistently from first half of last year to second half of last year to end of last year to now even increase very, very fast, continue to increase. We've been very forthcoming with our customers and given the really heads up on this.

And we're in relatively good shape, but there's obviously areas like substrate and assembly that they are very tight across the board. And we're working with our customers, we're our suppliers, our partners and the channel partners to make sure that we -- the right allocations and make sure that nobody is line down and -- but it's definitely a very tight situation right now. On the second part of your question, which you asked about the memory, we've asked this question many, many times to our customers, and they assure us that they've got supply agreements with their memory supplier that the forecast into Lattice are comprehending that memory supply situation. So we've obviously spent a lot of time on this to make sure that our demand was solid, and we are being assured by all the major cloud service providers, OEMs, ODMs that they've got the supply locked in with their supply agreements.

Got it. Got it. And for at a high level, strategically, you seem a bit more open to larger M&A. So just curious what the criteria would be for you because I don't know if there are a lot of FPGA companies out there that you could look at. So I'm just curious what are some of the metrics? And what sort of end markets, et cetera. And then also one of the concerns that in the industry is that we need approval, regulatory approval, particularly from China for any sort of larger deal. Do you think the environment is changing when it comes to the regulatory side .

Yes. Thank you, Srini. So I did answer the question I'll answer it again, which is we actually have done a number of small tuck-ins. We did 4 small tuck-ins on the software tools and IP that we haven't publicly announced because they're small and working on a couple more of those. So we really like the small tuck-ins because they do not require approval and a pretty strategic they're pretty accretive and are adding some very important skills in IP to our tool set and our team. So we've done them. We continue to do them and no regulatory approval needed.

For the large M&A, we've been open about the fact that we would do 1 in the future, but we're not in a hurry a whole bunch of factors to align the stars have to align, which is #1 and has strategic. Number two, it has to be accretive. Number three, the business model have to make sense. And number four, the teams have to get along and we need to have some synergies, both top line, bottom line, product road maps. So all these factors make it difficult and regulatory is one extra one. So we're going to be patient and right now, the good news is that the business is extremely strong on the organic side. So we're not in a hurry to do one.

There are no further questions at this time. I'd like to turn the floor back over to Rick Muscha for any closing remarks.

Thanks, everyone, for joining us on the call today. We'll be attending the following investor events this quarter. The Susquehanna Technology Conference in New York City on February 26 and the Morgan Stanley Technology, Media and Telecom Conference in San Francisco on March 4. This completes our call. Thank you very much for your participation, and have a great evening.

Good morning, everybody. We'll go ahead and get started. Welcome, everyone, to Needham's 28th Annual Growth Conference. My name is Quinn Bolton. I'm the semiconductor analyst at Needham. It's my pleasure to host this fireside chat with Lattice Semiconductor. The company is the leading provider of low-power programmable devices in the world that has driven innovation in the programmable market for over 40 years. Lattice's solutions span the communications, computing, industrial, automotive and consumer markets.

Joining me on stage from the company are Ford Tamer, President and CEO; and Lorenzo Flores, CFO. We also have Rick Muscha, VP of Investor Relations and David Pasquale in the audience. I'm going to hand the podium over to Ford. He's going to go through some slides, and then we'll come back for the Q&A. Thank you.

Good morning. We're just getting started in low programmable -- low-power programmable devices. The future is bright. We had a strong end of the year. We had a strong start of the new year. Those of you who are at CES, it sounds like physical AI is the new name of the show. So exciting trends that we're seeing across the board for us.

So what I'll do today is I'll give you the safe harbor that you've seen. And we'll -- quick intro on Lattice and a bit more about what we do and how we differentiate in the market. So a quick view on Lattice. We're very focused on a piece of the market, which is the small and midrange FPGA piece of the market, the most interesting, the fastest-growing piece of the market, I'll come back to why that is.

Going around from the top left, focused on communication computing being about over 50% of our revenue growing fast. The server business for us last year at the last 9 months report grew 80% year-on-year. The communication associated with that grew 63% year-on-year. Industrial and Automotive is bottomed out. The inventory in the channel is clear and go back to growth in 2026. And even consumer, we've got some interesting trends going on.

We are a high-volume supplier. So we shipped over 150 million units last year, expected to go to 180 million this year and awards recognized by awards across the industry. On the bottom right, global support to our global customers and partner and supplier and Tier 1 customer base. So what do we do -- the way I'm going to describe the business across 3 pillars: low-power programmable leader is our mission. Pillar #1, I'm going to tell you why we're the everywhere companionship. Pillar #2 is we have the best small and FPGA road map in the industry. And number three, all this should drive sustainable growth for the foreseeable future, hence, good investment.

So let's go on number one. We are -- the ASICs that we work with are the MVPs. So Michael Jordans of the world that are very powerful. These guys need a lot of OpEx to get to market, a long time to get to market. They're great, but they need a team around them. And so we are that team. We provide all the support functions around the MVP that are better done in these companion FPGAs, okay? So when you want to boot the server, we're the first thing to boots. We boot very fast. We have a very low latency. We have to boot securely. You've got to provide power sequencing, then you have to provide control. And there's a whole range of functions. When you -- these servers are being disaggregated. So you've got a board for the processor, another board for the security and I/O, another board for the switch, another board for storage, another board for power and cooling. We're on every one of those boards.

So providing a tremendous amount of function went from tens of FPGAs per rack to hundreds of FGA per rack in a data center. You look at humanoid opportunity could be very big. You've got anywhere from a few to 12 vision type of devices in a humanoid or robot. You've got anywhere from 40 to 70 motors. So again, large opportunity on a physical AI as well, providing all the support function for folks like the GPU, the NVIDIA, the AMD of the world.

But it doesn't stop there. So we obviously are on the GPU. We also are on very similar type of teams with the folks at Google and Microsoft, Amazon, Meta doing their own ASIC for AI acceleration. And so you've got the curry of the world where we -- also a different type of MVP, but very similar type of team, again, a basketball team. But we also do different types of teams. So here is Messy with the soccer team, the CPU, whether it's x86, AMD or Intel, whether it's ARM, we again provide the same support functions to those MVPs on a different type of team. How about networking? Sure, networking as well. Switch NIC, again, the Brady's of the world. MVP, we are the support team providing all the support functions.

So we do it across both data center and physical AI. So if you look first on the data center, we've talked about being companion to all of the GPU, TPU, CPU, switch and NIC on the networking, but also board management controller like the ASPEED and Nuvoton and NXP now with OpenBMC. We have recently announced some strong partnership with folks like NXP on the microprocessor. We do it across microcontrollers across these companies. And we are growing very fast in the physical AI, which will be a very strong growth vector for us in the industrial and automotive. And every sensor, you're going to find a Lattice FPGA next to them, whether it's an image sensor, whether it's a radar, whether it's a LiDAR, whether it's an infrared camera, whether it's some of these pressure and temperature sensor and industrial type of settings, we are the companionship that makes those ASICs work. So across the board. And even on the analog side, we've got strong relationship on the ADC and DAC with folks like ADI that we go together in some of these humanoid applications. So that's the companionship landscape that we focus on.

Now we do it across all these functions. We also do it across a number of markets. So we're everywhere on the functions. We're everywhere on the markets. So Communications and Computing, 53%; Industrial and Automotive, 40%. We do roll up aerospace, defense, medical under industrial automotive, but I'm highlighting them because there are some new areas of growth in those markets. We -- and then finally, consumer is also interesting. So enable innovation across all these large markets as well. So across functions, across markets, across applications, some of the world most interesting applications. Security is our bread and butter. So this is where we differentiate. We do extremely well. We're way ahead of the market on security, and this is what we were across all the data centers.

So the top 4 bubbles are data center, their rack management. The server is going from a server management to -- the new server is the rack, okay? So the rack is a new unit of compute. We actually help companies with now doing RMC, rack management controller, across the whole rack. Power and cooling is very important. I'm not sure how many of you realize how much power there is now. I'm sure you all know better than me. There is for every compute rack, there's now 1 power rack, and that's going to 2 power racks. And these racks all need control. And you don't need an expensive processor and an expensive board management controller on every one of these nodes. A small FPGA that's tailor-made for that function is perfect.

For cooling, we realized recently that we can be the fastest and lowest latency device to stop a leak, and that's spread like WiFi, this application, leak detection and shut down the water if there is a leak. FPGA is best for this. Quantum computing, we have -- we're the first introducing the post-quantum cryptography. The bad actors today are copying today to decrypt tomorrow. And what we do is we are in these devices, all the major switch OEMs, all the major security appliances are putting us in to do PQC, post-quantum cryptography, where we go in there and prevent future decryption by the quantum computer.

Moving on to physical AI. humanoids, we've done extremely well on vision. We're starting to make progress on motor. Robotaxi, we're in quite a few of those. Industrial HMI, we've had our first win using our own model in an industrial robot in Japan, very exciting. Edge AI, autonomous drone and fine AR/VR. So you could see innovation across, again, some of the world's fastest applications. And we -- most of the time are near the ASIC provide a companion function, but quite a few times when the application is small enough like industrial automation, we provide the whole primary function. So motion control, signal processing, image processing, et cetera.

Now companionship doesn't mean we're not powerful. Companionship means we're actually a very strong adder to some of these ASICs providing these functions, but we were very powerful. We do this everywhere across many functions, across multiple suppliers, which is very important because somebody says, well, I can replace you, okay, good luck trying to do this whole interop between all these different suppliers. We offload some of these very basic function into mature nodes. These ASICs typically are 2-nanometer, 3-nanometer, very expensive nodes. And we could do this in like 65-nanometer, 28-nanometer, older processes that are much more cost effective.

Accelerating the system design cycle, the system design cycle has gone from 2 years to 1.5 years, trying to go to 1 year. And so a lot of these functions don't need to be in these ASICs, get the ASIC to market and let us provide these cost-effective solution for the support function. You can program at hard -- basically program in high-level languages, but the performance we give you is hardware level performance. And what's really important, FPGA is programmable in the field. This could be programmable in the field in the future in case the application changes. A great example is this PQC, post-quantum cryptography. The algorithms are still changing. And so you're going to want to have the flexibility to be able to adapt and change in the future.

Bottom line, we feel we enable a faster pace of innovation. So that's pillar #1. We're the everywhere companionship. Pillar #2, we're very focused on the small and midrange FPGA, providing the best road map in the industry for these applications. We provide a platform. We call this our cube. So at the bottom, you could see silicon. On the vertical axis, we provide software tools, provide IP, provide solutions to make it easier for our developers to program using Lattice. And that's a gap that we still have compared to like a microcontroller or microprocessor. We're closing that gap by making it easier to use to program in our FPGA. The benefit is you get hardware-like performance versus if you are on a micro, you'll end up more like software-like performance.

We focus on 3 attributes: power, performance and size in order to win in the small and mid-range FPGA market. We've got 2 product lines: Nexus, which is our small FPGA; Avant, which is our mid-range FPGA. And this is how we differentiate. So the foundation of all FPGA is this bottom attribute, low latency. So if you're in a humanoid and it needs to stop, it could stop pretty fast. Deterministic means you provide the same answer every time. It's not easy to do in a micro or in software. Higher precision if you're trying to do screw driving or welding or bin picking in a humanoid, we're very precise. Parallel processing. So this is, by definition, we're a parallel processor engine, providing you higher performance for some of these edge AI applications. Connectivity is very key. So we provide all kinds of connectivity in a humanoid, for example, we provide EtherCAT and PROFINET that allow you to synchronize all these motors together. So the last thing you need is the arm moving at a different pace than the leg and not seeing what's happening in the torso and the vision. So we synchronize all this using connectivity and finally, future-proofing able to change application in the field.

Compared to other FPGA, as I said, we are very focused on the small and midrange hence, low power, small size, better value boot time. The last one is very important, longevity. So our fab partners commit to us to a 20-, 30-year lifetime. And our customers are asking us to sign on paper when we work together to these lifetimes. These applications, automotive, industrial, aerospace, defense, medical, they live in there for 30 years. The last thing they want is to be able to have a part that's not supported for that period of time. So that's very critical.

On the -- what's called SoC, our competitors integrate the processor inside the FPGA. And we believe that's not the right thing to do. At the time you integrate the processor and FPGA together, it's a marriage made in have. Everybody is in love, it looks fantastic. With time, you're going to see differences. And we have put partnership with folks like NXP, where you can marry big processor, small FPGA, big and big, medium, medium, small and big, it doesn't matter. We provide the best solution for the application at that moment in time. We work with NXP that has a vast catalog of parts that can provide providing a vast support on the software side where it allows you to call and see PyTorch, TensorFlow. And then the other functions are done in the FPGA where we provide some of the hardware type of functions in our FPGA with connectivity, et cetera. So -- this is how we compete, and we're very differentiated and very uniquely positioned in this market.

So everywhere companionship, best small to mid-range road map. All of this should drive growth. So we've been expanding the portfolio. We've been investing big. As you could see, '24, we exited with 91 devices. I've been in the company 1.25 years. We brought in a ton of stuff for execution. I can guarantee execution now is better than it used to be 1.25 years ago. We've released 54 new devices in 2025, 60% more that should translate into revenue. We're doing the same in '26, improving not just the silicon, but also the software, the IP, the solutions that go along with it.

We innovate across all these markets. So on the solution side, we're providing solutions to our customer and physical AI. On the left, you can see vision and motion control for humanoid robotics, industrial robots, robotaxis and anywhere that's vision. Vision is really our landmark, and we're expanding into motion control is the way to think about our physical AI type of portfolio.

On the Comms and Compute data center, security has been our very strong differentiation. We started with root of trust. We now went to PQC, this post-quantum cryptography. On the data center and communication, continue to innovate with new type of interfaces like this LTPI is the latest protocol and multiprotocol across all these applications.

And then we do this for the top customers in the world. We are in the top tier. We have 11,000 customers worldwide, about 200 is where we get most of the revenue from. But we do both a direct sale to these top 200, but we also do a channel sale to the 11,000. And all this should allow us to grow very fast. We've been growing the design wins. We've been growing our share. We're leading in the small and mid-range. And we do believe that we'll deliver faster than industry revenue growth for the foreseeable future, and we're committed to deliver EPS growing faster than revenue.

So with that, thank you very much, and we'll take questions.

Great. Thank you for this great set of slides. Thank you for walking through all the drivers of the business. I guess the first question I want to start with is one I've gotten from investors over the past year. And you look at the growth drivers, the applications you're targeting. As you step back, what do you think is a reasonable longer-term growth rate for the company? And how would that split between just TAM growth and then idiosyncratic factors, share gains, the launch of the Avant family that are specific to Lattice?

That's a great question. So this year, consensus is about 21%. We're very comfortable with that. We'll give you an update in a couple of weeks. The long-term growth of 15% to 20% should be very doable for us. And if we should grow hopefully towards the higher end of that range at -- right now, we're 69%, 70% gross margin, very strong EBITDA and free cash flow, very diversified business across all the various applications I've shown you. So that should be a very nice setup for the next few years.

We do believe in 2026 is going to be a year of the data center. We finished the year strong. We started the year very strong. The comms and compute is on fire, continue to see that -- those trends up and to the right. So 2026 will continue to be the year of comms and compute. It will be the year of the small FPGA for us. Nexus is our family that we said will inflect into '26. We're seeing exactly that, which is going to inflect in '26.

On the Industrial and Automotive, we'll see that come back to growth in '26, but really inflect in '27. So '27 then will be the year of growth in Industrial and Automotive, followed by Avant, which is our midrange FPGA business. So a nice setup here is '26 and '27 both have very strong growth vector in place with design wins going to production, products already existing. We really have to mess it up to not deliver this. So we're in very good shape for '26, '27, working on '28. And we do believe we'd like to get to that at least 15%, 20% year-on-year at high margin across the board.

Yes. And I'm sorry, I'll add. As you know well, Quinn, the -- and a lot of our investors know, this year, we've had a big headwind from the inventory that's primarily focused on the industrial and auto part of our business. And we think that's going to be normalized in '26 and set the stage for, as Ford said, the share gain and design wins that we have in place to ramp out of '26.

So your question is like there is TAM growth. Obviously, it's a large part of what we see in the data center-oriented business, but that's compounded by the share gain, the expanding applications that we just showed you guys, ASP expansion from those applications and the progression from our pre-Nexus to Nexus and Avant product lines. And then as we align with the industrial growth rate, and compound -- add that to the data center-oriented growth rate, we should be in really good shape those. We're highly confident.

And I think what you meant to say is we normalize end of '25 is...

That's right. Did I say that? I didn't say that. That's what I meant to say.

I was going to get there just to confirm that inventory in the channel.

Yes, we are on track to what we've said before.

Okay. Perfect. I wanted to start with the data center. Obviously, that's the big growth driver here in 2026. I think you had previously said on last earnings call, forecast to grow somewhere between 20% to 40% in calendar '26. What are the big drivers there? You went through some of the application security boot control. But can you talk also about what you're seeing in terms of attach rates, FPGA per server what the trend in ASP is, maybe what some of the new applications, FPGAs are...

Let me start and then Lorenzo will add. The biggest trend that is helping us is disaggregation. So on these AI servers, AI servers right now are estimate 12% of the total number of servers. These AI servers are being very disaggregated, which is a very helpful trend for us. So going from a single board to now you've got a processor board, you've got many satellite boards, you've got a board for networking, another one for storage, another one for power and another one for cooling. So this disaggregation is helping us multiply the number of FPGAs that we have per rack. So that's trend number one.

Trend number two, we continue to find these new applications, okay? And so as I just showed you. Trend number three is the CapEx continues to grow. So we were all pleasantly surprised last year at the growth that ended up happening in CapEx for the cloud service provider. Again, expecting another 50% CapEx growth this year, which is again on track and very powerful. So that's number three. Number four, this percent of AI server continues to grow as a total -- as a total number of -- as a percent of total server. And we have more content in AI server than in a regular server.

So our ASP are continuing to grow. So we go from older devices to newer devices that have more functionality, more capabilities, more connectivity and those typically have seen an ASP increase. And so you multiply all this, the CapEx growing, percent AI server growing, applications growing, attach rates growing, ASP growing, quite a few.

Pretty good tailwind.

No, I think that's good. captured it.

One of the questions we get, and I'm sure you get as well is just if you boil it down to rack is the new unit of compute, what's your average kind of FPGA content today in a rack? How many FPGAs are in those racks? And where might that go over the next generation or 2 or, say, over the next 2, 3 years?

Right. So we went from tens of FPGAs per rack to hundreds of FPGA per rack. We obviously have more in an AI rack than a traditional rack. So -- and then we have more depending on whether it's NVIDIA or a TPU from Google or Amazon or Meta, depending on the application, depending on the configuration of how people have the networking and storage defined. Storage is a big driver of attach as well.

So depending on how people configure around their AI, around their networking, around the storage, it varies, but it's all now in the hundreds. And we went from -- we have a few dollars per ASP per part. So you can see we went from low hundreds to now high hundreds of dollars per rack.

I just CFO temper this a little bit. It's very -- it's -- we get the question a lot because people want to take the number of racks and multiply and get to a number for us. And it's just -- we would say, because of the complicated way these are deployed, as Ford described, it's actually difficult for us to say, as these things ship, it looks like x units per rack and therefore, there's much revenue per rack. So we keep looking at it. We know it's growing, but how it actually manifests itself when different companies, different OEMs ship servers, it's actually quite difficult for us to pin down. So we're a little hesitant other than the general guidance that Ford provided to frame the dollars per rack.

Yes, it's a range from the traditional server all the way to the [indiscernible] server.

How much -- how many blades are populated in the rack. We don't necessarily know that kind of stuff.

Would it be fair just sort of to generalize, obviously, that business would grow with the number of rack shipments over time. The number of GPUs or XPUs in a rack is going up. And I would think given your attach rate trends and your ASP trends, you probably should be growing at least as fast, like if the number of GPUs in a rack double, you would -- would you double with the number of GPUs?

The best way to think about it is look at last year growth, right? The first 9 months -- we haven't reported Q4. The first 9 months of the year, we reported our server business growing 80% and our comms associated with that growing 63%. So we grew at least at the same rate in those segments as the rest of the industry.

Okay. I wanted to spend a few minutes on the com side of the business. I think it's mostly wired comms is also being driven by the AI trend, but spend a minute on wired comms. And then are you seeing anything interesting sort of happening on wireless infrastructure? Could that be an opportunity over the next year or 2?

Yes, absolutely. So on the wired com, there has been a lot of changes as -- what's nice now is an explosion, and you're going to have Credo and Astera Lab after us, right? So they're going to even talk to you about all the scale up and scale out and all of these. So there is a proliferation of scale out and scale up and all kind of new architecture being proposed, which is fantastic because the more diverse this is, the more you need somebody companionship across all this to make sense of all the disparity. So the variety of different comms, architecture are quite interesting.

I was pleased to see end of last year, growth being -- having this investment from NVIDIA I was born and early investor. And they also have a totally different way to approach how they hook up all these LPUs together. So you're going to see even more of this variety of how stuff gets hooked up, and that's good for us. On the wireless side, we definitely see new ASICs come to market from the major wireless providers, and that's going to be another attach rate that will help our wireless business for sure.

Great. The last question I had is if you look at the growth rate of the overall comms compute bucket in 2025, it looks like it will be mid- to upper 20s, where you talked about the server business growing 80% for the 9 months and comms up, I think, 63%. So there were some legacy business that I think you were fighting that headwind as it sort of rolled off. Is that legacy PC client business now effectively 0? Is comms and compute effectively all server?

Yes, all the growth will be really coming out of the server-oriented business. That headwind is gone for the future. Now we'd like to pick some of it back up, and we're still working on that. But right now, it's going to be the server-dominated business.

Got it. Okay. Just moving to the Industrial and Automotive segment. You confirmed you're sort of on track for the inventory normalization by the end of 2025. That business has been pretty slow for 15 months. I think you've been undershipping the channel. What have you seen in terms of POS trends or consumption that give you comfort of at least, I think, your 5% to 15% growth rate in '26 and probably better in '27?

Right. Obviously, as we came out of the peak of the industrial business that we had a couple of years ago and slid down to the baseline level where we are, we were taking inventory out of the channel, so POS was down in the second half of 2025, we'll get that year right, 2025. What we started to see is a significant uptick in the POS, which is consistent with what we expected in terms of depleting the channel inventory and setting the table for 2026, right? So it's been very good performance out of the industrial-oriented part of the channel.

Excellent. And I was going to ask any comments on what you're seeing in sort of the robotics side, factory automation? Are you starting to see demand coming back? It sounds like POS is picking up. Is it being led by those applications? Or is there other parts of the business...

I think it's the broader industrial. We're still very early on in the robotics side of the business, but it's a broader industrial play right now.

I mean the positive for me, when we look across geographies, all geographies are contributing. There were some geographies like Japan that were laggards in 2025, and even that has now come back to growth. So that's very positive.

In your Slide 4, you mentioned robots or humanoid robots. There are a couple of systems that are Lattice can play, the vision systems, the motor control. Can you sort of elaborate on how you're feeling about your design win position on vision and what the opportunity on motor control may be going forward?

On vision, we've done very well. Vision has been -- one of our strengths has been something we've done for a long time. So vision, we've really named Vision. Motion control is -- we're making progress. We've got a few headwinds -- beachheads as well, not headwinds, tailwinds and beachheads. So we are going to take those and try to see how much can we expand it on the motion control. So we're -- I think that's the summary.

Okay. And when would you expect those designs to start to contribute to revenue? Is that sort of a 2027, '28 time frame, just knowing that the design cycles are probably a couple.

Most of these human applications are going to market in '27. So across all geographies, U.S. and China.

Okay. I know investors track your new products, Nexus and Avant as a percent of sales. How are you tracking to that relative to your prior expectations of high teens in '25 and mid-20s in 2026? Do you feel comfortable with that progression?

Yes, I think that's pretty much where we are. And the products that we're bringing out, as Ford showed on this slide, we're expanding the product line significantly. Those do take a little bit of time to ramp in. But given the baseline we have, we're very comfortable with that new product ramp.

Just to give you a feel that we -- our new product growth in 2025 was about 65%. So we grew very nicely from '24 to '25. The new products grew very nicely in '25. We expect that to accelerate in '26.

Okay. And again, 2026 is really the year of Nexus. 2027, the year of Avant, Nexus more data center, Avant more industrial...

That's right.

Okay. Sort of a similar question, investors also track your AI revenue, which splits across both of your major end markets. What are you seeing on sort of the trends for the AI revenue in the business?

Our AI revenue has been 60% data center, 40% Industrial and Automotive. We're still tracking to that. We said we'll get to the high teens end of '25. We're on track to this. We said we'll get to the mid-20s in 2026. We're on track for that. So AI as a percent of revenue continues to increase. And this has been more like a companionship story. We -- as I said, the first -- for the first time now have a primary chip win that we've won in HMI. And we're excited about this because in the small type of models, they call tiny models that go into these robotics and industrial, we've got a solution that's differentiated as a primary function.

Excellent. Shifting to the competitive landscape. Intel has now respun out Altera under new management. What are you seeing on the competitive front from Altera? Do you think they get more aggressive? Do they try to get more active in the embedded markets and small mid FPGAs?

Yes. Look, I mean, competition is going to be around. I mean, it's good, there is competition. It means it's a good market. If there is no competition, there's not a good market. So yes, we do believe Altera would get more aggressive, and we do believe they'll come after us in the road map. But really, the major competitor we're worried about today is Xilinx. Xilinx is the one that has an UltraScale++ in our space. And Xilinx is a much bigger company in FPGA, and we've done very well against Xilinx. So we have fundamentally a very sustainable architectural differentiation in our -- in how we approach this low-power small FPGA architecture. And that we believe is sustainable for many years to come. And so recently, we've actually started making headway -- making progress in China against some of the China local FPGA. And these guys are very tough from a pricing and competition, and we're winning there.

So we -- our view is, look, focus on the customer, don't focus on the competitor. The competitor is going to be there. We're going to stay paranoid. But fundamentally, what is it that we need to do to leapfrog and make sure we provide benefits to our customers. I've always been customer-centric. I've always been customer-driven, and that's how we do the road map and the strategy. We don't -- we lead, we don't follow. So the last thing I'm going to do is worry about Altera and follow them. They're actually trying to follow us right now. We know the road map, and they're 24 months behind us. The guy that we worry most about, as I said, is Xilinx.

Perfect. Last question for me. Just sort of can you describe sort of your key priorities for capital allocation going forward and sort of thoughts on M&A. I know at Inphi and Broadcom, you were pretty active on the M&A front. Is that a big part of the strategy moving forward?

Let me -- I'll give you the overview and then Ford will elaborate on the M&A viewpoint. So look, we have been aligned from before the beginning of my time there that this is a business that is organically investable. We have really good opportunities in front of us, and that's where we focus primarily. And you saw that in our expansion of our product line. You'll see that again this year as we continue to drive our business into the expanding opportunity that we have. So that's the first thing we do. We will look inorganically at things to accelerate the business. And look across the board.

The cube that Ford showed gives you somewhat of an insight into the vectors that we look along, but we're definitely looking at that. And then we always have had a view on prudent return of capital to shareholders when we think we can keep the balance sheet flexibility and have generated cash. We generate a lot of cash for our size, right? So we are fortunate in that we have the investment opportunities in front of us.

Ford, do you want to elaborate on the M&A?

Yes. M&A, when I joined Lattice, the reason I joined is I think Lattice is a great foundation to build upon. So we're trying to build a big house. And the key is to have a foundation that's very solid. And so when I looked at the business, it's a very diversified business across many segments, across many applications, Tier 1 customers, a great team, a very strong strategy and high gross margin, high EBITDA and free cash flow, sustainable, high revenue growth for the foreseeable future. So that's what I was looking for, and that's what we have.

And so you can invest in this, and then we're going to build the house on top. But the key was to make sure that the foundation was very solid and had the growth vector and the margin profile that will allow us to actually add interesting stuff on top. And the goal here on M&A is to become more strategic to our customers, and we're looking at a few things that would help us do that.

We've done in line with investing in the business, we've done some small tuck-ins already and acquisition of IP and things like that. But that's just part of investing in this core business.

In the core business. Excellent. All right. Well, we are at the end of the session. So Ford and Lorenzo, thank you very much for coming at the Needham Conference. We really appreciate it.

Thank you.

Ladies and gentlemen, greetings, and welcome to the Lattice Semiconductor Third Quarter 2025 Earnings Conference Call.

[Operator Instructions]

As a reminder, this conference is being recorded. It is now my pleasure to introduce your host, Lattice Semiconductor's Vice President of Investor Relations, Rick Muscha. Please go ahead.

Thank you, operator, and good afternoon, everyone. With me today are Ford Tamer, Lattice's CEO; and Lorenzo Flores, Lattice's CFO. We will provide a financial and business review of the third quarter of 2025 and the business outlook for the fourth quarter of 2025.

If you have not obtained a copy of our earnings press release, it can be found at our company website in the Investor Relations section at latticesemi.com.

I would like to remind everyone that during our conference call today, we may make projections or other forward-looking statements regarding future events or the future financial performance of the company.

We wish to caution you that such statements are predictions based on information that is currently available and that actual results may differ materially.

We refer you to documents that the company files with the SEC, including our 10-Ks, 10-Qs and 8-Ks. These documents contain and identify important risk factors that could cause the actual results to differ materially from those contained in our projections or forward-looking statements.

This call includes and constitutes the company's official guidance for the fourth quarter of 2025. If at any time after this call, we communicate any material changes to this guidance, we intend that such updates will be done using a public forum such as a press release or publicly announced conference call. We will refer primarily to non-GAAP financial measures during this call. By disclosing certain non-GAAP information, management intends to provide investors with additional information to permit further analysis of the company's performance and underlying trends.

For historical periods, we provided reconciliations of these non-GAAP financial measures to GAAP financial measures that can be found on the Investor Relations section of our website at latticesemi.com.

Let me now turn the call over to our CEO, Ford Tamer.

Thank you, Rick, and welcome, everyone, to our third quarter earnings call.

As I marked my first year as CEO, I'm more confident than ever in Lattice's upward trajectory. Our strong Q3 performance and forward-looking guidance reflect the strength of our strategy and execution. With a world-class team and a robust innovation pipeline, we are well positioned to capitalize on the ever-expanding investments in AI and data center infrastructure.

At the Open Compute Summit last month, we saw increasing interest in Lattice's low-power data center offerings. We also witnessed accelerated momentum for Lattice's security and board management solutions that we provide to hyperscalers, neo-cloud and server and communications OEMs and ODMs.

And at the same event, we demonstrated our leadership position in post-quantum cryptography or PQC for short.

I'm pleased to report that the adoption of Lattice's PQC technology is also accelerating due to the NIST requirement that systems be CNSA compliant.

We exited Q3 with even higher confidence demonstrated by new use cases for Lattice products across our core markets as well as the acceleration of design wins, which were on pace for a record year in 2025.

This momentum highlights our differentiated value proposition, low power, small size, fast boot time and high reliability and sets the foundation for rapid growth into 2026. In general, we continue to see data center investments expand across all payers and applications in the AI infrastructure tsunami.

Lattice is benefiting from a corresponding revenue growth evidenced by increasing bookings now and into 2026. This is giving us increased confidence to invest further for growth in 2027 and beyond.

We firmly believe that delivering above-average revenue growth is consistent with why our shareholders invest in Lattice.

For Q3, we delivered revenue of $133.3 million, up 7.6% over Q2. This represents the highest sequential growth in more than 4 years. Our Q4 revenue guidance of $143 million at the midpoint equates to 22% year-on-year growth.

This estimate is the largest increase in nearly 2 years and shows our belief in a strong recovery and upward momentum. With respect to end markets in the third quarter of 2025, Communications and computing grew 8% sequentially and 21% on a year-over-year basis to a record level.

The computing subsegment growth is being driven by our expanding footprint and increased use cases in both general purpose and AI-optimized servers.

And the communications subsegment growth continues to be driven by wired data center infrastructure, including network interface cards, switches, routers and security appliances.

As expected, the Industrial and Automotive segment increased 6% sequentially. The growth rate is tempered as we continued to strategically shift under true demand to normalize channel inventory.

As we told you on prior calls, we are on track to normalize channel inventory by year-end, positioning us for renewed growth into 2026. We are confident that we're gaining share across smart factory, robotics, medical and aerospace and defense applications based on customer feedback and design win activity.

While we remain subject to macroeconomic and industry conditions, there are several factors fueling confidence in our prospects moving forward.

Lattice's addressable market is growing due to the size of the infrastructure capital expenditures growing fast, our diversified position in the largest and fastest-growing applications, growing attach rates, increasing average selling prices from our new products, broadening the application footprint of our small and mid-range FPGA portfolio; and finally, increasing AI usage.

Taken together, we believe these dynamic factors are expanding dollar content for Lattice per customer system. We also continue to win with pre-Nexus, Nexus and Avant products.

Customers are consistently choosing Lattice over our competition as evidenced by the growth in our design wins, which are on pace for a record in 2025.

Revenue from our new products continues to grow at a strong rate, and we are on track to exceed our 2025 goal that we projected in prior earnings calls.

Lastly, we estimate the percentage of AI usage across our products will be in the high teens in 2025 and mid-20% range in 2026.

In summary, Q3 was a strong quarter marked by consistent execution and strategic progress. We remain focused on delivering differentiated innovation, deepening customer relationships and driving long-term shareholder value.

We have increased confidence in our outlook, led by our leadership position and ability to capitalize on the compelling opportunities in front of us to drive accelerating growth in 2026 and beyond.

Let me now turn the call over to Lorenzo for a detailed review of our Q3 results and Q4 guidance. Lorenzo?

Thank you, Ford, and good afternoon, everyone. We will begin with a brief overview of our third quarter 2025 financial performance, followed by our fourth quarter outlook. We are pleased to report that Lattice again delivered on expectations with revenue, gross margin and operating profit all in line with our outlook for the quarter.

Revenue increased 7.6% quarter-on-quarter and 4.9% on a year-over-year basis to $133.3 million.

Overall, this was the highest revenue we have obtained in 5 quarters, and we are expecting continued growth in Q4 and in 2026.

We set a new record for communications and computing revenue, which grew 21% year-over-year and 8% sequentially. We expect strong growth in this end market in Q4 and in 2026.

Our gross margin expanded by 20 basis points quarter-over-quarter and 50 basis points year-over-year, 69.5% on a non-GAAP basis.

This performance continues to reflect the durability of our business model and the value and differentiation our products provide for our customers.

Non-GAAP operating expense was $53.9 million, in line with our guidance. OpEx was flat on a year-over-year basis. Roughly 4% sequential increase in operating expense reflects our strategy to invest in the products, infrastructure and talent that will further strengthen our leadership position and enable us to drive accelerating growth.

We have built confidence in our 2026 revenue expectations. Aligned with those expectations, this quarter, we accrued stock-based compensation expense for PRSUs. This accrual was the primary driver of the increase in our GAAP operating expenses.

Our non-GAAP operating margin expanded 150 basis points to 29%, and our EBITDA margin also increased 150 basis points to 35.6%. We delivered non-GAAP EPS of $0.28, which was at the midpoint of our guidance and represented 17% growth on both a year-over-year and quarter-over-quarter basis.

GAAP net cash flow from operating activities for the third quarter of 2025, increased to $47.1 million, up from $38.5 million in Q2, with a GAAP operating cash flow margin of 35.4%, up from 31.1% in Q2.

Free cash flow in Q3 was $34 million with a 25.5% free cash flow margin, up from $31.3 million and 25.2% in Q2. This remains a focus area for Lattice, and we expect to continue to generate strong free cash flow.

We are achieving these levels of free cash flow while strategically investing in CapEx in support of our product road map and operational improvement projects.

We are pleased with the continued improvement in industrial and auto channel inventory. We continue to track to plan and expect inventory normalization by the end of 2025. As we noted last quarter, comps and compute channel inventory has already normalized.

Now let me turn to capital allocation. Our balance sheet remains strong. We remain debt-free and have ready access to capital to support both organic and inorganic growth opportunities, and we remain well positioned to navigate macro uncertainty.

Given our balance sheet strength and our business model, returning capital to shareholders remains a key component of our capital allocation strategy. During the quarter, we repurchased approximately $15 million of common stock under our existing buyback program.

Through the first 9 months of 2025, we've repurchased approximately $86 million of common stock. We have $14 million left on our current authorization, and we will be reviewing our next authorization with our Board of Directors in December.

Now let me turn to our Q4 guidance. This guidance reflects the recovery of our business and sets the stage for continued growth into 2026. In Q4, we expect revenue to grow and be in the range of $138 million to $148. At the midpoint, this represents revenue growth of 22% over Q4 of last year. This would be the highest year-on-year growth in nearly 2 years and is supported by the strongest booking patterns we have seen in at least 6 quarters.

We expect gross margin to be 69.5%, plus or minus 1% on a non-GAAP basis. Non-GAAP operating expenses are expected to be between $54.5 million and $56.5 million. The income tax rate for Q4 is expected to be between 3% and 5% on a non-GAAP basis.

Non-GAAP EPS is expected to grow to be between $0.30 per share and $0.34 per share. In closing, Q3 was another strong quarter for Lattice. We delivered results in line with our guidance, and we expect further acceleration given our position in high-growth markets.

We are driving near-term operational improvements, investing to strengthen our leadership in small and midrange FPGAs and positioning the company for an even stronger 2026.

Operator, that concludes our formal remarks. We can now open the call for questions.

[Operator Instructions]

Operator, this is Lorenzo Flores at Lattice. I'd like to tell all our audience that we found a typo in our press release that relates to our operating expense guidance. The guidance that I provided in my prepared remarks of between $54.5 million and $56.5 million is correct, and we will correct the press release, which says between $54 million and $55 million immediately after this call. So with that, we can start our questions. Thank you.

We take the first question from the line of Kevin Garrigan from Jefferies.

Congrats on the results. So first question, you said you had increased confidence in your outlook for 2026. Is that confidence contingent on the expected normalization in industrial and auto channel inventory in Q4? Or can the strong comms and compute growth you guys are seeing more than offset any potential macro industrial and automotive softness?

Thank you, Kevin. This is Ford. As you've seen from our Q3 results and Q4 guidance, our comms and compute business revenue growth continues to accelerate, and it will accelerate further into 2026.

Our comms and compute business as a percent of total revenue went from 35% of total revenue in 2023 to 45% of total revenue in 2024 to an expected over 55% of total revenue in 2025, and we expect that to grow to about 60% of revenue into 2026.

With that growth, that segment becomes very significant to our acceleration of revenue growth. And in 2025, we have accelerated server and communications faster than the underlying CapEx growth.

So our server business has been up 85% year-to-date compared to 2024, and our communications business has been up 63% year-to-date compared to 2024. So you could see both server and comms are growing faster than the underlying CapEx, and that should continue into '26.

So you would expect that comms and compute business to drive accelerated growth in '26. Furthermore, as you indicated, we are on track to get industrial automotive inventory normalized by end of the year.

Our comms and compute inventory is already normalized. And our overall channel inventory is expected to get in the 3.x by end of the year as we had previously told you on prior calls.

And with that, we should also see the industrial and automotive shift to natural demand as opposed to chip to under natural demand like we're doing through 2025. So with both comms and compute accelerating and industrial and automotive recovering and shipping to natural demand, we expect 2026 revenue growth to be very strong.

Perfect. And just as a follow-up on the Industrial and Automotive segment, can you just talk about what you're seeing on a regional basis? Any geographies that are stronger than you had expected?

Yes. Thank you, Kevin. So we are seeing China automotive to continue to be strong. Automotive for us overall is less than 5% of our overall revenue, but China is probably the region where automotive is strong.

We are seeing the aerospace and defense to be strong worldwide, and we are winning increasing share of designs in that segment. We are also seeing this physical AI, a lot of design win activities, wins are companion chips for physical AI across all the various segments from industrial robotics to humanoids to automotive, to medical, aerospace, defense, test and measurement and all the various segments in industrial and automotive.

So we're pretty positive on where that segment is going to go into '26 and inflect even further into '27.

We take the next question from the line of David Williams from The Benchmark Company.

Congrats on the real shift in your tone. And I guess that's my first question is just around your confidence. Clearly, you outlined all the nice drivers there. But I guess what do you think has changed over the last 90 days that maybe has changed your confidence?

It feels like that inventory is clearing up. And so you kind of talked about that last quarter. But just wondering if there's any demand issues or demand things that have changed driving that greater confidence.

Yes. No, great question. David, thank you. We had a very successful Open Compute Summit. It was a show attended by most of the hyperscalers, server and OEMs and ODMs as well as the related wireline communication equipment OEMs and ODMs.

And at that show, attendees were noticing that somebody like flipped the switch. And all of a sudden, the past 90 days, we've seen definitely an increase in activity and spend.

As you've seen from the most recent hyperscaler earnings, the forecast for next year has been up, and we've seen that. We also are starting to work closer with some of the neo-cloud and enterprise vendors, and we're seeing them being more aggressive in wanting to increase their AI CapEx. One of the hyperscalers gave us a quite a big increase forecast for '26, '27 and even into 2028.

And we're seeing our adoption as a companion chip for some of these AI in both the data center and the physical AI accelerate. So for example, on the cloud data center, we're seeing our adoption as companion chip for CPU across x86 and ARM, AI accelerators from NVIDIA, AMD, Intel, hyperscalers, various accelerator ASICs across networking, across Broadcom, Mirvan, Mellanox and Cisco, security, board management, rack management, cooling, power management as a companion chip, the small and midrange FPGA from Lattice are doing quite well, and we're being adopted at an accelerated pace.

And we're seeing the same as a companion chip on physical AI around sensors for like camera, LiDARs, radars as well as various industrial and automotive applications.

I think -- and really finally, I'll repeat that what I said in my commentary, we've got the strongest book-to-bill that we've had in 1.5 years or so, and it's booked into the first half of next year.

So all the sentiment and enthusiasm that Ford has just described and the reasons for that are actually showing up in our orders. So we see not just the spirit, but the actual business coming.

Great. A lot of great color there. I certainly appreciate that. And maybe just from the Avant platform, you talked about seeing growth across all. But just kind of curious if we can get an update on the Avant platform, how the design wins are trending there and maybe what the expectations are for growth into 2026.

Yes, both Nexus and Avant are doing quite well as well as some of the pre-Nexus platforms as well. What we said in the past is that we believe that 2026 will be the year of Nexus and '27 will be the year of Avant.

We see 2026 as being the year of the data center with '27 being a big recovery in industrial and automotive. Nexus seems to be more related to the data center as Avant is more related to some of these midrange applications in industrial and automotive. And so that's how we're seeing the revenue stagger over the next couple of years.

David, does that answer all the questions you have?

Yes.

We take the next question from the line of Tristan Gerra from Robert W. Baird.

Thanks for the color about AI-related demand. And as we see an acceleration in AI demand, I guess we shouldn't read much into a little bit of a sequential slowdown in Q3 for communication and computing because it looks like that line was up 20% sequentially in Q2, up 8% in Q3.

So first, should we assume a re-acceleration of the growth in communication computing in Q4? And is the strength in Q2 perhaps linked to the fact that this is the first quarter where inventory levels normalized?

And as such, you kind of caught up with end demand and now you're back in line shipping with real end demand. Is that how we should read into this?

Yes. Let me handle parts of this, Tristan. I think the guidance we've given for Q4 actually shows significant sequential growth in comms and compute. I think there is some of the non-server business in our comms and compute business that impact the measured growth rates.

But I want to refer you back to earlier comments that we've had previous quarters and again this quarter about the real driver of the comms and strength is our server demand.

And that's over 80% growth year-over-year. I think that's a very strong indicator of what's going to drive that.

Now that may not stay exactly that high, but the point is it's growing faster than hyperscaler CapEx by a significant amount, and it is showing the expanding footprint in terms of the design wins we have, the number of chips in design, ASSP design, and we expect that to continue to grow, as David's question pointed out, as we go from generation to generation from pre-Nexus to Nexus to Avant.

So I think all those things point to a strong acceleration. So I don't think I would get too hung up on the couple of decimal points on a quarter-to-quarter change.

I think if you think of the overall trajectory through time, you see it's quite strong.

That's very useful. And then when should we look at gross margin picking up from current levels? And what will be the catalyst? I mean it sounds like with the growth that you're seeing in computing that there should be gross margin expansion next year.

And I understand you're not guiding, but is it fair to say that we should see gross margin expansion next year? Or are there any offsetting factors? And also wanted to understand the commentary about '26 being the year of Nexus.

Historically, and back in the days, as you know, high-end FPGA will typically see revenues peaking after about 7 years.

For Nexus, it looks like after a fairly muted ramp the first few years because you used to provide a breakdown. And earlier this year, we were kind of in the low teens as a percent of revenue for Nexus, which given the launch date of late 2019 was probably a little bit below what I would have thought at least at the time.

But now it looks like the momentum is accelerating, and I wanted to understand the dynamic of that. Why is Nexus getting after all those years, even more momentum?

Yes. So I will -- we're going to parse this question out probably amongst us here in the room. And I want to start with the gross margin question, which is actually one of my favorite questions because it gives me the opportunity to reflect back on the level we are actually at 69.5%, which is pretty good, and it takes a lot of work to get there.

Your question is actually in terms of the impact of comms and compute on the overall gross margin. I think we hear the other side of the question, whether industrial recovery will impact it positively.

I think you all have to take into account that there's a spectrum of profitability across our end markets and within those end markets, within particular design wins as they ramp up.

So we manage the overall portfolio to get to the margin, and it would be potentially plus or minus a few basis points going through time.

But I think we're pretty comfortable with this margin for now, anticipating several different mix scenarios. So I think we're not looking for very big improvement from one factor or another as we go through time, but pretty much staying where we are.

On the Nexus ramp, I do think the -- I'll start on this and Ford can fill in. One of the things to keep in mind is we are just now expanding the product line with Nexus, bringing out more products. The phenomenon you talked about before is more where there's a burst of products, but we are ramping new Nexus products through this year and next.

So Ford, do you want to fill in anything on the Nexus ramp?

Yes, absolutely. Tristan, good question. The Nexus products continues to be introduced. So if you think about it, we've introduced a new I/O optimized Nexus in Q2 of 2025. We've introduced 4 to 5 new SKUs of Nexus in '25.

We expect another 5 to 6 in '26. So yes, 2019 was the first product introduction, but we continue to roll new Nexus product. We've introduced a Nexus 2 that is going to, again, start to be rolled out.

So Nexus is the whole family with quite a few devices. And so when we say Nexus, Nexus really is what we call our small FPGA, and that's going to be going on for the next 10 years.

And we're seeing a tremendous bunch of new applications. So we went from booting and I/O expansion and buffer and control to now Root of Trust and to PQC and to board management and to rack management, the latest OCP, we had leak detection. So the number of applications that we are finding are immense. And you wouldn't be growing 80% -- over 80% year-on-year in server if we weren't widely adopted and continue to being adopted at even faster rate.

So we're very positive on Nexus for the next quite a few years here. And again, think of Nexus as what we're calling our small FPGA family. And we're leading there. We're going to continue to lead. We're going to continue to invest to continue to lead.

We take the next question from the line of Gary Mobley from Loop Capital.

Can you size in industrial and automotive, how much you're undershipping is embedded in your Q4 guide as you work down that inventory? And then on the other side of the coin, are you running into shortages or extended lead times for the comms and compute segment?

Yes. Thank you. So we don't break it up exactly, Gary, but we have said in the past that we're undershipping by about a couple of weeks a quarter, so call it $15 million to $20 million a quarter. And we expect that to be normalized by end of the year.

So you could see this being a headwind in '25 becoming a tailwind in '26. The lead time on the comms and compute are expanding, but we are on top of it and are very focused on making sure our customers get supply, and we have done a very good job so far.

Yes. We announced the extension of lead times a couple of months ago, actually in August and in parallel, work with both our suppliers and our customers to make sure that our deliveries were matching up.

And so we have the visibility from our order book to keep doing that. I think our customers have been satisfied with our delivery performance. And the extension of lead time is really actually helpful for us to plan the business and plan our loadings with suppliers.

And as I indicated earlier, we now have good visibility into a very strong book of business going forward for the next 3 quarters at least, including Q4.

As you think about your data center opportunity or sizing the served available market, do you use any sort of benchmark like what your dollar opportunity is per gigawatt of capital spending plans by the different hyperscalers or maybe differently, your content per rack scale solution, all-in GPU, CPU, networking? Anything there that you can share would be helpful.

Yes. No, we do have that, Gary. We have an attach rate per server, attach rate per rack, attach rate for different companion chip that we are, if we're companion chip to a CPU or a GPU or a switch or a NIC or storage or board management, these are all very different dollar per chip. So we are -- we have very detailed models.

We obviously don't share those.

What I can tell you is we're going to grow faster than CapEx. The CapEx is increasing fast, as you know. But we're growing faster because our attach rates are growing faster. The users of Lattice FPGAs and these data center applications continues to expand.

We're going to newer products like Nexus and Avant, hence, a higher average selling price for these newer products.

We have an increased AI server adoption, which again is driving higher content for Lattice. So overall, you could see we'll continue to grow faster than the underlying market.

We take the next question from the line of Christopher Rolland from Susquehanna International Group.

So Ford, I wanted to go back to your comment about data center at 60% of total for next year. So I think previously. 60%, right? Sorry, CNC, yes. So I think previously, you guys have said like 15% to 20% growth year-over-year that in I&A, inventories are normalizing by year-end.

I think you guys reconfirmed that. I would have thought that just as that normalizes for I&A, it would create a great deal of growth for next year.

So I would have assumed I&A is growing here. I mean there's 2 parts to that 60% comment that you made. CNC could be much higher, but also conversely, I&A could be much lower.

And so I'm trying to balance that like to understand like what I&A growth could be next year. You also made the comment that next year is really the year of CNC and I&A is 2027. So trying to read the tea leaves, particularly for I&A. Are you thinking it is a much more flatter year than perhaps we are?

So let me start at the beginning and at the high level, and you recall in my remarks, I said we're growing in confidence and with respect to our growth in 2026. And because of that, I had to make an accrual for some performance-based, revenue-based stock compensation. That threshold was 20% into 2026 from where we expect to end the year today at the midpoint of our guidance. So that's where we are. And Ford, do you want to add the color around the end markets? And go ahead or I can do it.

Yes. No. So at the high level, Chris, I&A is expected to go back to growth next year. We think it would be in the sort of mid-single digit to, call it, 15% range. But the comms and compute could be more in the 20% to 40% range. So that's how to think about these 2 segments.

Right. I wouldn't -- we wouldn't have the confidence in the growth if we were thinking we were highly dependent upon significant growth in industrial and auto because as you know that's more subject to the macro environment, and we are still working through the end of the year on the inventory.

So when we look out into 2026, we didn't want to drive our expectations assuming a bigger than mid-single-digit growth rate, as Ford, said in industrial and auto.

Excellent. And then maybe bigger picture, Ford, just thinking about your legacy kind of in AI, but also networking and maybe the next steps for Lattice, perhaps maybe even becoming more than just an FPGA company.

I think we've talked about putting hardened ARM cores into an FPGA before. But what about putting hardened MAC cores or an NPU into an FPGA and/or building kind of an ASIC capability to guide some of your customers to FPGA into a hardened ASIC? Or just any other thoughts on the next steps for this company and what you're exploring or thinking about without, of course, mentioning the specifics?

Yes. Look, we're very excited about where we can take Lattice into the future. On the hardened processing core, if you wish, we have adopted the strategy of being partners with microcontrollers and microprocessor companies. And we've had some very good success to date.

The most recent customer meeting we had in Asia, we had NXP Semiconductor be there with us. We worked together at OCP and the partnership with NXP is actually becoming quite strong and customers quite like the 2 of us together.

We're expanding it across others. So on the microcontroller, we're seeing interest from other players in that space to do the same. And you'll hear more partnerships in the future.

So we do believe that on the processing side, whether it's microprocessor or microcontroller, you're going to see 1 plus 1 equal 3, where we'll be providing joint solutions to customers that leverage the best out of FPGA and these processor next to us.

On where we're going to take this longer term, there's definitely quite a few discussions going on and exploration, but we're not ready to discuss at this point.

We're being asked by customers to do some specific stuff that we're investing in and that may be part of the increased investment into 2026 that Lorenzo has hinted to in his prepared remarks. So stay tuned on that. We'll have more detail as we go.

We take the next question from the line of Ruben Roy from Stifel.

Ford, I wondered if you could add a little detail to the AI usage comment. I think you previously said high teens going into the mid-20s for next year. How are you thinking about mix on that relative to compute and comms versus industrial and automotive? And maybe a little detail on how you're defining AI usage would be great.

Thank you, Ruben. We are still on track for the mid-teens in this year, 2025, going into the mid-20s by next year, 2026.

And the ratio is about 60% comms and compute versus 40% industrial and automotive.

On the comms and compute, we play a role of a companion chip where to the various ASICs and ASSP I've mentioned before.

On the industrial and auto, we can play a far edge AI near sensor intelligence, if you wish. And that's our focus, less than 1 tops, these very power-sensitive applications that are really tied to these camera, LiDAR, radar, other industrial sensors. I hope that answers the question.

Yes. That's helpful. And then just a follow-up on your comment regarding Nexus -- year of Nexus next year and then Avant '27. And you mentioned that Nexus sort of plays into data center and Avant may be a little more slated for industrial and auto.

And I'm just wondering why you wouldn't see some mid-range FPGA kind of usage in some of the, I guess, higher-end comm and compute applications that are coming up. Am I reading that wrong or listening to you wrong or anything to add on that comment?

No. I mean we would see Avant applications in the comms and compute. We'll see Nexus application in industrial and auto. So I wasn't trying to imply exclusion. I was trying to imply a majority, if you wish, of the design win will be Nexus on the comms and compute and Avant on industrial and auto.

And we, for example, had Ericsson at our developer conference, who was showing an Avant application. And so again, we'll have applications of Avant in comms and compute for sure.

We take the next question from the line of Melissa Weathers from Deutsche Bank.

Since there is no response, we'll move on to the next question, which is from the line of Joshua Buchalter from TD Cowen.

I guess I wanted to ask about some of the assumptions behind the growth rates you just talked about for 2026 or the revenue split, I should say. Any metrics you can give us on what would -- or details you can give us on what would allow you to come in at the lower end versus the high end of the 20% to 40% range, given it's such a wide range and maybe any details on how much contribution you're expecting from general purpose versus AI servers?

And then on the industrial and auto side, is that sort of assuming just shipping to normal seasonality with the 5% to 15% and no restocking?

And I'm trying to understand what's the sort of normalized sell-through for auto and industrial at this point given all the volatility.

So I'm going to -- Josh, I'm going to go kind of a little backwards on this. The industrial and auto, like I said, for us to be confident, given what's happening in the world and what could happen, we thought that it would be prudent really to, as you indicated, just say we're back to kind of a normal -- I don't know, seasonal is the right word, but normal demand cycle as we start shipping to where consumption and our revenue are aligned versus undershipping demand as we have been. That's not -- I think that's not aggressive at all.

On the comms and compute side, I think we are pretty confident that the $20 million to $40 million range that Ford said earlier, we have good support in with the fundamental driver, which is the CapEx of the industry.

And that range aligns with what we see the CapEx for the industry being. And as Ford said earlier, we could have upside as our footprint actually grows against that CapEx based on the design wins we have and the opportunities that we're beginning to ship to now. So that's -- hopefully, that answers your question.

I don't I think that AI mix, and this is something -- you can jump in and correct me if I'm wrong. The AI mix will continue to grow just as the nature of the prevalence of AI in the overall footprint in both comms and compute and industrial grows as well. We haven't broken that out separately. Is that helpful?

Yes, it was. I appreciate all the color there. And then for my follow-up, I mean, you guys called out a bunch of different applications for companion chips and use cases that you're seeing in both general purpose and AI servers. Could you maybe speak to which ones you're seeing the most traction for now and which ones you expect to grow more as Nexus and Avant product portfolio layers into the mix?

Yes. Josh, I'm not sure we're prepared to break it up to such a level of detail by application. I think as Lorenzo was saying, what gives us confidence, if you see the overall CapEx expected to grow 20% to 30% next year, we expect to grow faster than that underlying CapEx. Hence, the numbers that we're mentioning. I think starting to break this down by application is beyond this call. We can discuss on post calls, if you like.

We take the next question from the line of Melissa Weathers from Deutsche Bank.

Can you hear me now?

Yes.

Great. Sorry, having some phone issues. Maybe my line needs more FPGAs. I guess -- sorry if I missed this in an earlier question, but you guys have been talking about PQC a lot. And I'm still a bit uncertain like how big that market could be.

So could you just talk about like what -- how big is that opportunity for you guys? Is it growing? What are you seeing? Just anything else on the PQC side?

I mean PQC is a big driver of our security adoption. We do believe we're ahead. It is being mandated in all systems right now. So we're being designed in now into both comms and compute type of applications.

And we haven't broken it up, Melissa. So again, I think we're probably not going to be ready here to break down percent of applications per use case on this call.

That's fair. And then on the pricing side, I just want to make sure as we're kind of bottoming out in the semi cycle, are you seeing anything interesting on the pricing side either from your industrial and automotive customers or maybe on the content compute side, a little bit more leverage? Just anything on pricing?

No. So our pricing strategy is obviously to price to value, as you can tell by our gross margins. And as I said earlier in the commentary around gross margin, that varies a lot between the segments and within the segments between the different design wins depending on how the customers are going to use it.

We also have had consistently a strategy of working with our customers to provide a long-term value proposition. So we're not tactically taking advantage necessarily of certain demand trends because we think we're going to work a long time with these customers and over time, we'll get the full value of what we offer from our product set to these customers. So we do see, at the same time, candidly, pressure from our suppliers based on demand, and we're managing through that on both sides of the supply chain with them and with our customers to mitigate the impacts of that.

We take the next question from the line of Quinn Bolton from Needham & Company.

I just want to come back to the auto industrial. That business has been running at about $50 million a quarter all year, and that's been very consistent. You've also said you've been undershipping by $15 million to $20 million every quarter.

So consumption feels like it's been very consistently at $65 million to $70 million for the last 4 to 5 quarters. I guess I really struggle to see why would that drop to something in the $55 million range implied by your 5% to 15% guidance in 2026, that just seems like there's a massive change in the consumption level implied by your guidance.

Yes. I'm -- I think the challenge in reconciling these -- what you're saying and what we're seeing is there is a direct aspect, meaning non-channel piece of our industrial and auto. And what we're talking about in terms of where you see the phenomenon of undershipping demand, it's really primarily in the channel.

But I don't want to underwrite, if you will, that level of increase into 2026 until we actually see the strength underlying the overall industrial and auto business.

And the backlog coverage -- sorry, go ahead Ford.

No. What is the number that you're mentioning then for....

He's saying 52%, 47%, 50% something like that in terms of the industrial auto through the year. Is that right, Quinn?

Yes, I'm just -- industrial and auto.

That's our number.

So to answer your question -- so your number for -- well, maybe we should do this reconciliation after the call? Quinn?

Yes. Yes, we can take offline. The other question I had just on the comms and compute, guys, you talked about servers up over 80% year-to-date, and I think the comms up over 60% year-to-date.

Yet if I look at the total comms and compute in '25, even with a healthy increase Q-on-Q in December, kind of looks like you're going to be up mid-20s for the year.

And so it feels like there must be a portion of comms and compute that's down pretty substantially to bring the total bucket to mid-20s when I think the 2 biggest components of comms and compute, servers and wired comms are growing 3 faster. So what's the offset?

Yes, yes. Yes, the offset is the client business. So we had a big client business last year that pretty much has disappeared now.

So we've had a big headwind in the client business, Quinn, with 3 client OEMs, large 1 and 2 medium-sized ones. And that headwind disappears now.

We take the next question from the line of Chris Myers from Rosenblatt Securities.

This is Chris Myers on for Kevin Cassidy. First, can you guys just share your current view on the automotive market right now? Are you guys seeing any inventories come down or any backlog start to build up again?

And then on the data center side, what's the trend that you guys are seeing in terms of your dollar content per traditional server versus AI servers?

And then do these devices play different roles in the 2 different types of systems?

Yes. We -- on the AI server versus traditional server, we definitely have more content in AI servers because of the aggregated nature of these servers. So we -- so yes. And then on the...

On the auto, first, let's remind everybody that the automotive business is a very small part of our business. And what we see on a global basis is not -- we don't see any tailwinds there yet, except the place that seems to be moving the fastest is in China automotive.

But again, that's a portion of our overall business, which is a small part of our overall industrial and auto business. So that's what we see. We don't see auto strength anywhere but China.

Ladies and gentlemen, this concludes the question-and-answer session. I would now hand the conference over to Lattice Semiconductor's Vice President of Investor Relations, Rick Muscha, for closing comments.

Yes. Thanks, everyone, for joining us today. We'll be attending the following investor events this quarter: the Stifel 2025 Midwest One-on-On Conference in Chicago on November 6, The AllianceBernstein 2025 Buy-Side Small and Mid-Cap Summit in New York City on November 18. And then lastly, the UBS Global Technology and AI Conference on December 2 in Scottsdale.

This completes our call. Thank you very much for your participation, and have a good evening.

Thank you. Ladies and gentlemen, the conference of Lattice Semiconductor has now concluded. Thank you for your participation. You may now disconnect your lines.