TeraWulf Aktie

Greetings, and welcome to the TeraWulf 2026 First Quarter Earnings Call. [Operator Instructions] As a reminder, this conference is being recorded.

It is now my pleasure to introduce your host, John Larkin, Senior Vice President, Director of Investor Relations. Thank you. You may begin.

Thank you, operator. Good morning, and welcome to TeraWulf's First Quarter 2026 Earnings Call. Joining me today are Chairman and CEO, Paul Prager; CTO, Nazar Khan; and CFO, Patrick Fleury.

Before we begin, please note that our remarks today may include forward-looking statements. These statements are subject to risks and uncertainties, and actual results may differ materially. Words such as anticipate, expect, believe, intend, estimate, project, could, should, will and similar expressions are intended to identify forward-looking statements. For a discussion of these risks, please refer to our filings with the SEC available at sec.gov and in the Investor Relations section of our website. We will also reference certain non-GAAP financial measures. Reconciliations to the most comparable GAAP measures are available in our earnings release and filings.

With that, I'll turn the call over to our Chairman and CEO, Paul Prager.

Thanks, John, and good morning, everyone. The first quarter of 2026 was about execution. We exited 2025 with an established platform, including sites, contracts, capital and strategy. And what you are seeing in Q1 is the early conversion of that foundation into operating performance and recurring revenue. That strategy continues to guide everything we do, controlling energy advantage sites, engineering infrastructure around power and contracting long-term credit-backed AI capacity. At this point, we are moving from formation to delivery. You can see that most clearly in our operations.

During this quarter, we continued scaling our HPC platform at Lake Mariner with 60 megawatts of critical IT capacity energized and generating revenue as of March 31. Importantly, this is the first period, where HPC leasing is meaningfully reflected in our financials, contributing $21 million of lease revenue during the quarter. At the same time, we are transitioning portions of our legacy mining footprint to support higher-value HPC workloads. That transition is deliberate. Mining served its purpose. It enabled us to build infrastructure, monetize power and develop operational expertise. But the future of this platform is contracted long-duration compute infrastructure. That same focus on execution carries through to development.

We continued advancing construction for Fluidstack and Google at Lake Mariner, while incorporating customer-driven design refinements. These are not disruptions. They reflect the reality of building infrastructure for sophisticated counterparties. We are building to evolving hardware and tenant requirements, not in anticipation of them, and that discipline reduces execution risk and improves long-term outcomes.

CB-3 at Lake Mariner remains on schedule, and we are working closely with Fluidstack and Google to coordinate energization with hardware deployment. Execution at this scale requires tight alignment between infrastructure readiness and customer deployment, and that coordination continues to progress well. At the same time, we are expanding our platform. Since year-end, we have added meaningful new power-backed capacity, including the Hawesville, Kentucky site, a large-scale campus with immediate power availability and significant expansion potential.

In parallel, we're progressing the Morgantown acquisition in Maryland, which remains subject to regulatory approval. We currently expect a FERC decision in the mid-summer time frame. Strategically, Morgantown is a highly attractive asset in one of the most power-constrained regions in the country. And that ties directly to how we think about power. We are not pursuing power as an input cost. We are structuring power as a core asset.

If you step back, the broader AI build-out is accelerating, but it is increasingly constrained by power, including interconnection delays, transmission limitations, the need for new generation and transmission and the capital and credit required to bring these assets online. The constraint is not GPUs. It is power. And in that environment, the industry is moving toward integrated campuses, where generation, storage and compute are designed together. That's exactly how TeraWulf operates.

We are fundamentally a power company that builds digital infrastructure, not the other way around. We understand how to source and control power, how to permit, operate generation, how the grid behaves and how to integrate these systems at scale. There are very few teams that can do all of that credibly, and that continues to differentiate TeraWulf. Against that backdrop, demand remains very strong. We continue to see active engagement from hyperscalers and AI compute platforms across that portfolio. At Lake Mariner, we are executing against existing contracts. And in Kentucky, we are engaged in advanced negotiations.

More broadly, when you look at the pipeline, the opportunity set continues to evolve. Given the state of interconnection queues, we believe there will be an increasing opportunity over time to partner directly with utilities to develop new sites and aggressively advance existing sites. Utilities are going to be looking for partners with the experience and credibility and the capital required to actually deliver on these projects, and we believe that TeraWulf is very well positioned.

We also believe that the market is moving towards a higher bar for execution certainty around grid access itself. Securing queue position alone is no longer enough. Utilities and regulators increasingly require confidence that projects can deliver the infrastructure, equipment and financial support needed to come online at scale. Over time, factors like procurement capability, delivery credibility, financial assurance become increasingly important differentiators. That dynamic favors scaled, well-capitalized operators with real development and power experience.

Our platform is designed to operate across all three paths to power, immediate access as we have in Hawesville, bring your own generation as we are pursuing in Morgantown and increasingly, utility partnerships as interconnection queues are rationalized and prioritized. That is a structural shift in how this market will develop, and it plays directly to TeraWulf's strengths.

Now let me spend a moment on Kentucky. I have said that we expect to have a customer in place in the second quarter, and I remain highly confident. We are in late-stage negotiations. The process has been extremely competitive and engagement is exactly where we want it to be. At the same time, we recognize that projects like these raise real questions around environmental impact and power costs. We take those seriously. Our approach is to engage early and transparently and to help communities understand how these facilities serve as long-term economic engines through jobs, investment and infrastructure, while being thoughtful and responsible to their impact. That is core to how we execute.

And importantly, even with that confidence, our approach remains disciplined. We do not build on speculation. We contract first, deploy capital second. That discipline shows up in how we evaluate opportunities, focusing on durable power control, scalable development, credit-backed counterparties and capital efficiency. We review a significant number of opportunities, but only a small subset meets that high bar. And that same discipline applies to capital.

We entered 2026 with substantial liquidity and a fully funded development pipeline with $3.1 billion of cash and restricted cash on the balance sheet at quarter end. That capital is being deployed into contracted or actively commercializing assets, not speculative builds, and our financing strategy continues to align capital with long-term cash flows. You can see the transition of the business model clearly in our financials. In the quarter, digital asset revenue was about $13 million, while HPC leasing contributed $21 million. That shift will continue, and the business will be increasingly driven by stable, contracted, high-quality credit-backed revenue. That is the end state we have been building toward.

So when you step back, the TeraWulf story is consistent. The strategy is unchanged. The platform is in place. The sites are secured. Demand is strong. Capital is in place. We are locked and loaded. From here, it is about execution, delivering capacity, energizing megawatts and converting contracts into durable recurring cash flow. That is what will define 2026.

I'll now turn it over to Nazar to provide an update on construction and delivery.

Thank you. As Paul mentioned, execution across the platform continues to progress well. Let me provide a brief operational update. Starting at Lake Mariner, execution continues to progress well. During the quarter, the second data hall in CB-2 came online, bringing all Core42 capacity into service and generating revenue as of the end of the first quarter.

Turning to the Fluidstack deployment, including CB-3, CB-4 and CB-5, all major project time lines remain unchanged from our prior update and execution across the campus continues to track well against schedule. For CB-3, we are on track to complete our defined scope by the end of May and are actively coordinating with Fluidstack and Google on final energization and lease commencement. CB-4 and CB-5 remain on track for delivery in the third and fourth quarters of 2026, respectively.

In parallel, we continue to advance development and construction activities across the broader platform, including our joint venture in Abernathy, Texas. As a reminder, the Abernathy joint venture project is being delivered under a lump sum EPC agreement with Hypertec with Fluidstack leading development and construction. While execution remains on track relative to our scope commitments, we are working closely with Hypertech and Fluidstack alongside the ultimate compute end user to finalize deployment sequencing and delivery timing. As those milestones are finalized, we will update the market accordingly.

More broadly, these are large-scale infrastructure deployments that require an immense amount of integration. Success depends on maintaining tight coordination across the power infrastructure, equipment procurement, construction delivery and customer hardware deployment. As these campuses scale, execution certainty increasingly matters just as much as site selection or interconnection access. And that is an area, where we believe our team is highly differentiated.

With that, I'll turn it over to Patrick to review the financial results for the first quarter.

Thank you, Nazar. The first quarter 2026 results reflect a business in transition with a meaningful contribution from the rapidly growing HPC lease segment and strong operating profit at the Digital Asset Mining segment due to its flexible load profile and demand response participation. In the first quarter of 2026, revenue was $34 million, down from $35.8 million in 4Q '25, primarily driven by lower Bitcoin production. Importantly, HPC lease revenue increased to $21 million in Q1, up 117% from $9.7 million in Q4. In March, CB-2 achieved RFS or Ready for Service and the lease with Core42 commenced. We have now delivered all 60 critical megawatts of capacity to Core42.

As additional buildings come online in 2Q, 3Q and 4Q, revenue mix will continue shifting towards stable contracted HPC revenue. Cost of revenue, exclusive of depreciation, decreased 88% from $18.9 million in Q4 to $2.4 million in Q1. Demand response proceeds recorded as a reduction in the cost of revenue increased to $14.1 million in Q1 from $4.4 million in Q4. Operating expenses increased as we scaled the platform to support HPC deployment. Quarter-over-quarter operating expenses rose to $11.2 million from $8.8 million.

Regarding the HPC leasing segment profitability, as presented in Note 17 of our 10-Q, the as-reported quarterly segment profit margin is approximately 50% versus our long-term guidance of approximately 85%. That difference is driven by three factors: first, $2.1 million of tenant fit-out revenue and associated costs during 1Q. TFO carries a modest margin as provided for under the HPC leases. Second, $3.5 million of pre-revenue operating costs at WULF Compute. And third, $2.1 million of development costs across our 1.75 gigawatt portfolio of uncontracted development sites.

Adjusting for those factors yields approximately 85% segment profit margin in 1Q. SG&A expenses also increased as we scaled the platform to support HPC deployment. Quarter-over-quarter, SG&A expense rose to $127.8 million from $66.6 million. After adjusting for stock-based compensation, SG&A decreased from $60.1 million in Q4 to $26.3 million in Q1, in line with our prior guidance of $75 million to $100 million for 2026. Depreciation was flat at $28.5 million in Q1 versus $27.7 million in Q4. Impairment of PP&E in 1Q was $25.7 million, $8.9 million of which was driven by the shutdown of a mining facility and cessation of Bitcoin mining operations to support the HPC operations and $16.8 million related to acquired assets of Hawesville, Kentucky associated with asset retirement obligations.

Interest expense in Q1 was $67.1 million compared to $62.4 million in Q4, and we recognized interest income of $29.4 million in Q1 compared to $31.5 million in Q4. The increase in net interest expense was expected following our capital raises at TeraWulf and WULF Compute in the second half of 2025. Actual interest paid during Q1 was $5.3 million compared to $6.9 million in Q4. Change in fair value of warrant liabilities in 1Q and 4Q was losses of $216.3 million and $5.2 million, respectively, related to the Google warrants. This is a noncash loss and does not affect our liquidity.

Equity in net loss of investee, net of tax for 1Q was $11.5 million compared to a net loss of $4.1 million in 4Q, which represents TeraWulf's 50.1% share of the net loss of the Abernathy joint venture, which was formed in October 2025 and has not yet commenced operations. Our GAAP net loss in 1Q was $427.6 million compared to a net loss of $126.6 million in 4Q, with the increase primarily driven by noncash fair value adjustments related to the Google warrants and noncash stock-based compensation. Our non-GAAP adjusted EBITDA in Q1 was a negative $4.1 million, an improvement from negative $50.9 million in Q4.

Turning to the balance sheet and liquidity. As of March 31, 2026, cash and restricted cash totaled $3.1 billion. Total assets amounted to $7 billion with total liabilities of $7.1 billion. Regarding liquidity, as detailed in our 1Q investor presentation on the slide titled Capital Structure, as of March 31, the TeraWulf parent entity held approximately $300 million of available unrestricted cash. This figure increased to approximately $1.5 billion when incorporating the impact of the equity we raised in April. As of March 31, 2026, WULF Compute remains on budget and had approximately $2.8 billion of gross cash or $2.3 billion net of debt service reserve and interest during construction accounts, with $1.5 billion of CapEx spend complete and $2.2 billion remaining.

With regard to the Abernathy JV, as of March 31, 2026, the JV had approximately $1.4 billion of gross cash or $1 billion net of debt service reserve, interest during construction, letter of credit and HoldCo LockBox accounts with $0.4 billion of CapEx spend complete and $0.9 billion remaining.

With respect to Kentucky, subsequent to the quarter, we repaid the $100 million draw on the bridge credit facility and terminated the facility. We expect a portion of the approximately $1.2 billion of equity raised year-to-date will fund TeraWulf's equity contribution to the Kentucky project. Demand for near-term power remains strong, and we are targeting 480 megawatts online in the second half of 2027.

In summary, 1Q reflects a business in transition from volatile Bitcoin mining revenue to stable contracted HPC revenue. Mining continues to strategically support this transition. Contracted HPC revenue is ramping. Liquidity at the parent and its subsidiaries remains strong.

With that, operator, we are ready to take questions.

[Operator Instructions] The first question is from Mike Grondahl from Northland Securities.

Two questions. One, Paul, could you talk a little bit post Kentucky, what sites are you most excited about? And then secondly, kind of getting at a pipeline of new sites, would you expect to add any new sites in '26, '27, like you've added Kentucky and Maryland so far?

Mike, thanks very much. Yes, I'm very excited about Kentucky, not only for what we've been talking about and the customer that we'll be signing, but it has expansion possibilities as well, working very well with the utility there, Big Rivers. And so I think there's great opportunity for expansion. I think that Kerri Langlais run our development and M&A teams. I think we indicated last year, we looked at hundreds of possible sites and selected a few just to move on with. One of those was Morgantown. We're really excited about it. We anticipate mid-summer FERC decision. It's a very exciting site because of its location for data center use, but also because they need generation. Power is really constrained there.

So we're going to be doing something that's really good for the grid, really good for state ratepayers and really good for our next data center customer at that site. So I think we're really, really constructive with respect to it. We're going to continue to perform on Lake Mariner and build that out. And then, yes, Kerri's team has worked hard. I think there's a couple of things that they like a lot. It would be premature to talk about them right now, Mike. But again, I think Kerri's experienced teams -- their capabilities will be demonstrated in our ability to add a couple of more pretty compelling sites to the queue. I think we're very, very excited about building out Kentucky right now. And our focus has to continue to be on execution because you're only as good as your last job, and we need to perform for our customers. So that's my primary focus. But yes, I'm very excited about what we have in the pipeline.

The next question is from Stephen Glagola from KBW.

Can you elaborate on the opportunity to partner directly with utilities on future sites? And are there specific markets or power types, where you see the greatest opportunity? And what could this mean for -- whether it's project economics, development time lines and scalability?

Stephen, it's Nazar here. So with respect to where we see things going, more and more so, you want certainty throughout the chain. So that's whether you need new transmission or generation, who's going to fund, who's going to bring the customer and who's going to fund that build-out. So what we're seeing is utilities may have sites that used to have generation or interesting sites within their territories or their jurisdictions. They want load and they may need generation to kind of come with that load. And some of the utilities can do that on their own.

You're seeing more and more so of the integrated utilities, I think, taking a more aggressive approach to trying to bring load in and say, "Hey, we're a one-stop shop, come here." And there are other folks who may not have those tools or may not be in the financial position to support all of that generation on the front end. And so they're looking to work with partners like us, who can help bridge that. So there are a number of discussions that we're having kind of across the country with respect to those. Again, it's how do we bring generation into a territory so that we can both bring the supply that's needed as well as bring the load that would use that generation as well. And can we be either the bridge solution to get that generation going or can we be the solution to get that generation going.

Okay. Great. And then if I could just ask one more follow-up on Cayuga Lake. Just an update on the site plan review process there across both Phase 1 and Phase 2, including maybe expected timing for the initial power availability for the 150 megawatts and then the path to expanding to 300 megawatts?

Yes. Cayuga, we're in the planning permit process. We have another meeting coming up that we will be preparing for. The first planning meeting was, I think, really good. They were asking a lot of constructive questions. They're really getting into the details of what we plan to do on site. I can't give you a window into their approval process beyond that it seems to be, in my opinion, going according to the rules. We're engaged. They are engaged. We have homework to do to get back to them with all the responses to the questions they've asked. They're asking good questions. They're asking about noise. They're asking about security. They're asking about fire security measures. They're asking about traffic on the roads. They're asking about all the things that they should in the best interest of their community, and we're trying to be responsive to that.

We're still very, very excited about that site. It's pretty great. There are a couple of other things that are going on at that site. There's also going to be a solar facility from AES. A battery storage group is interested in that. So I think the town is -- the Town Planning Board is trying to sort of integrate all those groups in terms of what they're trying to do at their various locations on the lakefront. But I think at the end of the day, we're making good progress. They're being a good partner, and we are excited about our progress there.

The next question is from Brett Knoblauch from Cantor Fitzgerald.

Paul, on, I guess, both Lake Mariner and the Maryland site, maybe just starting with Lake Mariner, I think the slide deck shows kind of 750 megawatts. I guess, is there an update to the additional power that you guys were in queue? Have you received that yet? Or when should we expect to receive that?

And then on Maryland, maybe just the cadence of what that progress looks like. Would you look to sign a contract before procuring the generation equipment? Would you look to maybe partner with an IPP at that site or kind of do everything yourself? How should we think about just the cadence of that site, which obviously can be very big?

Yes. Maybe I'll start Naz, and then I'll turn it over to you. In terms of Maryland, we haven't received FERC approval yet, right? So we are receiving a lot of inquiries from customers with respect to the site just simply because of the sheer scale of it. And again, its location in that Washington, D.C., Northern Virginia that Maryland corridor is very exciting for data center use. And it's rare to find it in a brownfield site as opposed to some farm field out in Virginia. This is an industrial site that's going to be reused. And so the county is really excited about it as well about the jobs and the idea that, that site is going to be mitigated and reused.

So the answer is, we got -- we have to wait on FERC approval. We're not going to spend a whole lot of money until we have that. We're not speculators. We're highly confident that FERC approval will come. And at that point, our first focus will be on ensuring that we're in compliance with our grid obligations because she is a generator into the grid with her existing units. And our intention is, as you know, to build a much larger gas facility at that site, so we could be a surplus generator.

Naz, do you want to take it from here?

Sure. Brett, on Lake Mariner, the incremental 250-megawatt interconnect, we're expecting feedback from the ISO here midyear. So we haven't received anything yet, but we've been told from the ISO as recently as last week that we should be getting that feedback on schedule. So as soon as we have that, that will then drive timing on when that incremental 250-megawatt will be available. And pretty soon thereafter, we'd be looking to market that capacity as well.

The next question is from Nick Giles from B. Riley Securities.

Just wanted to ask on Justified. What are really the key items still being determined there? I mean, should we expect that the customer could be determined at this stage and you're just negotiating terms? Or would it be that you still have multiple customers kind of in the mix that are fighting for that capacity? Just kind of any more color on where things stand would be really helpful.

Yes. I want to be careful here. We've run a very competitive process. And as you know, the space is all about immediately available power, prompt power. And so that's what excited people about the site, plus it's mid-country location. I think that the result of a very competitive process are some terms that we're excited about that we think you will find competitive in the market today. And now we're working through the details of what that looks like in terms -- from the customer side. I remain highly confident that we're there, as I said we would be before the end of the second quarter.

But these deals are constantly evolving. There's a lot of disruption in the space, and we want to make sure that we get the absolute best contract terms for our investors and shareholders. On the development side, we're in great shape. We -- our team turned to and we got that agreement with the utility, and so we're all set to go. We've, as you know, initiated limited notice to proceed to Fluor. That relationship continues to grow in the most constructive way. So we're really, really excited about what we have going on at Hawesville.

The next question is from Tim Horan from Oppenheimer & Company.

Are there any terms you'd rather get in the new contracts that maybe you didn't get in the past? Maybe just how you're thinking about how the terms change? And then Justified, if you get everything you need here in the next few months, when do you start construction? When do you think that comes online? And then just lastly, it sounds like there's more locations or power in that area. I mean, just any more color on that.

Tim, it's Nazar here. With respect to just the contracts themselves, I think we've -- the initial contracts we had were 10 years. We were some of the earlier contracts that were signed. More recently, the term has been 15 years. And so that's been, I think, a term that's become pretty well accepted within the market. When we look at the kind of the construct of this, we have a customer that's going to be there for 15 years. Inevitably, there will be some hardware changes that occur during the course of that time. And so making sure that there's an alignment on how that transition will occur and again, whether it happens halfway through it, 5 years through it or 10 years through it, is not the point. It's more of how do we manage that.

So when we look at these things, it's really putting together both on the front-end scope. And as Paul mentioned earlier, I mean, things change on that side month-to-month as well. The hardware that's being deployed here depending upon when you're coming online is going to be different if it's 6 months sooner or later. And so that's one thing that feeds into the overall basis of design. And given, again, just how quickly things are moving, we thought we had a final design that was going to work for a number of years, and we did it with Fluidstack and Google towards the end of last year. Again, there's been a couple of iterations on hardware from vendors, and we work with all of the vendors, right?

We've got AMD, NVIDIA, Google TPUs. And so we have a number of different OEM hardware equipment. And again, each one of those vendors have an evolution in the hardware they're delivering. So that's another piece that works through it. And then underlying it all is ensuring that the creditworthiness of the counterparty that we have is paramount. And so that's the final piece that we look to in putting these things together is to ensure that, that counterparty credit is aligned with what we've done in the past. And so those are some of the things that go into it. But hopefully, that gives you a sense of just some of the changes and nuances we've had here over the past couple quarters.

The next question is from Chris Brendler from Rosenblatt.

Just wanted to ask from zooming out a little bit, sort of thinking about the demand environment and certainly seems to continue to get stronger and stronger. As you look to the next several years, do you think that we're in a window of opportunity here for sort of higher profitability on these contracts that will dissipate over time? Or do you think that we'll see continued upward pressure on yields, cost and lease rates and terms given the lack of power that's available for this massive AI build-out?

This is Paul. I appreciate the question. Demand is real. I mean the customers are pushing hard to get more prompt power, but they're willing to contract for longer periods of time. I think that the disruption in the space will be in the technology, but it won't be in the fact that they require energy use. Some units may end up being more efficient, but that will just enable more compute time on the same amount of energy. So I think at the end of the day, what we are anticipating is continued pressure from customers based on the strong demand for quality energy. And they're focusing on sites where they think they could be sustainable. They want to know that they're going to places where the regulatory framework is going to be stable, where the energy costs are going to be okay and not in excess of where they anticipate the market is going.

I think for us, what we need to do is continue to ensure that we provide low-cost energy, great regulatory framework in our sites, sustainability in terms of the energy mix and stability of the energy price. And then some sort of regional diversification for a lot of our customers so that they're not all focused just in one place like West Texas, but they had places where they could do their high-power compute throughout the country. I think that the market will continue to improve for some time because you just can't bring power online that quickly.

We have a lot of hopes in our country for continued power development, but it takes time. It takes time to permit. It takes time to refine the energy mix for the customer. I just think it all takes time. So I don't think the window is too narrow. I think Steve Byrd's talked about the shortage of electricity in this country for a couple of years now being pretty significant. I don't see that shortage. I don't see supply coming online fast enough to meet the demand. And so the need is not abating. So we're pretty constructive.

But that doesn't drive our strategy. Our strategy is more driven on what we know we can do. And that's why we've told the world we're out there to put 250 megawatts to 500 megawatts in the ground in any given year. We're highly constructive and confident about our abilities to do that. And if there are some larger projects that come along and customers want to participate in that, then we'll work with them, and we will. But we think slow and steady is going to win this race. We anticipate to be doing exactly what we're doing several years from now.

That's very helpful, Paul. I appreciate that. And a question for Patrick, kind of a similar question, certainly seeing increasing acceptance and demand from debt investors for companies who have pivoted from Bitcoin mining, which was not that financeable to HPC. And I think at first, it was a little tougher to get debt deals done, and now it seems like it's getting easier, terms getting better, spreads are tightening. Just wanted to hear Patrick's perspective on the financing outlook and how things have improved.

Chris, it's exciting, right? I mean come a long way in a very short period of time. And now all of us are starting to look more like traditional S&P 500 companies, right, with project finance debt, real revolvers. I think one thing we put in the release today that I think our whole team is really proud of is $250 million revolver from 8 different banks, that a group of global leaders and financial institutions. That's a big accomplishment.

And I know some of our peers are doing that, too. So yes, I think as a whole, the entire space is just becoming more legitimized. And not only that, what's exciting for me, I mean you've seen us use a bridge, a revolver, high yield bonds, right? Project finance bonds. I think you'll see us, one of our peers just opened up the loan market. So, yes, I think it's becoming more common place with the rating agencies, with high yield investors, with IG investors.

And, I constantly come back to, for us, like, where we are trying to go is, I want flexible debt structures, meaning structures that I can call in a couple years because where I -- ultimately I'm trying to drive things and us as a team are trying to drive things is we want to be an investment grade company. And, I had a front row seat early in my career to watching, Cheniere go through the capital raising process and ultimately, financing projects at the project level is very efficient when you're first getting started. But then once the projects are up and running, right, the way the massive unlock for the equity happens is you take out those project financings with long-term bullet investment-grade debt.

And that's where I think we are trying to drive our bus. And I think the space will follow because there's a lot of cash flow that comes out of these projects. And once you've amortized the debt down to a comfortable level, there is a massive unlock in free cash flow that happens for the equity, and that's only a couple of years away for us.

The next question is from Brian Dobson from Clear Street.

So as you're in conversations with new clients, how would you characterize potential contract terms compared with, call it, similar conversations last year? Have your hurdle rates changed? And what do you view as the biggest constraints on future development at this point?

Brian, it's Nazar here. So there's a handful of customers that we all want at these sites. And so in the near term, assuming you have access to the power, right, that's where the discussion starts. I mean there's one constraint is just people's time. The 6 customers, 8 customers everyone wants are doing a number of different things. For us, we have one site that we're focused on at one time. And so that one site is 100% of what we're doing and it's probably not 100% of the customers that we want of what they're doing. So one is just, again, finding that alignment between the customer and where they're at, what they're looking for and where we're at.

And I would say that's a more meaningful part of the art of getting this stuff done than people think. And so that's one is just, again, kind of the timing alignment on where you are. The other piece is these larger customers are looking to build zones of capacity. So it's not just a site in isolation that they look at in a state, but it's really once they go into a place, they're looking to have a footprint. And so they also want to have a view to -- can in those locations, can they continue to expand. And so whether that's with you or with others, but in those areas, can they kind of expand a footprint. So it's not just a single site that they're at, but it's multiple sites that they're at. So that's another piece that gets back to this overall alignment is, different folks are in different phases in different parts of the country. And so one group may be ramping up in the Southwest, another may be ramping up in Kentucky.

And so that's where part of our work is really to kind of find that overlap as well with what we have and what we're bringing to market and where those folks are. So those are some of the things I just think from a market and demand perspective that we're working through. I think as Paul mentioned earlier, the 250 megawatts to 500 megawatts that we have guided the market to is well in line with what we can deliver. And again, I think across the 6 or 8 customers that we want, it's also meaningful enough but not so large portion of what they want as well. And so there's ample room for us to be able to deliver that with the capacity that we have.

And so -- and the other piece that I'd just kind of to finally add in is in trying to build a model with a couple of folks, where we can maybe even kind of be ahead of the game in understanding where their growth plans are. And so again, as I was saying before, if there's a group, one of the folks is looking to build out a zone in one part of the country, can we be the resource for them to kind of identify the next site within that zone versus us just coming to them when we have something in hand. And so that's another kind of discussion that's ongoing with counterparties that we have is not just kind of focused on what's in front of us right now and negotiating a lease agreement, but it's also a little bit more strategic and saying, "Hey, in this zone, should we be looking for other things for you as well."

And Brian, it's Patrick. Just to follow up on your economic question. It's the same. We've kept the same sort of discipline and principle from day 1, which really, again, is yield, right? And that's just your NOI yield effectively over your cost to build. That's what we try to target. But I think the added element, right, is that has to align with also the credit support that you're getting from the counterparty, right? Is it direct on balance sheet? Is it credit sleeved? How does that look? But that's really the focus because as Nazar mentioned, I think a lot of the other lease terms have moved to a market rate across multiple providers, multiple sites, others, but that's, I think, still the economic focus for us.

The next question is from John Todaro from Needham & Company.

Congrats on the progress so far. I guess just on Kentucky, can we say if there's kind of a frontrunner at that site and if we're thinking direct IG hyperscaler there? And then I have a follow-up.

Yes, we like somebody a lot for that site. It was a very competitive process, and I would assume it's going to be investment-grade, super high-quality customer.

Okay. Great. And then just housekeeping modeling item. Can you remind us on the cadence of the Bitcoin mining ramp down maybe on a hash rate basis?

Yes. John, it's Patrick. I think I'll get back to you on the specifics, but I think today, we're somewhere probably between 5 and 6 exahash. And we won't be putting any significant amount of money into that business. So as some of the miners kind of come offline because we're repositioning buildings or power feeds for the HPC leasing segment, we'll kind of slowly eat into that capacity. I think generally speaking, for the year, though, we're going to try to be in that sort of 5 to 6 exahash range.

And then I think beyond this year, it's kind of wait and see. But I think we will likely be out of that business certainly by the next halving. But it is -- this quarter is a great example. I mean it does provide a very significant service to the grid and provided us with some nice cash flows this service -- excuse me, this quarter as well, and we'll continue to do that in the future. So yes, that's kind of what I would use for the modeling.

The next question is from Martin Toner from ATB Cormark.

Congrats on a great quarter. The political risk of the AI data center build-out seems to be like slowly but surely picking up. And just can you talk a little bit about how you guys are mitigating that sort of like NIMBY development risk that comes from the political risk related to concern over AI buildup?

Sure. Appreciate the question. Listen, people are concerned when somebody comes into their community and wants to put up a big building, and there's a lot of misinformation out there about whether or not they're going to take water from a lake or they're going to be radiating or they're going to drive up electricity prices in the community. So the first thing to do is to help educate the community. And we're very focused on that. We've been developing power plants for over 25 years.

So before data centers were the big ugly, power plants were the big ugly. And then people realized Holy cow, we really do need energy. So we need to be thoughtful and responsible about the energy that we build, but we do need it. We can't just be in denial about that. Likewise, data centers, they're really important. They serve an important goal for humanity. And I think that they do use electricity. And so the market has started to evolve to this notion of you got to bring your own generation or in order to be connected to the grid, we want you to show us your nameplate capacity or we want you to be a surplus generator.

And so the answer is that different communities are going to define what their needs are. We're never going to solve the problem for the person that's just an absolute -- we don't want any progress, we don't want a data center. But what we can do is educate the community on what the data center will bring from a constructive perspective in terms of jobs, economics, what it won't do in terms of -- it's not going to harm the environment, it's not going to drive up electricity prices. And once we work on that education, then we work that all the way through from the everyday citizen in the town hall to their local representatives in government.

I think that it's important for folks to appreciate that we're not all the same. Not every site is equal, not every developer will bring the same degree of thoughtfulness or responsibility or responsiveness to these developments. But that's TeraWulf's mission. We are -- we have customers that are world-class and socially very, very conscious and want to contribute to the benefit of the communities. And that's what we're all about. So I think that you got to pick your spots, you have to understand what you're walking into, and you have to be -- lead with transparency because in that transparency, people will ultimately get very comfortable because we're doing the right thing.

So we're dealing with it in Maryland, where, as an example, it's a very -- it's a state that wiped out its generation, unfortunately, and now it's paying the price for it. So they've come back and they've been very supportive. Governor and the state, the MDE, they're very supportive of building new generation that is thoughtful and responsible and clean. And they're excited about bringing jobs to a county, where -- there was a former plant and now people miss those jobs. So I think you could work with it, but it requires work. It's nuanced, and it will be -- the teams that win here are going to be the teams that have gone about this in a mature and thoughtful and constructive way.

That's great. What caused you to -- what caused EBITDA to be a bit lower than the given range?

Yes. Martin, it's Patrick. So closing out the books on our business has gotten increasingly complicated as we get a lot bigger. And so when we gave that guidance in April, we weren't done as far as inputs into the general ledger and other things. So there were two specific items. One is tenant fit out and the other is there are costs that we incur up at WULF Compute that have to get split between PP&E, meaning like a capitalizable cost and an operating expense, right, as OpEx.

And so in my remarks, as you saw, there was about, I think, $3.5 million or $3.2 million of operating expense at WULF Compute that when we were putting together the preliminary guidance effectively was in PP&E. And as we kind of went through and closed out the final books, that kind of came out of the CapEx basket and went into the OpEx basket. So a lot of that stuff, honestly, will normalize, Martin, as we bring operations up and online, right? We don't -- we won't have to kind of keep going through those buckets carefully every quarter, but that was what drove it.

The last question is from Michael Donovan from Compass Point.

For Maryland, my understanding is that the site currently has live peaker plant capacity and that the independent Market Monitor has raised questions on that front. Obviously, your planned battery capacity appears to more than address that issue once completed. But can you walk us through how you expect to manage the existing capacity during the build-out and what steps you're taking to address regulatory concerns?

Sure. The existing capacity will continue to operate. So there is no plan to do any -- make any changes whatsoever to the existing capacity that's there. So as a part of our plan, all of the things that we've been discussing with respect to battery storage, gas generation are incremental to that. So that's an important point to note that existing 210 megawatts or so of operating capacity will continue to be bidding into the PJM market as a peaker. And incremental to that, we'll be looking to add battery storage capacity, efficient gas-fired generation capacity and load.

So that's kind of the setup there. And so both in discussions with -- if you have read through the filings that we did at FERC, we've kind of reiterated that point. And also with the various discussions we've been having with the local utilities as we've been discussing the potential for adding capacity there, both on the generation side as well as the load side, we've highlighted to them that, that capacity is staying.

This concludes the question-and-answer session as well as today's call. You may disconnect your lines at this time. Thank you for your participation.

Thank you, operator. Good morning, and welcome to TeraWulf Wulf Compute Lender Update Call. Joining me today are CFO, Patrick Fleury; CTO, Nazar Khan; and COO, Sean Farrell.

Before we begin, please note that our remarks today may include forward-looking statements. These statements are subject to risks and uncertainties, and actual results may differ materially. Words such as anticipate, expect, believe, intend, estimate, project, could, should, will and similar expressions are intended to identify forward-looking statements. For a discussion of these risks, please refer to our filings with the SEC available at sec.gov and in the Investor Relations section of our website.

We will also reference certain non-GAAP financial measures Reconciliations to the most comparable GAAP measures are available in our earnings release and filings. This update is intended to provide transparency on construction progress, liquidity and schedule relative to our Wulf Compute financing.

With that, I'll turn the call over to our CFO, Patrick Fleury.

Thank you, John. As of January 31, 2026, Wulf Compute had approximately $3 billion of gross cash or $2.6 billion net of debt service reserve and interest during construction accounts, with $850 million of CapEx spend complete and $2.38 billion remaining. That leaves approximately $200 million of cash cushion, which is incremental to the substantial contingency embedded in the financing structure.

Schedule sequencing adjustments have shifted approximately $16 million of projected revenue in years 2025 and 2026 into later periods. Importantly, design optimization has increased critical capacity from 162 to 168 megawatts across both CB4 and CB5 generating approximately $200 million of incremental revenue over the initial lease term.

The net effect of these adjustments improves projected cash flows and reduces expected debt outstanding at maturity of the senior secured notes by approximately $45 million versus prior projections. I will now turn the call over to our CTO, Nazar Khan.

Thanks, Patrick. Wulf Den and CB1 were delivered in the third quarter and generated revenue throughout the fourth quarter. CB2A is operational and CB2B is expected online in March. By the end of the first quarter, all CORE42 capacity will be energized and revenue producing. .

Following contract execution, Core42 requested certain fit-out refinements typical for hyperscale developments. These were incorporated into the existing construction envelope with a revision to the monthly recurring charge. No penalties are triggered and revenue commencement remains aligned with the customers' deployment schedule. Regarding the fluid stack buildings, CB3 is expected to be delivered in May. After signing tenant-driven layout refinements, typical of late-stage deployment planning were incorporated without changing building footprint or lease economics. The associated revenue timing impact has been fully reflected in the updated financial model.

Importantly, CB4 and CB5 were designed collaboratively with the tenant from inception. These buildings reflect a fully standardized, repeatable design and represent the majority of contracted Wulf Compute capacity.

Several structural improvements materially reduce execution risk. First, electrical redundancy has been optimized and standardized. Second, trade stacking and sequencing have been refined to minimize rework Third, long lead equipment was procured after final design alignment; and fourth, mechanical and electrical systems now follow a repeatable installation model. Execution risk declines as the design standardizes and CB4 and CB5 reflect a mature optimized build. Both buildings remain on schedule with targeted lease commencement dates of the third quarter and fourth quarter, respectively.

Through design optimization, we increased critical IT capacity from 162 megawatts to 168 megawatts per building without impacting the base construction budget. The incremental 12 megawatts across the campus is expected to generate approximately $200 million of additional lease revenue over the initial term.

I will now turn the call over to our COO, Sean Farrell, who will provide more details on each specific building.

Thank you, Nazar. From an execution standpoint, each successive building has incorporated lessons learned and increasing standardization. CB2 transitioned from structural elements to a pre-engineered metal building reducing both scheduled duration and labor variability while maintaining data center performance requirements. Electrical distribution was simplified, lowering capital intensity and improving constructability.

CB3 further standardized UPS architecture and run time parameters, delivering electrical cost efficiencies while maintaining required redundancy and tenant specifications. Integrated skid-based systems improve installation sequencing and schedule certainty.

CB4 reflects full tenant collaboration at the engineering stage. Layout optimization, reduced conduit and piping runs, minimize underground work and shifted redundancy to higher voltage distribution levels, reducing equipment melt and field labor. Containerized UPS systems provide plug-and-play installation further accelerating execution.

CD5 replicates the CD4 design an execution model and benefits from procurement scale, labor familiarity and established sequencing. As a result, we expect continued improvements in schedule predictability and cost control.

Overall, the project has transitioned into a standardized and repeatable development program, materially reducing variability relative to early phase builds.

Thank you, Sean. This ends our call. Thank you for your continued partnership and support.

Greetings, and welcome to TeraWulf Fourth Quarter and Full Year 2025 Earnings Conference Call. [Operator Instructions]. As a reminder, this conference is being recorded. I would now like to turn the conference over to your host, Mr. John Larkin Senior Vice President, Director of Investor Relations. Thank you. You may begin.

Thank you, operator. Good afternoon, and welcome to TeraWulf's fourth quarter and full year 2025 earnings call. Joining me today are Chairman and CEO, Paul Prager, CTO, Nazar Khan, and CFO, Patrick Fleury.

Before we begin, please note that our remarks today may include forward-looking statements. These statements are subject to risks and uncertainties, and actual results may differ materially. During this call, we may use words like anticipate anticipate, expect, believe, intend, estimate, project, could, should, will and similar expressions are intended to identify forward-looking statements. For a discussion of these risks, please refer to our filings with the SEC available at sec.gov and in the Investor Relations section of our website.

We will also reference certain non-GAAP financial measures. Reconciliations to the most comparable GAAP measures are available in our earnings release and filings. With that, I'll turn the call over to our Chairman and CEO, Paul Prager.

Thank you, John, and good afternoon, everyone. 2025 was a defining year for TeraWulf. We said we would transition this company into a scaled, power backed AI infrastructure platform and we did it. Our strategy is simple and disciplined. Control energy advantaged sites, engineer infrastructure around power and contract long-term credit-backed AI capacity. Everything we did in 2025 supports that strategy.

First, we acquired 100% of [indiscernible] electricity and data, eliminating related party complexity and fully integrating power generation expertise into our platform. In today's market, power is the gating factor. If you don't control power, you don't control your destiny. We do.

Second, we secured long-duration site control at Cayuga, up to 400 megawatts at a retired coal facility with real grid infrastructure already in place, brownfield, power backed, scalable. That's our model.

Third, we signed a 450-megawatt lease with Fluidstack supported by Google's credit. That was a platform-defining deal. It validated our model, our execution capability and ability to contract at scale. With Google's warrants, it will be our largest shareholder. That alignment speaks for itself.

Fourth, we replicated the model in Texas through the Abernathy joint venture, proving portability across power markets.

And fifth, we executed the WULF Compute and Flash Compute Financings. These transactions demonstrated that contracted credit-enhanced AI infrastructure supports scalable and repeatable capital structure.

Operationally, we delivered WULF Den and CB-1, began recording HPC revenue and have now delivered CB-2A for Core42. We are building delivering and contracting simultaneously. And since year-end, we added approximately 1.5 gigawatts of additional power back capacity in Kentucky and Maryland. Now let's talk about what actually differentiates us. Power and execution.

The AI build-out is not constrained by GPUs. It is constrained by power, interconnection transmission and increasingly new generation. That's why Morgantown matters. Morgantown is not just another data center site. It is a former coal generation facility in the Washington, D.C., Northern Virginia corridor, one of the most power-constrained data center markets in the world. Our Phase 1 vision includes approximately 500 megawatts of new dispatchable generation, 250 megawatts of battery storage and 500 megawatts of data center load, followed by a similar Phase II. Critically, the site is being engineered to operate as a net generator to the state. We are not just consuming capacity. We are adding it in constrained markets that is the only sustainable model at scale. This industry is moving towards integrated bring-your-own generation campuses. We are well ahead of that curve because we are fundamentally a power company that builds and operates digital infrastructure, not the other way around.

We know how to permit generation. We know how to build generation. We know how to operate generation. We understand grid behavior, and we know how to integrate generation, storage and compute in a way that works for customers and regulators. There are very few teams that can do that credibly and at speed. We are premier among them.

Turning to Kentucky. Demand is extremely strong. We are engaged with every major hyperscaler and several large AI compute platforms, and data room is open, diligence is active, conversations are robust and substantive. This week, we met directly with Governor Beshear and state leadership. The alignment at the state and local level is clear and constructive. Kentucky understands the economic and strategic import of power backed AI infrastructure. This is 480 megawatts, a campus with immediate power availability, expansion potential and strong state support. We are excited and highly confident in the long-term value of this asset.

While we execute on the platform, we are consistently evaluating a constant pipeline of additional opportunities. We review hundreds of sites and most do not meet our standards. We are disciplined around four key elements: power control and durability, scalability in 250 to 500-megawatt phases, signed credit-backed contracts, we do not speculate, and fourth, capital efficiency. We turn sites away all time. But when we do find one that meets these criteria, we move decisively.

Importantly, we already control the sites necessary to deliver our targeted 250 to 500 megawatts of contracted capacity annually through the end of the decade. The runway is in place. Growth from here is execution.

Finally, we are building the team to match the ambition of the platform. We recently added a senior data center construction lead for Meta and have strengthened origination, project management, commissioning and cybersecurity capabilities. This business is one in the trenches of engineering detail, construction discipline and operational rigor, we are staffed accordingly.

So in summary, the strategy is clear. The sites are secured. The capital is in place. Customer demand is strong. The team is built. Now it is about disciplined delivery. Turning contracted megawatts into energized capacity and durable recurring cash flow. With that, I'll turn it over to Nazar.

Thank you, Paul. Let me start with delivery and risk compression. WULF Den and CB-1 were delivered in the third quarter and generated revenue throughout the fourth quarter. CB-2A is operational and CB-2B is expected to be fully online in March. By the end of the first quarter, all Core42 capacity will be energized and revenue producing. Following contract execution, Core42 requested incremental fit-out enhancements. We adjusted sequencing accordingly. The monthly recurring charge was revised, no penalties were triggered and revenue commencement remains aligned with the customers' deployment schedule.

Turning to the Fluidstack buildings. CB-3 is expected to deliver in mid-May. After signing the tenant finalized certain design optimizations, which we incorporated without changing building footprint or [indiscernible] economics. The associated revenue timing impact has been incorporated into the financial model. Importantly, DB-4 and CB-5 were designed collaboratively with the tenant from inception. These buildings reflect a fully standardized, repeatable design and represent the majority of contracted WULF compute capacity. Several structural improvements materially reduce execution risk.

First, electrical redundancy has been optimized and standardized. Second, trade stacking and sequencing have been refined to minimize rework. Third, long lead equipment was procured after final design alignment; and fourth, mechanical and electrical systems now follow a repeatable installation model. Execution risk declines as designed standardizes and CB-4 and CB-5 reflect the mature optimized build. Both buildings remain on schedule with targeted lease commencement dates of third quarter and fourth quarter of 2026, respectively. Through design optimization, we increased critical IT capacity from 162 megawatts to 168 megawatts per building, without impacting the base construction budget, that incremental 12 megawatts across the camp is expected to generate approximately $200 million of additional lease revenue over the initial term.

Finally, Abernathy, our joint venture remains aligned with its fourth quarter 2026 lease commencement and continues progressing under fixed EPC structure, which further limits construction cost variability. Execution requires managing scope, timing and cost in real time. We have incorporated adjustments transparently and remain focused on disciplined delivery. Large-scale AI infrastructure requires active management of scope, schedule and cost. We have incorporated refinements transparently, preserved economics, increased capacity and maintain budget integrity. Execution remains disciplined and on track. With that, I'll turn it over to Patrick.

Thank you, Nazar. The second half of 2025 was transformational for the company. We secured over $12.8 billion of HPC lease agreement, executed $6.5 billion of debt and equity-linked financing and materially strengthened our balance sheet liquidity while carefully managing and minimizing dilution.

Let's begin at a high level before diving into financial details. We are a business in transition and executing on rapid growth. The 2025 results still reflect a meaningful contribution from Bitcoin mining and its inherent volatility, including commodity pricing and complex network difficulty dynamics. Over time, that volatility will decline as long-term credit-enhanced HPC revenues become the dominant driver of results. Importantly, while mining introduces revenue volatility, its flexible load profile has been strategically valuable at Lake Mariner, supporting demand response participation and power cost management. Mining is not a long-term growth focus, but it has enabled the transition. The 2025 results of operations reflect that transition in motion and the balance sheet reflects that we have the capital structure to execute.

In the fourth quarter of 2025, revenue was $35.8 million, down from $50.6 million in 3Q '25, primarily driven by lower Bitcoin production. Importantly, HPC lease revenue increased to $9.7 million in Q4, up 35% from $7.2 million in Q3. For the full year, revenue increased 20% to $168.5 million from $140.1 million in 2024. With Digital Asset revenue of $151.6 million and HPC lease revenue of $16.9 million. We commenced HPC leasing in July 2025 and had energized 18 megawatts of critical IT capacity as of year-end. As additional buildings come online, revenue mix will continue shifting towards stable contracted HPC revenue.

Cost of revenue, exclusive of depreciation, increased 10% from $17.1 million in Q3 to $18.9 million in Q4. Demand response proceeds [indiscernible] reduction in cost of revenue decreased $4.4 million [Audio Gap] prices. Demand response proceeds also increased year-over-year from $8.6 million in 2024 to $17.7 million in 2025. Operating expenses increased as we scaled the platform to support HPC deployment. Quarter-over-quarter operating expenses rose to $8.8 million from $4.5 million. Full year operating expenses increased to $19.7 million in 2025 from $7.6 million in 2024, reflecting staffing and operational readiness.

For context, TeraWulf finished 2024 with under 100 full-time employees and will exit 2026 with close to 300 full-time employees. Let me address the HPC Leasing segment profitability as presented in Note 19 of our 10-K.

The as reported annual segment profit margin is approximately 42%, versus our long-term guidance of approximately 85%. That difference is driven by three factors: First, $1.2 million of tenant bid-out revenue and associated costs during 2025. [indiscernible] carries a modest margin as provided for under the HPC leases.

Second, $4.1 million of development in pre-revenue operating costs. And third, partial period revenue contribution as building [indiscernible]. Adjusting for those factors, yield approximately 77% segment profit margin in 2025, which is consistent with ramp expectations and converging towards our 85% steady-state margin guidance as utilization stabilizes.

SG&A expense also increased as we scaled the platform to support HPC deployments. Quarter-over-quarter SG&A expense rose to $66.6 million from $16.7 million. Full year SG&A accounts for 2025 totaled $147.8 million from $70.6 million in 2024. After adjusting for stock-based compensation, SG&A increased from $39.7 million in 2024 to $94.5 million in 2025. This increase is primarily attributable to an incremental $47.5 million of new hires, strategic growth performance and milestone-based employee compensation in 2025, reflecting the notable scale of execution achieved during calendar 2025. Adjusting for this item results in total SG&A of approximately $47 million in 2025 in line with our prior guidance of $50 million to $55 million.

Depreciation increased to $88.6 million in 2025 from $59.8 million in 2024, reflecting infrastructure placed into service, an accelerated depreciation of $19.6 million associated with certain mining assets transitioning to HPC use. Interest expense in Q4 was $62.4 million compared to $9.8 million in Q3, and we recognized interest income of $31.5 million in Q4 compared to $4.1 million in Q3. Annual interest expense for 2025 and 2024, was $80.2 million and $19.8 million, and we recognized interest income of $39 million and $3.9 million, respectively. The increases in net interest expense were expected following our capital raises at TeraWulf and WULF Compute in the second half of 2025. Actual cash interest paid during Q4 and calendar year 2025 was $6.9 million and $13.9 million, respectively.

Change in fair value of warrant and derivative liabilities in 2025 was a loss of $429.8 million, primarily related to the Google warrants. This is a noncash loss and therefore, does not affect our liquidity. Equity and net loss of [ investee ] net of tax for 2025 was $4.1 million, which represents TeraWulf's 50.1% share of the net loss of the Abernathy joint venture, which was formed in October 2025 and has not yet commenced operations. Our GAAP net loss in 2025 was $661.4 million compared to a net loss of $72.4 million in 2024, with the increase primarily driven by noncash fair value adjustments related to the Google warrants and noncash depreciation.

Our non-GAAP adjusted EBITDA in 2025 was negative $23.1 million, down from positive $60.4 million in 2024. As a reminder, these results are inclusive of significant increases in SG&A and operating expense over the past 12 months as we invest heavily in our HPC business.

Turning to the balance sheet and liquidity. As of December 31, 2025, cash and restricted cash totaled $3.7 billion. Total assets amounted to $6.6 billion, with total liabilities of $6.4 billion. Regarding liquidity. As detailed in our fiscal year-end 2025 investor presentation on the slide titled Capital Structure as of January 31, 2026, and the HoldCo parent entity had approximately $500 million of available cash or approximately $300 million pro forma for the Kentucky acquisition announced on February 2. Regarding project-level capital positions and construction [indiscernible]. Both WULF Compute and Abernathy are fully funded through substantial completion with long-term fixed rate financing, eliminating construction funding uncertainty and reducing reliance on near-term capital markets access. Importantly, we do not anticipate the need for additional equity to fund our currently contracted development.

As of January 31, 2026, WULF Compute had approximately $3 billion of gross cash, or $2.6 billion net of debt service reserve and interest during construction accounts, with $850 million of CapEx spend complete and $2.38 billion remaining. That leaves approximately $200 million of cash cushion, which is incremental to the substantial contingencies embedded in the financing structure. As Nazar noted, schedule adjustments resulted in approximately $16 million less projected revenue in years 2025 through 2026. However, design optimization has increased capacity from 162 to 168 critical megawatts across CB-4 and CB-5, generating approximately $200 million of incremental revenue over the initial lease term. The net effect improves projected cash flows and reduces expected debt and maturity by approximately $45 million versus prior projections.

Finally, with regard to the Abernathy JV, as of January 31, 2026, the JV had approximately $1.5 billion of gross cash, or $1.2 billion net of debt service reserve, interest-bearing construction, letter of credit and HoldCo lockbox accounts, with $268 million of CapEx spend complete and $1.1 billion remaining. That leaves approximately $70 million of cash cushion at Flash Compute with a further $100 million liquidity reserve at the parent JV supported by a $1.35 billion lump-sum CPC contracts with Hypertech.

With respect to Kentucky, we have proposals in hand for secured loan facilities to fund pre-leased development to preserve HoldCo parent liquidity. Demand for near-term power remains strong, and we are targeting 480 megawatts online in the second half of 2027. We do not build on speculation.

In summary, 2025 reflects a business transitioning from volatile Bitcoin mining revenue to stable contracted HPC revenue. Mining has strategically supported that transition. Contracted HPC revenue is ramping, liquidity and contingency are strong. With that, operator, we are ready to take questions.

[Operator Instructions] Our first question comes from Mike Grondahl with Northland Securities.

First, I just wanted to start with Kentucky. That site sounds like the reception has been strong. Could you give us a few more details on the site? And what an ideal customer or lease would look like there?

Mike, this is Paul Prager. Its a fantastic site. This was the site of a former smelter. It's at a transmission superhighway, multiple utilities can service the property. What was most compelling about it was its central location and the immediate availability of power and scale. Demand for the site is extremely strong. Our data room is open, every major hyperscaler and several large AI compute platforms are doing diligence. The conversations to date have been substantive with some written term sheets already coming over to the desk. Earlier this week, I was down in Kentucky. I met with Governor Beshear and state leadership. The alignment at the state and local level is clear, very constructive, very, very, very pro business.

The importance of this project for the local community, particularly the public schools, which Kentuckyians are super proud of is massive. We held a community info session 2 nights ago, we had a job fair yesterday. We filled the auditorium and then some. We're extremely excited and confident in the long-term value of the asset.

As Patrick has said from day 1, we're all focused on the best financial credits to be our long-term customers. I think we've operated that way in the past, and that's the message going forward. I think you'll see a world-class credit as our next customer, for what we're hoping to be a 10- or 15-year deal. I think we'll see that deal happen pretty soon. I don't want to give you a drop dead date, but we're in very, very active and substantive discussions.

Great. And then secondly, the Maryland site. Seems like a lot of potential up to 1-plus gigawatts but a little bit more complex and kind of playing into what you guys have talked about bring your own power. How does that play into some of TeraWulf strength?

Sure. Just to be clear, Chesapeake data, it's about the power and load differentiation to gig site, certainly a gig data center load capability, 500 megawatts of battery storage capability. It's a former coal generation campus. It's in the Northern Virginia corridor, which means prices are exceedingly competitive. It's designed to be a net contributor to the Maryland grid. It's very well supported by Governor Moore and MTE. Again, both Maryland and Kentucky, by the way, a very sophisticated brownfield programs, which make it really easy for a guy like us who's been around the block and owned and operated coal-fired power plant, shut them down, mitigated them, did everything the right way. These two states have very, very progressive programs on how we do that at commercial feasibility.

Listen, the market is moving to bring your own generation. We said that a year ago. In December, Alphabet bought [ Intersec ] for almost $5 billion. In January, Microsoft raised dedicated generation as part of their 5-point infrastructure strategy. President Trump in January floated the notion that PJM emergency auctions needed to incentivize new generation, new generation projects would go to the top of the queue for interconnect. And then just the other night in the state of the union, the President stated data centers need to build and fund their own generation. So that's where the world is going. We have a real and growing power shortfall. Morgan Stanley says potentially 47 gigawatts, 2025 to '28. The hyperscalers are openly stating that power is the mining constrain.

Look at anything, recent public commentary by Collete Kress, NVIDIA's CFO, Sundar, Alphabet CEO, certainly Jensen Huang, NVIDIA's CEO, I mean it's all about we need power. So delivering generation alongside that load solves the problem. We could bring incremental megawatts to the grid or the system, dispatchable generation using CCGTs, not just bridging power, and we have a history of partnering with grid operators to solve reliability and adequacy challenges. What's our core competency.

We've been talking about this. This is the only team out there. For 25 years, we have been developing, building, renovating, rescuing generation. 6 gigs of power generation experience on this team together on this team and it's been over 25 years. We have deep expertise in siding, interconnection, generation development and -- and that's why we could go and take on a project like Morgantown and have the support that we can from both the local and state communities as we pursue this. We're really excited about Morgantown. It's a big job, but it fits into our schedule. And again, we've told the world, we're 250 to 500 incremental megawatts of data center load every year, and this fits right in.

Our next question comes from Tim Horan with Oppenheimer.

Do you think the pricing in terms and conditions in Kentucky can be materially better than what you've done in the past? And the construction schedule seems pretty ambitious. Do you have the labor and all the equipment you think you'll need in, I guess, the permits to get it done?

So I think I'll go first and maybe Naz, you can go second. In terms of the labor, yes, Kentucky is a fantastic, it's just a fantastic place. You have not only the construction expertise and the trade expertise, but you've got people that really want to work. We brought on for that job floor, which I consider to be a world-class contract party. We've dealt with them in the power space for a long time. They have relationships along in our C-suite for many, many years at every level. We are already ahead of the curve and sort of procurement. We have a proven design that our hyperscaler customers really like. And we think because it's immediate available power.

Well, we don't think we know because we've got this very robust conversation going on right now with folks that want to come in and be our customers there, for competitive pricing. Naz?

Yes. And just to add to what Paul said. As Paul said, we brought in floor on the EPC side. So we're already hitting the ground running with respect to the overall development of the project. And so the time lines that you see for the second half of '27 are reflective of ongoing and meaningful discussions already with the EPC. So that's -- so we feel pretty good about the overall time line. The other big component is gathering the labor that's going to support the project as well. And so there have been efforts already underway for both the mechanical, electrical and civil scopes, to get the requisite labor to support the project as well. So that data is informed by quite a bit of discussion that we've had with floor, who is our EPC as well as the work that we've done on site since the acquisition.

In terms of just the overall economics, I think we've said this on prior calls, we really think about these projects on an un-levered yield basis. And so if you look at where we've been historically, we think we'll be in that ZIP code or maybe better over time. But again, as we think about the overall economics of the project, we're always kind of zeroing in on what's the un-levered yield that we're developing at and pushing to kind of maintain and continue to increase that as well.

Our next question comes from Chris Brendler with Rosenblatt Securities.

And congratulations on the progress here. I wanted to ask on a couple -- on the power side. First, I noticed that the PUE across all these sites is right in line with the initial deals at [ 1.25 ]. And you mentioned in the slides that that's best-in-class. My understanding was there were certain aspects of Lake Mariner and Cayuga that drove that 1.25, but maybe I'm mistaken, it seems like it's standard for TeraWulf to operate that incredible efficiency. Can you just give us a little color there on why you're able to sort of run circles around your competitors when it comes from a PUE standpoint?

Sure. Chris, it's Nazar here. So there's a couple of factors in that. As you noted, one is just the geographic location. And again, as we are more in the northern half of the country versus the southern half of the country there are benefits that we get from ambient conditions perspective with respect to the design and being able to meet that peak PUE. You'll see the Abernathy site is at a 1.4 PUE, again, which is reflective of just the geographic location.

In addition to that, we have invested heavily kind of on the cooling side well. And so we're giving ourselves extra room on the design that we have on cooling as well to maintain that lower PUE. So general, as we think about where we are for generally kind of in the northern half of the country where we have these off seasons during kind of the winter and the spring and where the summer isn't sustained heat, we think we can kind of maintain that 1.25 PUE.

Okay. Was there also a redundancy aspect to it? Are you still not using big diesel generators in these sites?

That's correct. And again, that gets back to the brownfield sites. So typically, you're seeing us play at brownfield industrial sites. And the way to think about it is when the smelter was there, there was a significant investment to design and build a smelter at that site. The last thing that Century wanted was a single point of failure on power coming into the site. No different than a data center, right? So there's 5 different lines coming into that site, which provides a considerable amount of kind of redundancy. So when we look at sites like that, we see that there are multiple independent pathways for the delivery of power. -- obviate the need for on-site kind of backup generation.

So again, Morgantown, similar. It used to be a former coal-fired power plant. That 1.5 gigawatt coal-fired power plant did not want to have a single point of failure in getting power out. We will utilize that same system now to kind of bring power back in. So the big benefit with these industrial -- former industrial brownfield sites is that usually, they have that built in redundancy that data centers want as well.

I just wanted to thank for that question regarding PUE because we put a page in the deck called critical IT capacity, the metric that matters. And what we're going to be trying to do and really sort of ensure that The Street and our retail shareholders understand, gross megawatts measure scale, but it's critical IT megawatts that measure monetized execution. And we think of TeraWulf as an execution story. So we're really going to be reporting more in the context of critical IT as opposed to just gross megawatt capacity. We think it's far more...

Great. And one more quick question is the how big of battery power -- sure caught my eye. Can you just give us a little color on why and what that means?

So as we're adding these large loads and if you see in the statements that we've made, do we want the site to be overall a net supplier of energy back to the grid into the state. So having that battery storage allows the load at the site to operate in a way we're really impacting peak demand. And so it's kind of a critical peak demand shaver, which we think again, provides makes the loads out of there, passes back to the grid rather than burn it. So we think the right composition is about 0.5 megawatt of storage per megawatt of load. And so that's why you see in each of the phases at Morgantown for every megawatt of load, there's about half a megawatt of battery storage.

Our next question comes from Nick Giles with B. Riley.

Guys, I want to congratulate you for the progress across all fronts. Maybe just following up on that last one. Nazar, you mentioned the battery storage, there just appears to be some different moving parts at Chesapeake. So can you just give us a sense for how CapEx could differ from Mariner and what you're doing to really de-risk that development?

Sure. So when we think about -- as we think about the composition of Morgantown, as you noted, there's a few different legs here versus what we're doing at Lake Mariner or other sites. On the data center side, we're in that $8 million to $10 million per megawatt range for CapEx. And so that's what we've done at Lake Mariner, Abernathy, and as we look to contract Kentucky, we're in that same range. In Morgantown, now we're adding two incremental legs. One is the power generation side and the second is the battery storage. On the power gen side, we're in that we're going to be somewhere in the $2 million to $3 million per megawatt range for the fully delivered power plant. Part of that will be time. Part of that will be the type of turbines that we are deploying, and we're working on a number of different alternatives for that site as we speak. But that will be around $2 million to $3 million per megawatt for that capacity.

And there's about another $1 million or so kind of dollars per megawatt on the critical IT load. So kind of the battery storage is usually kind of priced in a megawatt hour basis. But when we translate it back to just kind of the overall load, that's around another $1 million. And so when we think about what we're offering back to our customer, it's the data center that they're already paying for, in other deals that we have, the $8 million to $10 million per megawatt range. In addition to that now, it's the components on the generation side, the $2 million to $3 million per megawatt on the the dispatchable gas-fired generation and then another $1 million or so in the battery storage. So all in around $13 million to $14 million per megawatt on a fully loaded basis.

What that does kind of embedded within that, we will be seeking to kind of charge that full cost back in the capacity payment, back to the underlying customer. So they'll effectively now be long, the value of the generation as well as long the capacity in the data center. With the grid connectivity, we will have now kind of the ability to both bring the power in from the grid as well as generate more than enough power to offset that load kind of going out. And so on a net basis, we think over time, the net cost of power that the tenant will realize will be meaningfully lower than just a grid solution only. So kind of on the ins and outs, again, they'll be kind of paying a capacity payment for all of the CapEx, both on the gen side as well as the data center, but they'll [indiscernible] long kind of the value of that gen, the value of that energy, and that will be offset by whatever they pay for the power coming in. So net-net-net, we think it's a very strong position that gives the tenant a, A, quicker path to scale up in scale that gigawatts, a large amount. And, B, we think over time, their net cost of power will be meaningfully lower than a solution that's relying on the grid only.

Nazar, I really appreciate all that detail. I'm sure others will have follow-ups. I just wanted to squeeze one in on the regulatory side. I think you still need approvals from FERC at Morgantown. Can you just walk us through what the timing looks like there and how we should think about what you ultimately need to get across the finish line for a third-party perspective?

Yes, it's pretty pro forma approval process that we've done countless times. It happens anytime you transfer an existing power facility to another party. We're certainly they tend to look for things like are you a monopoly in the area? Stuff like that, we are not. We consider this to be just pretty routine. We would expect FERC approval within 3 to 6 months.

Our next question is from Stephen [indiscernible] with KBW.

Can you update us on any remaining zoning requirements or state and local approvals needed to move forward with the [indiscernible] data center campus? That's one. And then two, separately, while Cayuga already received zoning approval? Are there any additional permitting or regulatory steps still outstanding for that site?

We'll go backwards. I'll start with Cayuga. And in Cayuga we had our first informal session with the planning Board a couple of nights ago, it went very well. They were engaged. They asked good questions. And so that process is one where we go ahead and finalize their application. That will take another month or so. They then meet on it. They review it. We come to an agreement together on what that permit should look like. They'll be solving for certain conditions like are you drawing water from the lake? What will noise levels be, in the midst of operation? What are you doing for the site in terms of beautification? Things like that. These are all reasonable things. It's what we do. It's what any company in redeveloping a former power plant or industrial complex does. The Cayuga process, as you know, first was had to go through a zoning board of approval to ensure that it was consistent with the definition required to achieve a permit. It did winded down that, and so we're just ordinary course routine working together with the planning board of town to move forward.

With respect to Kentucky Naz?

Yes. So on Kentucky, Stephen, from a permit perspective, we have to obtain the requisite building permits. That being said, we had a town hall this week in Hawesville. We had a workforce session for potential employees. And so the entire Judge Executive, Economic Development Director in Hawesville are all kind of fully onboard and very eager to kind of get us going. As Paul mentioned earlier, it's an extremely supportive environment. And so while we don't have the permits in hand, everyone is fully aware of what we're doing. We briefed them on both the scope, size and scale of the buildings that we're bringing.

Importantly, as a part of what we're doing at Hawesville, we're also decommissioning and taking down the old aluminum smelter. So there's a big cleanup that's happening at the site as well, which the town is very eager to have happened. So we got to do all that, but that's, I think, going to come in time. We're expecting as soon as we can get this lease signed up, we'll be in a position to kind of submit all those things. We have previewed a number of those things, and we're hopeful that it should be a pretty smooth process to kind of get approvals around that.

Just two more things. One is with respect to both of these projects, there's not just one permit you get. There's a principal notion of a permit for a facility. But all along the way, whether it's a specific demolition permit or a building permit for a particular structure, you need to get local permits.

The second thing I just wanted to mention about Cayuga which is kind of interesting, as the state has become really a popular place for people and they look forward to doing more data centers. They are starting to come with this notion of what about power. The beautiful thing about the Cayuga site is, it was a former power site. It can be, again, we have the ability both in terms of the landlords, huge site, which is 400 acres. We have the ability to sort of bring in our own power generation, if that was ever something important to state or local constituents. So Cayuga, I think, really a much better site than just -- any other one in terms of that part of the world because of its ability to have a power plant on it if it's needed.

Next question is from Darren Aftahi with ROTH.

First one, if I may, could you kind of characterize the maybe competitive process at Hawesville, maybe comparing that to Lake Mariner. And are we talking about one entity taken, three, four, or is it going to be multiple entities? And then I've got a follow-up.

I'll start maybe just by giving you -- in the environment that we were contracting for Lake Mariner, I think we were very confident we had a really special site. But I think the hyperscalers were on the growth curve in terms of education. They were on the growth curve in terms of figuring out design when we started Lake Mariner, I think people were not as rock solid in their design notions or in terms of how it would fill out the customers, their customers ledger. So I think we're in a very different environment now.

We're in an environment where the hyperscalers are -- I mean, this is a [indiscernible] I've been in over the course of the last 45 years. I mean you've got Meta is very aggressive. Google's got a great program, Amazon's got a great program. Microsoft come back into the market. They're extremely competitive. You've got some of the Neos that are really seeking larger and larger facilities. So we're in a much more competitive environment where the customers actually have much more definition to what it is they want.

The second thing is when we started out with Lake Mariner, we had already built some buildings a little bit. And so we had to sort of work with our customers for instance, in WULF Den and CB-1 and 2 on sort of figuring out the design elements that they wanted, even in CB-3, whereas 4 and 5, we built these buildings from the ground up with the customer, which made the whole procurement and build process that much more efficient. So I would tell you, we have a much more committed and sophisticated customer right now, who knows the design needs that he or she wants, and we're in a much more competitive environment. So we're pretty excited about -- that was what -- the beautiful thing about Kentucky is its immediate availability has just enabled a very robust dialogue for us in terms of filling out our customers. But Naz, do you want to add something?

Yes. The only thing to add, Darren, is we're targeting one, maybe two customers for the entire site.

That's helpful. And then just as a follow-up. I mean, so you've given -- you've kind of raised your bogey and given this 250 to 500 megawatts range. I guess, what are the one or two deterministic factors that kind of drive that range? And given your background in Power as a team. I mean, is there any reason we could think maybe there's upside to that range? I'm just trying to get an understanding of what that context really means.

I can start here, and then Paul can jump in. The 250 to 500 megawatt is a very kind of calibrated range that we're giving you. It factors in the operational and just deployment capabilities. Again, this is getting hundreds of people on site across multiple trades. It's a procurement process around equipment and ensuring that we're not the last buyer, but we have quite a bit of room between what we need and where the market has availability. And then from a financing perspective, it's what the company can bear from an accretive perspective. So when we talk about 250 to 500 megawatts per year, that's $2.5 billion to $5 billion of incremental capital per year that we are guiding the market that we're going to spend every year for the next few years.

And so when we talk about the 250 to 500 megawatts, really is a calibrated guidance around all of the various facets that are required to properly execute and deliver upon capacity. So that's where we are. That's what we see is available. 1.5 years or so ago, we were guiding at 100 to 150 megawatts as we've been able to now deploy capacity and understand the needs of customers. We've opted that guidance to 250 to 500 megawatts. We remain to be confident that every year for the next few years, we will continue to sign up another 250 to 500 megawatts of critical IT load, with customers, deliver that 12 to 18 months hence, and do that year-over-year and continue to kind of grow shareholder value as we're doing that.

Yes, I would want to add a few things. Number one is we selected floor to be our contractor in Kentucky. Why? Because we think they're the most skilled contractors in the country, they're particularly good on the front end projects so that the execution side goes flawlessly. I think -- as we move forward, the notion of the selection of [ Floor ] was scalable growth so that we could work with them on additional projects on a national level. The second thing is we're all about execution. Again, whether it's on the development side or the execution side, we could talk as much about our pipeline as you want. We could talk as much about all the projects that we're reviewing before we decide to make a move. But at the end of the day, if we don't deliver for our customers, we are out of business, and we have failed our investors.

So we need to be conservative here because we're still on the -- we're still a nascent business. I mean, we've seen changes in the design of our facilities just between CB2A, B, 3, then 4 and 5. And so I think as we're learning and growing in partnership with our customers and now our leading contractor in Kentucky, we will continue to enhance our execution capabilities, and if there is an opportunity to grow beyond what we've said is 250 to 500 megawatts incremental critical IT load, then we will. But for right now, we have to keep our eye on the ball. If the table is full of lots of cookies and cakes and candies, but we have to stay focused on the meat and potatoes and deliver for our customers.

Our next question comes from Brett Knoblauch with Cantor Fitzgerald.

Paul, I guess you have multiple attractive sites that you could see really kind of go and lease out now. I guess I'm talking to prospective customers. Is there a reason why they would want maybe the Kentucky site over like Cayuga or maybe the mega campus that you're building in Maryland? And has in your mind, what maybe site is next up for grabs change, as kind of Kentucky jumped to the top of the list?

Listen, I -- so first off, -- the answer is the demand is so significant. I think that a party would be happy to take everything off the table at any one time. But it's about time to power, and that's why Kentucky has become so important to the customers that we're in discussion with. The ability to have that kind of scale within this short period of time where [indiscernible], if you will, just makes it one of the most exciting sites in the country.

The one thing that we've been pushing, Brett, is we like regional diversity. We think hyperscaler customers are getting focused on that, too. We think -- they've seen now. They've experienced what happens now first in Ohio and then now down in Ercot. When you get everybody all in one place, it's a question of security, but it's also a question of what can the grid really handle. And so I think having regional diversity is something that our customers find compelling. And that's a good reason for them to really focus on Kentucky, that and the immediacy of its power.

Awesome. Then I think in the prepared remarks, you said that Kentucky could potentially expand beyond the 480 megawatt, I guess to what extent have you guys looked into that? And how much do you think it could expand down 480 megawatts.

I think -- listen, we've talked to our power suppliers there. We understand the grid there. I think that -- and again, I was just with the governor a couple of days ago, and he's really terrific. He runs DGA he's very commercial. He's very pro business for the state. I think that we will have opportunities to expand our footprint both in the generation side and on the customer side in Kentucky. But again, I have to stay focused, right? My job is execution. I've got to deliver these facilities to our customers, and we're doing that at Lake Mariner. And Kentucky is going to be a very prompt bit. We've already issued a limited notice to proceed to [ Floor ] just because our customers really want to get on to that property. So we're doing everything we can to facilitate that. Just got to really focus on execute.

Our next question comes from Michael Donovan with Compass Point.

To follow up on what you're talking about Paul, on sizing and design plans. For Maryland and Kentucky phase build-outs, what is the target critical IT megawatt per building or hall you're thinking about today? Should we think about repeatable modules similar to Fluidstack? And what drives that sizing decision?

This is Nazar here. So in the Fluidstack context that 168 megawatts of critical IT load is the base design that we worked closely with them on developing. You've seen that in the number pop up subsequently with other of our peers as well. But in that context, that's kind of the base design. That puts you a little over kind of 200 megawatts of gross capacity. And so in general, as we're having discussions with various customers, we typically look at that 200-megawatt gross 160-megawatt net, as kind of a building block that we're building off of. And so the design that we're working on with [ Floor ] really kind of incorporate that, again, roughly 200 megawatts, plus or minus gross 160 plus or minus net critical IT megawatts as the base building block from which we're kind of deploying that.

And so when we talked about Kentucky, for example, in that 380 megawatts of net, that's a couple of those buildings would kind of consume that capacity.

Our next question is from John Todaro with Needham & Company.

First one is more high-level political regulatory. It sounds like really good conversations with the governors in Kentucky and Maryland. Just wondering more at a broader level, how you are viewing some of the pushback at the state level to data center builds. Just any commentary there, whether sort of the media articles might be overblown or if there is some risk there it?

It all requires, I would say, thoughtful and careful engagement. How you bring your load on is critical to how you're perceived and what impacts you have. And so if you have been listening to us for the last year when we talk about Bitcoin mining, we've said from the beginning, there is a way to do Bitcoin mining that's accretive and an asset back to the grid, and there's a way to do it, where you're not. And so -- we have been very engaging with -- I mean, for example, in Kentucky, we spent a tremendous amount of time with the energy supplier there, big Rivers and have developed a very close and strong working relationship with them, where we are kind of aware of the challenges they have on committing to supporting a large load, and if it's there, it's great. And then when it's gone, it's not so great. That's A. And so kind of ensuring that our interests are aligned with them, and they have clear visibility on where we are.

And then B, just in terms of the overall load profile and when you're consuming power and what happens at those kind of peak demand hours. And so we've been very, again, thoughtful in thinking through how does our load and where we're locating these loads tie back to what's happening in the grid around it. And how do we ensure that, again, that our loads can be overall assets and kind of a beneficial back to the grid and to the local communities versus just kind of coming in and being burdens. And so I think the articles, the news is news and kind of it comes out, how it comes out. But we pride ourselves in trying to be very thoughtful around the issues that we think are pertinent and properly addressing them. And so hopefully, over time, you'll see even in Kentucky, based upon how we're structuring things with the local utility down there, that it hopefully becomes a model for how things should be done. An example of how data centers should be integrated back to grids.

Great. And then second one is just on kind of where we are in the head count growth for Kentucky and Maryland and just, I guess, your [indiscernible] of the G&A guardrails around there.

So we are in the early stages on both. We have kind of ramped up the Kentucky team to half a dozen or so folks over time, including the on-site personnel, where that's going to be over 100 people for Kentucky by itself. Each of our sites, we are targeting somewhere in that 100 to 120 people range for fully loaded staffing, once the site is fully operational. And we're probably a couple of people in Maryland right now.

As was noted earlier, we're pending kind of FERC approval to take over the site. There's existing generation at the site. And so we're ramping that team up to kind of support the operations as well. So I'd say we're in a couple of dozen people between the two sites now, but that's quickly increasing. And again, we should be hitting towards the end of the year, early next year, we should be hitting pretty sizable numbers in Kentucky and we'll kind of be in that 100 people range at Kentucky by this time next year I guess.

We've also, of course, added at the corporate level so that we can manage the much larger portfolio and stay on top of everything from procurement. And so it's legal, it's accounting, it's IT. It's everything that you need at the corporate level to manage projects of this scale.

Our final question is from Martin Toner with ATB Cormark.

When do you think Phase 1 of Morgantown might be able to be turned on?

We're working through that as we speak. So at Morgantown, in addition to the 3 legs we mentioned earlier, just kind of around the load, the data center, the gas gen and the battery storage. There's also a remediation that goes along with that. We were with MDE just this afternoon and kind of scoping that out. And so once we get definition around the timing and scope of that remediation plan, that will then kind of feed into, just the timing.

If you look at what we said kind of in the deck, we've kind of positioned this as kind of end of '28 kind of in '29 and beyond. So generally, that's where we are. But over the next, I'd say, a couple of quarters here, we'll have greater definition to provide around the specific timing on Morgantown.

Coming at it from 10,000 feet. The State of Maryland had some challenges because they've shut down a lot of their great generation. They were -- they weren't really as pro generation as -- or as prescient to what would happen as a lot of other states, which have been very, very supportive, like Pennsylvania next door. So the Governor has really changed policy. We've received written commitments for expedited permitting for our site. So I don't think this is going to be business as usual. I think they're really keen to have us come there, create jobs both at the county level and the state level. The reception has been amazing, but they have literally given us sort of an office to work with, to expedite all sorts of permitting from the repowering to the remediation.

We have reached the end of the question-and-answer session, and this concludes today's conference. You may disconnect your lines at this time, and we thank you for your participation.

Greetings, and welcome to the TeraWulf 2025 Third Quarter Earnings Conference Call. [Operator Instructions] Please note that this conference is being recorded.

I will now turn the conference over to John Larkin, Senior Vice President, Director of Investor Relations. Thank you, Mr. Larkin. You may begin.

Good afternoon, and welcome to TeraWulf's 2025 Third Quarter Earnings Call. Joining me today are Chairman and CEO, Paul Prager; and CFO, Patrick Fleury.

Before we get started, please note that our remarks today may include forward-looking statements. These statements are subject to risks and uncertainties, and actual results may differ materially. During this call, we may use words like anticipate, could, enable, estimate, intend, expect, believe, potential, will, should, project and similar expressions, which indicate forward-looking statements. For a more comprehensive discussion of these and other risks, please refer to our filings with the SEC available at sec.gov and in the Investors section of our website at terawulf.com. We will also reference certain non-GAAP financial measures today. Please refer to our 10-K and 10-Q filings on our website for a full reconciliation of these non-GAAP measures to the most comparable GAAP measures.

We will start today's call with prepared remarks from Paul and Patrick, followed by a Q&A session. Now I'd like to turn the call over to our CEO, Paul Prager.

Good afternoon, and thank you for joining us today. The third quarter was truly transformational for TeraWulf, both operationally and financially. During the quarter, we executed one of the most significant steps in our company's evolution, signing approximately 360 megawatts of critical IT load with Fluidstack backstopped by Google at our Lake Mariner campus in Upstate New York. This 10-year agreement representing average annual revenue of approximately $670 million and average annual net operating income of more than $565 million before extensions, firmly validates our high-performance computing hosting strategy and establishes TeraWulf as a leader in designing, building and operating low-carbon enterprise scale compute infrastructure.

In October, we reinforced that leadership by closing $3.2 billion in senior secured financing backed by the Google credit enhancement to fully fund the Lake Mariner high-power compute build-out. This transaction is a milestone for the broader industry, demonstrating a repeatable end-to-end development model that begins with design and site control, extends through customer contracting and construction and culminates in long-term credit-enhanced lease revenue.

The third quarter also marked an operational inflection point for TeraWulf as we recorded our first HPC revenues with lease commencement at WULF Den and CB-1. We remain on track to deliver CB-2 near year-end, subject, of course, to tenant fit-out requests, which will complete our delivery of 60 megawatts of critical IT for Core42. Across our platform, these early deployments are proof points that our strategy is working and our execution is disciplined.

At Lake Mariner, our team continues to perform exceptionally well. In terms of executing for Fluidstack in Google, the majority of long lead items have been contracted through CB-5 and construction progress is both visible and measurable. CB-3 is more than 50% directed. The final concrete pour is scheduled within two weeks, and the structure will be fully enclosed before year-end. CB-4 and CB-5 are already well underway with underground work beginning next week, field deliveries arriving in early December and building erection expected to begin before Christmas.

The progress our construction and operations teams have achieved and with rigorous quality standards reflects TeraWulf's deep experience in developing and delivering large-scale energy and data infrastructure. We also continue to expand our geographic footprint and customer base. Just two weeks ago, we expanded our partnership with Fluidstack and Google, announcing our joint venture to develop and operate the Abernathy HPC campus in Texas within the Southwest Power Pool market. This project adds 168 megawatts of new HPC capacity with expansion potential up to 600 megawatts and replicates the same credit-enhanced structure proven at Lake Mariner. This joint venture with Fluidstack and Google leverages our collective expertise, incorporates Hypertech as EPC partner and includes two additional options to expand the joint venture, one for future phases at Abernathy and another for a separate site elsewhere in the United States. This partnership represents the next evolution of our growth model, scalable, capital efficient and backed by world-class partners.

And while we've made tremendous progress executing the business we have, what's equally important is how we're positioning TeraWulf for the next wave of growth. Our approach remains disciplined, expanding only where we have clear structural advantages in power, permitting and partnership and our opportunity set continues to broaden.

In August, we signed an 80-year lease at the Cayuga site in New York, laying the groundwork for large-scale high-power compute deployment beginning in 2027. As just mentioned, the Abernathy joint venture offers meaningful embedded expansion potential, both on campus and across future projects with Fluidstack and Google. Meanwhile, our in-house development pipeline continues to mature with several high-quality opportunities now approaching realization. Together, these initiatives form the very foundation for TeraWulf's next phase of growth, executing today while methodically building the platform for tomorrow, scalable, low carbon and designed to meet the accelerating demand for high-performance compute.

Reflecting that confidence, we recently increased our annual target for new HPC signings from 100 to 150 megawatts per year to 250 to 500 megawatts per year. We did not make this decision lightly. It reflects the tangible progress we've made in advancing our development pipeline and the strength of customer demand. Over the past year, we've evaluated over 150 potential sites, narrowing that list to a select group that meets our strict criteria, grid redundancy, minimum power thresholds, attractive geographies for end customers and time to power.

To support this next phase, we've expanded our site acquisition and development teams, strengthening what is already the most capable organization in the sector. Our deep understanding of what hyperscale and AI customers need, combined with our access to scalable, low-cost power, positions TeraWulf at the forefront of the infrastructure transformation now underway. We are proud of what our team accomplished this quarter, but we are even more excited about what lies ahead.

With that, I'll turn the call over to our CFO, Patrick Fleury, to discuss our financial results in more detail.

Thank you, Paul. The 3Q 2025 results reflect a strong contribution from our legacy Bitcoin mining operations and more importantly, the start of HPC Leasing segment revenues. On our 2Q 2025 earnings call, we discussed a series of capital markets initiatives in the second half of 2025. I'm proud to report that with the benefit of our new financial support from Google and help of our partners, including Morgan Stanley and Paul, Weiss, we've executed beyond our expectations, raising over $5.2 billion at incredibly attractive rates, creating durable equity value for our shareholders.

Now let me turn to the results. In the third quarter of 2025, GAAP revenues increased 6% quarter-over-quarter to $50.6 million from $47.6 million in 2Q '25. We recognized $7.2 million of HPC lease revenue at WULF Den and CB-1 with intra-quarter lease commencement resulting in 22.5 megawatts of energized hosting capacity. Continuing with our long-term commitment to financial transparency, we've added a page in our investor presentation detailing lease accounting nuances, which we hope you find helpful. We self-mined 377 Bitcoin at Lake Mariner or approximately four Bitcoin per day, a 22% decrease compared to the 485 Bitcoin mined in 2Q '25.

Our GAAP cost of revenue, exclusive of depreciation, decreased by 22%, $22.1 million in 2Q '25 to $17.1 million in 3Q '25. Power prices in Upstate New York normalized in 2Q '25 and continued to decline in 3Q '25 to $0.047 per kilowatt hour, in line with historical levels and our previous guidance of $0.05 per kilowatt hour for the second half of 2025.

Proceeds from participation in demand response programs, which are recorded as a reduction in cost of revenue during the period in which the underlying program occurs, increased to $7.4 million in 3Q '25 from $3.1 million in 2Q '25.

Operating expenses increased 28% quarter-over-quarter to $4.5 million in 3Q '25 from $3.5 million in 2Q '25. This trend higher throughout 2025 is primarily the result of increased staffing levels at Lake Mariner necessary to support our entry into HPC leasing.

SG&A expense for 3Q '25 was $16.7 million, a 17% increase from $14.3 million in 2Q '25. After adjusting for stock-based compensation, SG&A increased quarter-over-quarter from $10.6 million in 2Q '25 to $12.3 million in 3Q '25.

Depreciation increased quarter-over-quarter from $18.8 million in 2Q '25 to $26.5 million in 3Q '25. The company recorded accelerated depreciation expense of $7.8 million related to a certain minor building and related miners, of which the company shortened its useful life based on expected shutdown of operations for purposes of supporting the HPC operations.

Change in fair value of contingent consideration was $8.8 million in 3Q '25 related to fair value remeasurement of contingent consideration liabilities based on milestones achieved during the quarter related to the acquisition of Beowulf E&D.

Loss on disposals of property, plant and equipment net was $2 million in 3Q '25, down from $3.8 million in 2Q '25. These losses related to the sale of 8,900 and 2,900 miners, which were sold or otherwise disposed of for proceeds of $6.9 million and $1.9 million in 3Q '25 and 2Q '25, respectively.

GAAP interest expense in 3Q '25 was $9.8 million compared to $4.0 million in 2Q '25, and we recognized interest income of $4.1 million in 3Q '25 compared to $1.2 million in 2Q '25. Cash interest paid during 3Q '25 was negligible compared to $7.1 million in 2Q '25 as the 2.75% interest on our $500 million convertible notes is accrued and payable in 2Q and 4Q of each year.

Change in fair value of warrant and derivative liabilities in 3Q '25 was a loss of $424.6 million related to the Google warrants and the conversion feature of the 2031 convertible notes, which was originally accounted for separately as a derivative liability.

Our GAAP net loss in 3Q '25 was $455 million compared to a net loss of $18.4 million in 2Q '25.

Our non-GAAP adjusted EBITDA improved 25% quarter-over-quarter, totaling $18.1 million from $14.5 million in 2Q.

As a reminder, these results are inclusive of significant increases in operating expenses and SG&A over the past 12 months as we invested heavily in our HPC business. These incremental costs have been entirely borne by our legacy mining business until now.

Turning our attention to the balance sheet. As of September 30, we held $712.8 million in cash and restricted cash with total assets amounting to $2.5 billion and total liabilities of $2.2 billion. In October, we closed over $4.2 billion in capital markets transactions, including $3.2 billion of 7.75% BB-rated senior secured notes due 2030 and $1.025 billion of 0% convertible notes due 2032.

As seen on Page 14 of our 3Q '25 investor presentation, with these financings complete and the La Lupa and Akela data center construction projects at Lake Mariner fully funded, our pro forma liquidity totals over $1 billion, which provides cash for three important initiatives: one, TeraWulf's cash equity contribution to the Abernathy JV with Fluidstack; two, the acquisition of key sites in our pipeline that have advanced to the final stages of diligence and negotiation; and three, excess cash to create a fortress balance sheet to weather any storm.

With regard to the Abernathy JV, we anticipate coming to market before year-end with a senior secured notes financing similar in all respects to the offering we recently completed at WULF Compute. As a reminder, the Abernathy JV benefits from $1.3 billion of Google credit support over a 10-year period.

The second half of 2025 has been nothing less than extraordinary for TeraWulf and its stakeholders. We have secured over $16 billion of HPC lease agreements and executed over $5.2 billion of financings at incredibly attractive rates, added significant liquidity to the balance sheet and shown we have a deep multifaceted pipeline to grow the business at 250 to 500 megawatts annually in the future.

With that, I'll turn it back over to the operator, and we look forward to answering your questions.

[Operator Instructions] Our first question comes from Mike Grondahl with Northland Securities.

First question for Paul. Paul, it was noted that there's some key sites that you're close to closing on. Can you talk a little bit about those sites?

No. Good question, Mike. There are at least two sites that we're very, very close to sorting out. We're going for some regional diversity where we think our customers are inclined to enter into long-term agreements.

We've built out our team on the front end here to focus. We've looked at over 150 opportunities. I would not be surprised if by year-end, we announced at least one, possibly two additional sites.

That's great. And then a question for Patrick. Patrick, I noticed in the slides you're breaking out segments now with BTC and HPC. And I also noticed on the HPC side, the margins look like they were about 72%. And I think in the past, you've talked about roughly 85% margins. Can you kind of reconcile that?

Yes. Yes, sure. Thanks, Mike. Yes, I think you'll see when we file the Q, full detail on the segments, obviously, which we're really proud of given the start of HPC leasing revenue. But yes, if you -- the actual margin was about 72%. There -- in the operating expenses, there's about $700,000 of expense development expense at Cayuga. So if you back that out, you'll see that gets you to about an 82% margin for the quarter. So obviously, much closer to that 85%. And then the quarter is a little off just because we didn't have full revenue. It was a stub period for CB-1 in particular. So I think you'll see that normalize here very quickly in the fourth quarter to right around that 85% that we've guided to.

Great. And congrats guys on all you've accomplished the last 90 days or so.

Our next question comes from Nick Giles with B. Riley Securities.

Congrats on all the progress. Obviously, your agreements to date involve top-tier credits. Just was curious to hear how you're thinking about customer diversity from here. Is there a desire to expand the customer base? Appreciate any comments.

Sure. As you know, we have two/other world-class credits as our customers, Core42 backed by G42 and then the Fluidstack Google deal and those that may be associated in that deal, if you've taken a look at the recognition agreements, you're aware of that. And so I could not be happier with the credit quality of our customers, which is the critical element here and something Patrick pounds the table on because, obviously, we want to be able to get ideal credit terms for our transactions and the result in financing.

So the answer is I would expect we'll continue to grow with the customers that we have. And certainly, we're continuing to have dialogue with a couple of others. But again, the key for us is credit quality. And so it's a little bit of a smaller universe as we move forward.

But yes, the sites that we have and the sites that we are anticipating bringing home, all would be very compelling to a great quality credit in addition to, of course, the ones that we already currently have on the books.

Thanks, Paul. That's helpful. My next question was just on the JV. Can you just talk about how that opportunity came about? And then just to clarify, should we consider these type of deals as part of your new updated kind of megawatt per annum guidance? And should we think about that on an attributable basis? Just appreciate any clarity there.

I'll start, maybe Naz or Patrick can take it from there. We have built a great working relationship with the Fluidstack Google team. And if you think about it, there's no sort of reference models for what we're building. These are design build opportunities so that the teams -- listen, you're either going to succeed because you work together or it's going to be a disaster. So we like succeeding. So we work hard with the Fluidstack Google teams. They're on site. We meet constantly. I would tell you that they're part of our team now as we look to how we want to consider financing Abernathy.

So it is in the course of that dialogue that they indicated that they had the site, they had the credit quality and that they were thinking about using an EPC, but that they recognized given our experience in energy infrastructure and financing energy infrastructure that it would be additive to the endeavor. And so everybody decided this made a lot of sense.

Our primary strategy as we move ahead is to continue to do what we've done, which is have great sites and look for the right customer. But certainly, we're delighted to partner again with Fluidstack Google. And like I tell everybody, when Google says, "Hey, we'd like you to come into this project with us, you say, yes. You find a way to say yes, so that it makes sense for them and that it is a rewarding experience as well for your shareholders. That's what we did here in Abernathy.

And I -- in the Abernathy deal, there's room to grow at Abernathy itself. But separately, as you are aware, we have a going-forward relationship on the next project of similar credit quality. So yes, that will be a project that we continue to evolve into, but our primary focus is additional sites and additional high-quality credit customers for those sites.

I think Naz or Patrick, you may want to respond to how do we think about it from a financial perspective?

Yes. Nick, it's Patrick. I don't really have much to add. I think as I said in my prepared remarks, we intend to finance this, be in the market, financing it before year-end, and it's substantially similar to the deal that we just printed at WULF Compute.

Got it. And just to clarify, we should think about this on an attributable basis. If we see another one of these announced and we're trying to pair that against your 250 to 500, it would make sense to look at it that way.

Yes, we own 51% of the JV. So, yes.

Again congrats on the transformational quarter. Keep it up.

Our next question comes from Dillon Heslin with ROTH Capital Partners.

Just a follow-up on sort of how you're talking about your power strategy. The non-JV sites you're looking at, are you procuring those sort of in absolute for marketing? Or do you already have customers in mind that you could basically go to right away?

So we have an active dialogue with our customer base. And this is a very competitive market. So we sort of -- we're constantly discussing with them what their needs are, what it is they're seeking at the same time that we're trying to complete our diligence and negotiation over some additional sites. If you remember, we've just got one quarter of a century in energy infrastructure development and operation. It's been very helpful in identifying sites that we think will be in the next wave of data center development for these hyperscalers and high-quality credits. I think in the case of Abernathy in the Fluidstack Google relationship, I think as we look at sites with them, we're obligated, obviously, to work with the customer.

In the case of TeraWulf identifying and determining that another site makes a lot of sense for us. We're going to want to make sure that it's a competitive process that we end up with a great quality credit, but that we also get rewarded for the unique and wonderful qualities of that site and that we get the most return we can for the shareholders. So it's a little bit of both.

Great. And as a follow-up, how are you guys seeing the market in terms of build costs of -- like the sites you've been building so far have been at existing sites where you've had redundant power and then the Fluidstack Abernathy site, you're not building a substation, so the CapEx per megawatt is a little bit cheaper. And then you're talking about you've got long lead items procured for -- through CB-5. But how do you see sort of the market in general maybe beyond Lake Mariner?

Naz?

Dillon, this is Nazar. Generally, what we try to do is be able to deliver capacity or the bulk of the capacity within 12 months of signing the customer. And so that often requires engaging with folks kind of on the electrical gear and the coolers and chillers in advance of that. And we've developed relationships with a number of different vendors where we have a rolling 12-month projected schedule with them, which gives us positions in queues. And then we can -- depending upon which -- how many megawatts of capacity we sign up, then we take those down.

So we've tried to be ahead in being able to procure the equipment, which allows certainty to our customer when we engage with them on discussions. So that's been going well. And given the capacity that we've signed up and procured, we've been good partners for the vendors and vendors have been good partners for us as well on that equipment procurement side. So that's generally how we're approaching things. And as we look forward, I think the forecast and guidance we provided for that $250 million to $500 million, that general procurement strategy covers that capacity as well.

Our next question comes from Chris Brendler with Rosenblatt.

Thanks and also congratulations, amazing amount of progress and looking forward to more. I wanted to ask on the -- actually a Bitcoin question. I noticed the operating hash rate was sort of 70%, I guess, of the nameplate hash rate this quarter and expect it to go down even further next quarter. Can you just give us a little color on what's going on there?

Yes. Chris, it's Patrick. So we're running our sites for our HPC clients now. And as we get closer to bringing all of Core42's capacity online, there are some things electrically that we have to do at the site. So you noticed in my remarks, I talked about accelerated depreciation on one of our minor buildings of $7.8 million this quarter and sales of miners.

So as we sort of reposition the site, particularly for two lines of power and redundancy for the HPC buildings, we're making some changes on the Bitcoin mining side. And that's reflected in those numbers. It's a combination of kind of calling our fleet to be more efficient. And then again, just operating the site really for HPC. So you'll see -- I mean, obviously, site is still very profitable. We had a great quarter from a Bitcoin mining perspective. But I think going forward, that's our approach. Hence, the sales of miners, calling the fleet and then the accelerated depreciation on the minor building.

Makes sense. And I guess from that comment, it sounds like that continues in 2026.

Yes, I think so. I think like we said we kind of gave you some guidance in the deck for what we thought we'd have here in the fourth quarter. And then, look, I think beyond that depends on market conditions.

But yes, I think our intent is to keep mining Bitcoin through certainly the end of 2026 and then dependent upon when the next 250 megawatts that we've requested from the New York ISO comes, I think we could -- you could see us operate beyond there until the next halving. But again, I think that will be a function of additional megawatts at the site as well as just Bitcoin profitability itself.

Excellent. Great. Speaking of additional megawatts, I like the slide that broke out how you think about the pipeline on the power side on Slide 8 and some great details there and helpful to think about where that comes together. The gigawatt plus of development pipeline, is that in the Phase 4 or the Phase 2? I wasn't quite clear where that comment on the following page fits relative to Slide 8 on the phases.

I think, Chris, that's really just going to Paul's comments, which, again, when you think about our funnel, which is really what Page 8 is meant to kind of show you, just the amount of time and effort that goes into cultivating that pipeline. And again, I think unlike some of our peers, we're not telling you a fictitious pipeline of thousands of megawatts all in the same region. We're telling you about stuff when it's literally imminent and ready to go.

So I think as Paul mentioned, we're getting very close on a handful of sites that hit on Page 9, the development pipeline of a gigawatt.

Our next question comes from Stephen Glagola with Jones.

On the raised AI capacity growth targets to 250 to 500 megawatts net annually, are there any structural or operational constraints, like whether, I guess, like EPC capacity, financing availability or like internal bandwidth that could just limit the number of projects you can execute simultaneously?

This is Paul. I'll start. The answer is yes. I mean, I think we've always tried to emphasize here a focus on our ability to execute. And I think that we're more than capable of meeting that goal of 250 to 500 megs. But it takes a lot of time to really get to the bottom of these sites. it's a very competitive market. You need to sort of look near and far afield. You have to make determinations on site suitability for the customer, but also what's power like at the node, what are the environmental and regulatory considerations. It's just a lot. So we're very confident we could do what we now say, which is an increase from where we're at.

I'm not terribly worried about the EPC side. I feel pretty good about that and procurement capability and supply lines aren't what they were. I feel very good about that. I think that the key is going to be our ability to meet schedule and price. That's what the Street is looking for. That's what our customer wants. That's what we promised to our shareholders. So I'm very comfortable at 250 to 500. And as we grow, listen, we're building, as Patrick used to say, serial model # 6. As we get down to 10 or 11 and we find more efficacious ways to do this and needer ways to scale, then we could grow from there. But I think 250 to 500 is the right way to think about us for the coming year.

And if I can just ask one more. Are you seeing any meaningful demand from tenants for the AI capacity with ready for service dates beyond 12 months? And how is that shaping your pipeline site acquisition priorities?

Yes. Yes. Demand is real, and it's a constant. And I think that -- listen, I think there was a site out in Ohio the other day. They got a letter from AEP saying they were in the queue and they were in the queue for '26. And now you should probably not think about that power in '26, but you should think about it for like '29 and '30. And that is a way of saying that you've got to pick your sites really carefully. You have to understand what the grid is capable of. Are you in an area where the whole grid is only X and the demand is 3x that.

So it goes to the notion that you've got to have a very good handle where you cite these things. But that then -- when you go back to the customer and you say, hey, how do you want to think about it if you want to be in this region, you're okay moving from '26 to '27. The answer has been yes, universally. The answer from '27 to '28 is yes. I don't think you get the power problem solved by then. You've got hyperscalers now looking at island generation, which means they're going to bring their own power to the table, and that's at least four to five years away.

If you look at the deal that was signed with NextEra for bringing back the nuc, they looked at a 30-year transaction, which doesn't come online until '29, and I don't think there's any way that nuc is back online in '29. So the demand, I think, is just increasing. And it took a little while to get here. I think everyone was waiting to see who was going to make that first move. But now that we hear, the demand from the hyperscalers and the cloud companies is very, very significant.

Our next question comes from Justin Pan with Clear Street.

This is Justin on for Brian. Obviously, the increase in incremental HPC guidance underscores a lot of optimism in demand over the next two to three years. You guys have had a great couple of months. Some of your peers have had a great couple of months. But could you dig in a little bit deeper into what gives you the confidence in the heightened demand outlook over the medium term? It seems like a pretty interesting dichotomy in the market at the moment, right? We're seeing some talk over AI valuation frothiness. But at the same time, there's been a ton of material success momentum in your space. So...

I'm happy to start, maybe you could follow up. But as I just said, we're looking -- we've been asked by customers to look at opportunities where we have to bring our own generation to the table. And you have to assume that if you're bringing your own generation to the table, you're at least four years out. And so when you have those kinds of questions coming in from world-class credits, that just speaks to just massive amounts of demand.

I think the shortfall in energy is real. It's greater than I think originally forecast and the demand for energy by users like these high load data centers is more significant than was originally forecast. So we we've been getting calls since -- certainly since May, and they haven't abated. Every time we come up with a site, we have at least 5 phone calls that we could go to ask, would this kind of be something that is of interest to you.

So I think if you just look at the hyperscalers and two or three of the big leading cloud players, they're being very, very aggressive in how they're looking at further locations and sites. And again, a lot of these sites wouldn't have any availability for electrification for two, three years out. So I don't know what to say other than to tell you the phones ringing, they show massive demand. And -- we don't see that abating anytime in the near future.

Our next question comes from John Todaro with Needham & Co.

First one here, as we kind of move from a phase of signing some of these leases to executing on them, can you just give us kind of any color as it relates to penalties if you missed time line of delivery and then just kind of frame up your confidence in hitting delivery dates? And then I have a follow-up question.

Yes, this is Patrick. Sorry, Nazar, go ahead?

No, go ahead, Patrick.

Yes. So, John, just with regard to penalties, I'm not going to tell you the specific ones because that's confidential to the lease. But I would tell you, given the accelerated build time lines and all the work we've done with our customers here, there are pretty significant grace periods. So the leases cannot be terminated until we're over 180 days late. Obviously, as Paul said, we're meeting weekly, daily, monthly, like both in person and via teleconference with our customers. So there's no surprises. So folks are well aware of budgets, time lines, et cetera.

In general, we have very minimal penalties for the first 30 to 90 days. Generally, there's a penalty for the 30 days, then there's another one for 60, another one for 90, and then the penalties start to accelerate from day 90 to 180. But those are relatively de minimis for us through 90 days and then scaled from 90 to 180.

Great. That's super helpful. And then second question, if we do just take a step back, I guess, how are you guys able to add more of the power pipeline? Like some of the stuff was procured pretty quickly like Abernathy. I would just have to think major hyperscalers, Neo cloud, maybe private equity, everyone is competing now. Just, I guess, give us -- frame it up a little bit more for how you guys are able to win that.

Yes. I'm not sure I understand the question. I mean -- Abernathy didn't -- I wouldn't look at that as came on real quickly. I would -- again, we've had a long-term relationship now with Google and Fluidstack. And so we are aware of the strategy here, and they decided that bringing us alongside would be additive to the overall effort. But I'm not sure I understand the balance of your question.

I guess just the main crux of it is if we take a step back and there's such a power constrained environment, one of the biggest questions we get from investors is just how these guys are able to continue to procure capacity like that 250 to 500 megawatts you talked about when we are in still a constrained environment, and there's just likely so many bidders for these assets.

Yes. I think the answer is -- so some of them are looking at island generation where they bring their own power. Some of them are looking at high electrification sites that had former industrial uses and they're looking at repositioning them into data centers. And some of them are talking to utilities about figuring out if there's a way that they could work out a deal like the NextEra transaction.

I think they're following multiple strategies to get to the answer of they have long-term demand, and it's near term in terms of its immediate urgency, but they're looking at the 25- and 30-year deals. If you take a look at the Abernathy deal, it's 25 years.

So I'm -- I can't tell you or opine to what the long-term answer is other than United States needs to build more generation. But I think everyone's figured that one out. The question is, are there sites that one can discover in the right regulatory frame set and from an environmental perspective, not too injurious to a customer that could enable a high-quality credit to come along and be a customer. And I think the answer is yes, but you got to know where to look.

I guess I should emphasize TeraWulf where to look, which is why I think prior to year-end, we'll be bringing on at least one, maybe two other sites.

Our next question comes from Tim Horan with Oppenheimer.

Was there a specific trigger that caused you to kind of basically guide to more than doubling your incremental capacity per year or your customers? I mean it seems like demand is much stronger than maybe you were thinking about we were six, nine months ago. Yes, anything that really drove that?

Is that you Naz?

Maybe I can start, Paul, you can jump in. Yes, I think there's a few things, right? So one is if you go back a year ago, we hadn't gone through the full cycle, right? As we have laid out in the deck, there's not just the site, there's not the design, there's the engineering, there's a financing and then there's a construction. So we've been through that full cycle now. And so sitting 18 months ago, looking forward, there was pieces of that process that we had not completed. Pending this quarter, through this quarter, we've completed all of those cycles. And so now we have a much better nuanced understanding of what's required for each one of those phases, what we can get done. There was a question earlier on capacity.

So all of that is factored into that 250 to 500-megawatt forecast, which, again, 1.5 years ago, we didn't have that visibility. So I think that is a reflection both of our capacity capability to get each of those phases done as well as the size of the demand that's coming from the customers.

Yes. I would just want to add two more things. One is, obviously, after we had 42 online, that enabled us to sort of do show and tell to other customers. So I think the level of credible incomings to us just -- it was multiples of what it was prior to that. Secondly, Patrick, who was the architect of our financing strategy from day one, we didn't want to sort of have an arrangement where our customer was also financing our CapEx. He wanted to go about it in the way which we have, which was to say we would do it, and we would require credit support to enable good financing. In our case, Patrick has been able to get absolutely great financing.

So once we went through that, that also opened up the doors to our ability to sort of get the capital we needed to build these things out and also showed customers, hey, this is how we do this. And Patrick led the way. I mean everyone's followed suit since then, but we were the first through that door and it was on the back of Patrick's original vision for that.

So, I think, both what Nazar and Patrick sort of were prescient in thinking about once we were able to execute on both those visions, I think that just led to increased demand that we -- you're correct, we hadn't quite anticipated.

And just two quick questions. Abernathy, do you have a sense how much equity you're going to have to put up to finish that project? And are we talking like $8 million per megawatt for the build-out? And then Lake Mariner, can you talk about what items or item is on the critical path on the construction schedule, please?

Yes, I'll answer the first question. So, on Abernathy, we've given you guidance of $8 million to $10 million per critical megawatt. If you do that quick math and again, take the wide end of the range, it's roughly $1.7 billion, and there's a $1.3 billion Google backstop.

So, again, you can kind of do that quick math. And I would, again, just say stay tuned. We're kind of sorting out the details with Morgan Stanley and our partners right now, and we'll be in the market as soon as we can be.

And then the construction item critical path.

Nazar, do you want to take that?

Sure. On the second question on critical path, there are -- from an equipment perspective, we are on schedule or ahead of schedule for all of the long lead time pieces. So those -- mostly on site, CB-4 and CB-5 are on schedule for construction. We are currently ramping up labor at the site. We're going to peak probably sometime in the month of March. And so ramping that labor up is probably the near term, the biggest item that we're focused on.

And so that's, again, we've got a number of different things ongoing to accomplish that, but that's the big driver, I think, over the next couple of months is getting that ramp up as CB-4 and CB-5 get through civil and get into kind of full swing on mechanical and electrical.

We have reached the end of our question-and-answer session. I would now like to turn the floor back over to Paul Prager for closing comments.

Listen, everybody. I really appreciate you joining us today. I think if there's one takeaway from the quarter is that TeraWulf is executing. We're methodical, we're consistent, and we're doing this at scale. We are building a differentiated platform at the very intersection of AI, power and infrastructure, supported by long-term contracts, strong partners and a proven ability to deliver.

I'm convinced we have the right strategy, the right team and the right assets to continue this momentum well into '26 and beyond. My focus and our focus remains disciplined execution, thoughtful expansion and creating sustainable long-term value for both our shareholders and partners. I want to thank you for your continued support and confidence in TeraWulf. Thank you again.

This concludes today's call. You may now disconnect your lines.

Greetings. Welcome to the TeraWulf 2025 Second Quarter Earnings Conference Call. [Operator Instructions] Please note this conference is being recorded.

I will now turn the conference over to John Larkin, SVP, Director of Investor Relations. Thank you. You may begin.

Good morning, and welcome to TeraWulf's 2025 second quarter earnings call. Joining me today are Chairman and CEO, Paul Prager; and CFO, Patrick Fleury. And for Q&A, we will be joined by Co-Founder and CTO, Nazar Khan; TSO, Kerri Langlais; and COO, Sean Farrell.

Before we get started, please note that our remarks today may include forward-looking statements. These statements are subject to risks and uncertainties and actual results may differ materially. During this call, we may use words like anticipate, could, enable, estimate, intend, expect, believe, potential, will, should, project and similar expressions, which indicate forward-looking statements. For a more comprehensive discussion of these and other risks, please refer to our filings with the SEC available on sec.gov and in the Investors section of our website at terawulf.com.

We will also reference certain non-GAAP financial measures today. Please refer to our 10-K and 10-Q filings and our website for a full reconciliation of these non-GAAP measures to the most comparable GAAP measures. We will start today's call with prepared remarks from Paul and Patrick, followed by Q&A with the full management team.

I will now turn the call over to our CEO, Paul Prager.

Good morning, and thank you for joining us. We moved this call from Friday to this morning so we could share a complete update. Over the past few days, we finalized transformative agreements that meaningfully advance TeraWulf's strategy and reinforce our position as a leader in next-generation digital infrastructure.

This morning, we announced 2 major transactions. First, we have a new tenant at Lake Mariner. We signed a 10-year 200-plus megawatt hyperscale AI hosting agreement with Fluidstack, a premier AI cloud platform that builds and operates GPU clusters to some of the world's most innovative companies. This agreement represents approximately $3.7 billion in contracted revenue, with the potential to exceed $8.7 billion if lease extensions are exercised. Fluidstack will utilize more than 200 megawatts of critical IT load, about 250 megawatts of gross site capacity. We also granted Fluidstack a 30-day exclusivity on CB-5, which would add another 160 megawatts of critical IT load, on similar terms, including Google's participation.

Deployment will be phased, with the first 40 megawatts expected online in the first half of 2026, and the full deployment by year-end. The lease is expected to bring in over $350 million a year in revenue, with site level net operating margins of roughly 85%. Importantly, Google is providing a $1.8 billion backstop for Fluidstack's lease obligations in exchange for warrants representing about 8% of TeraWulf's equity, an extraordinary vote of confidence from one of the most influential players in AI.

Second, we brought in Cayuga. We executed an 80-year ground lease with a purchase option, securing exclusive rights to develop up to 400 megawatts of digital infrastructure on a fully equipped site, with high-capacity transmission, industrial water intake and redundant fiber. We expect to bring more than 130 megawatts online in 2027, with substantial expansion potential beyond that.

Together, these transactions increase our total platform capacity to over 1-gigawatt, firmly positioning Lake Mariner and Cayuga as cornerstone assets for the future of AI infrastructure.

Our first HPC tenant, Core42, continues to be an outstanding partner, and we anticipate growing that relationship. The WULF Den is fully operational and generating revenue. CB-1 begins generating revenue within the next week, and CB-2 is on track for Q4. We are hitting our milestones on time and on budget. For months, I've highlighted 3 key priorities: execute for Core42; secure our next tenant; and expand capacity. These announcements deliver on all 3.

Looking ahead, our focus is on financing the HPC build-out efficiently and in a shareholder-friendly manner. With this new customer and the $1.8 billion Google backstop, our credit profile is significantly enhanced, enabling us to pursue low-cost, scalable capital solutions that align with our growth trajectory. Therefore, my immediate focus is execution, execution, execution.

Finally, I want to thank my team, our partners at Fluidstack and Google, and our advisers, Morgan Stanley, Paul, Weiss, and Noah Hansford at Stutzman Bromberg, for their exceptional work in making these milestones possible.

With that, I'll turn it over to Patrick for a quick look at our second quarter results.

Thank you, Paul. I'll briefly cover the financial highlights for the second quarter before diving into the transactions in partnership with Fluidstack and Google announced this morning, and our objectives for the second half of 2025.

In the second quarter of 2025, we self-mined 485 bitcoin at Lake Mariner or approximately 5 bitcoin per day, a 30% increase over the 372-bitcoin mined in 1Q '25. Our GAAP revenues were up 38% quarter-over-quarter at $47.6 million in 2Q '25 from $34.4 million in 1Q '25. Meanwhile, our GAAP cost of revenue exclusive of depreciation decreased by 10% from $24.5 million in 1Q '25 to $22.1 million in 2Q '25.

Power prices in Upstate New York normalized in 2Q, and we expect pricing to remain in line with historical levels for the rest of 2025, guiding at $0.05 per kilowatt hour for second half of the year. SG&A expense for 2Q '25 was $14.3 million. After adjusting for stock-based compensation, SG&A decreased quarter-over-quarter from $11.5 million in 1Q '25 to $10.7 million in 2Q '25.

I'm also pleased to report our non-GAAP adjusted EBITDA showed significant improvement in Q2, totaling $14.5 million, up from a negative $4.7 million in 1Q. As a reminder, these results are inclusive of significant increases in SG&A and operating expenses over the past 12 months as we've invested heavily in our HPC business. These incremental costs have been entirely borne by our mining business until now.

As Paul mentioned, we're on track for the WULF Den and CB-1 leases with Core42 to start generating revenue in Q3. We remain on schedule and on budget for the delivery of this capacity.

Looking ahead to the second half of 2025, we've updated our guidance in our investor presentation. At current bitcoin prices and network cash rate, we expect our mining operations to contribute positively to EBITDA in the second half of the year. We've also slightly adjusted our annual SG&A guidance to $50 million to $55 million from $40 million to $45 million, reflecting the accelerated growth in our HPC business.

Now moving to the transactions announced today. These are truly a game changer for TeraWulf, and I want to highlight some of the financial implications. As Paul noted, the Fluidstack lease and Google support agreement are carefully structured to enhance our credit profile and position us to scale quickly.

Google's partnership and support is multifaceted. First, Google is backing Fluidstack's lease obligations, which include early termination protections for the first 6 years. Second, Google is providing $1.8 billion of credit support over a 10-year period. Third, Google is pledging its equity stake in TeraWulf to support the construction phase of our data centers.

Given the expected improvement in our credit profile, we've refined our financing strategy to focus on a series of capital markets initiatives in the second half of 2025, with the benefit of our new financial support from Google and our updated lease agreements.

Additionally, the long-term ground lease at the Cayuga site adds significant future capacity and upside value for TeraWulf shareholders. We plan to develop up to 400 megawatts of HPC hosting at Cayuga, a site with many of the same advantages as Lake Mariner, including low cost, predominantly 0 carbon power and strong existing infrastructure. This acquisition and the structuring of these transactions were designed with careful consideration of shareholder alignment. As Paul mentioned, we have structured the terms with long-term shareholder value in mind.

The acquisition of Beowulf Electricity and Data in 2Q further streamlines our structure and consolidates expertise in power generation. This acquisition not only simplifies our corporate structure, but also strengthens our ability to execute on future projects. As part of the acquisition, 94 employees from Beowulf, including key personnel from Lake Mariner and corporate support functions have transitioned to TeraWulf.

Finally, regarding our growth pipeline. We are constantly evaluating additional sites to add to the TeraWulf portfolio and maintain an extremely rigorous approach to this process. In 2025, we have evaluated over 75 potential expansion sites, and from that, we have a handful of progressing through negotiations. Given our HPC customer base and electrical infrastructure experience, we will maintain discipline as we evaluate future expansions.

With that, I'll turn it back over to the operator, and we look forward to answering your questions.

[Operator Instructions] Our first questions come from the line of Mike Grondahl with Northland Securities.

Congratulations. A couple of questions. If you could maybe talk about why Fluidstack, kind of the pros and cons there, and the demand for that power? And then how are you thinking about the 30-day exclusivity?

Naz, you want to get that.

Mike, it's Nazar here. Why Fluidstack? Fluidstack, as Paul had mentioned, delivers compute clusters to some of the largest companies in the world, and they've been active with a number of different customers, both domestically in the United States, as well as in Europe. And so they made a couple of announcements in Europe earlier in the year. So they are very active and are in deep and detailed discussions with a number of different counterparties in delivering optimized compute.

And so as we thought about our trajectory and our ability to grow, adding somebody like Fluidstack to Core42, who also has grand ambitions in building a large platform around compute made a ton of sense. And I think as we've mentioned on prior calls, having more than 1 customer at the site has been a goal of ours. And this accomplishes that.

On the exclusivity, today, we signed up a deal for CB-3, which is 42 megawatts of net critical IT capacity, and CB-4, which is 162 megawatts of net critical IT capacity. The discussions that we have ongoing are for a copy of CB-4, which would be CB-5. And given the discussions that we've had, given the intense amount of effort that's been put into customizing this design for Fluidstack, we have a relatively tight window here, so we are hopeful. But again, there's work to do there. But again, that tight window should give an indication of how much work is left.

Yes, that's really tight.

And then Mike, I'll just add to that. This is Kerri, Mike, is that -- the terms on the additional building, CB-5, are on the exact same basis of the terms that we just signed. So the same guarantee from Google and the same economics between TeraWulf and Fluidstack.

Great. And then maybe just one more. The $1.8 billion backstop from Google, how was that number sort of decided on? Is that basically a 6-year protection?

Yes. Mike, it's Patrick Fleury. So yes, the backstop amount is roughly 50% of the payments over the lease term. So yes, it's approximately 6 years. And that support, Mike, is actually in place for up to a -- the 10-year term of the lease, and it declines from the beginning over time.

Congratulations, guys. Great progress.

Our next questions come from the line of Brett Knoblauch with Cantor Fitzgerald.

Congrats on both of the announcements today. Really excited about those. Maybe first, on Lake Mariner, the 200-some-odd megawatts we have allocated towards bitcoin mining, does the Core42 plus the new deal M&A kind of impact the electrical capacity for bitcoin mining? Would you envision shutting bitcoin mining down for Fluidstack? Or is that something you don't really have to do quite yet?

Brett, it's Kerri. So as you know, right now, we've got about 200 megawatts of near-term power available at Lake Mariner. And if you look out over the next 18 months, we're optimistic that we can bring another 250 online there. So I think that -- your question here leads us to the other transaction that we announced today, which is the Cayuga deal. That's why moving on that so quickly was very important to us, to keep the momentum in our customer conversations and have that capacity available across multiple sites. So as you think about it, once we sign a deal, we're typically delivering the facilities within a year. So we're already thinking ahead to meet that time line.

Okay. That's very helpful. And then on the Core42 build-out, can you maybe just remind us of how much CapEx is remaining for that first deal?

Yes. Brett, it's Patrick. I think the precise -- I might be off by 1 million here. But I think through the second quarter, we had spent approximately $200 million. That -- I think if you looked at our presentations in quarters prior, that total spend was about $430 million. And so I think you might say, well, that doesn't seem right, but a lot of that spend is very back-ended. So for example, things like UPS, which is a very large number there, is really back-ended spend that we'll be spending here in the sort of late third quarter and fourth quarter.

That's very helpful. And then maybe if I could just add one more. The new CapEx for Fluidstack, $8 million to $10 million range. It's a bit higher than Core42 range. Can you maybe just talk about the difference in the build-out and cost for that versus what you did with Core42?

Sure. This is Nazar. There's a couple of factors here. One is scale and the time associated with that scale. So for the Core42 discussions, we're delivering 60 megawatts of total critical IT capacity. And with this announcement here, we're delivering over 200. And so -- and the time lines that we're doing so are relatively similar. So there's a need for significantly more labor, which is requiring us to pull from further and further away, which has an impact on our overall cost. And so that's one component to it.

Second is, as we mentioned earlier, we spent quite a bit of time with the Fluidstack team and coming up with a design that really was customized to what they were looking to deploy. And so within those design considerations were tweaks off of what we've built in CB-1 and CB-2. So when you put those things together, as you noted, we're a little bit higher than what our CapEx was going into Core42.

I really appreciate it, guys. Congrats on the deals.

Our next questions come from the line of Darren Aftahi with ROTH Capital Partners.

It's Dillon Heslin on for Darren. The first one, it seems like the yield or the unlevered yield to cost on the build is roughly in the same range of Core42. When you start to look at other sites within Lake Mariner and then Cayuga, and people you're at the negotiating table with, like how much room do you think there is to either expand or maintain that yield to cost?

Yes. It's Patrick Fleury. I'll answer that question. I mean, look, I think we've been pretty transparent about what we're targeting in the past. And I think we will maintain that discipline going forward. And we think -- Paul can talk to market demand, but I would say it's very, very strong.

And so I think you'll -- as Kerri mentioned, and Nazar, and on the expansion capacity that we're talking about, it's on the same terms. So I think we will -- we feel very strongly that we'll be able to maintain that discipline and keep achieving positive economic results for the shareholders.

Yes. And I would just add to that, that since early May, demand has felt almost urgent. We're not chasing yield on cost deals. We just think we can get better returns. Seeing very strong enterprise and hyperscale demand in several formats with more enterprise direct talks, especially in financial services.

On pricing, we're very happy with the economics from Core42 and the Fluidstack Google deal, and think shareholders will be rewarded if we just keep replicating them. I think there's a good argument that the market might even be tighter in 2026 than in 2025 given ongoing power constraints and rising hyperscaler CapEx.

Got it. And as a follow-up, is there a way to earmark potentially how much capital you might need from the capital markets?

This is Patrick. So look, we're really excited about the structure of this deal and the partnership here. So I think given the expected improvement in our credit profile, we've really refined our financing strategy with our advisers at Morgan Stanley to focus on a series of capital markets initiatives in the second half of this year. And those initiatives are going to really benefit from the new financial support from our partner at Google and the updated lease agreements.

So we think that strategy will really afford us a lower cost of funds going forward and increased flexibility going forward, and are just excited because we feel like we've cracked the code. As you know, one of the hardest things about these deals is financing them. And our strategic alignment with Google supporting us both on the debt and as one of the largest shareholders in our equity going forward, we think, is incredibly novel, and we're really excited about continuing to grow that partnership as we move forward.

Yes. And just to put a finer point on it or to underscore what Patrick said. I mean this is going to lower the cost of financing that we do, do, and it enables us to really approach things from a very shareholder-friendly perspective.

Our next questions come from the line of Nick Giles with B. Riley Securities.

Guy, congratulations here. It's really great to see. I think the $3.7 billion implies average revenue per megawatt of around 1.9. At what level does the contract start? And what's the annual escalator?

Nick, it's Patrick. So those details are confidential. I think we gave you enough information that -- you're a real smart guy, I think you can probably back into it. But the data that we gave you is the data. And I think as we move forward and get bigger as a company, we've been incredibly transparent in the past. And I think this is a really competitive space both for us and for our customers. So I think you will continue to see transparency from us. But the data we gave you today is the data that we're going to give you.

Okay. Fair enough. I appreciate that. And then just back to the project financing. I mean, should we expect to see something announced for the Core42 and Fluidstack in kind of one single announcement. Or are those separate conversations? Just appreciate any color on kind of what we should be looking for.

Yes. Nick, like I said, I think it's going to be a series of transactions. Obviously, this transaction fundamentally changes the credit profile of our company in a very positive way. So I think you will see a series of transactions that we're working on with our advisers at Morgan Stanley here in the second half.

Got it. One more for me. Just on the 30-day exclusivity. Do you feel that Google would presumably backstop this capacity as well if executed?

Yes.

Yes.

Okay. Guys, congrats again. Keep up the good work.

Our next questions come from the line of Brian Dobson with Clear Street.

Just on the Google partnership, do you think you could maybe opine a little bit on how you think that that's going to change discussions with future clients? Do you see this as a net long-term strategic positive? And if so, what kind of synergies do you think this agreement would yield, in terms of generating new business further down the road?

Naz, do you want to start, and I'll follow up?

Sure. So I think the hyperscalers and Google, in particular, have put out pretty significant complex forecasts both kind of balance of the year as well as looking into next year. And if you look at the structure of this arrangement, it's, I think as Patrick mentioned earlier, unique, novel. And so what we have found is the demand for capacity from hyperscaler exists, as Paul had mentioned, and how they take that capacity down will take a number of different forms. And so this is an example of that.

And with -- as we think about it, there's the demand of who's going to use it. And then there is the: how do you pay for it and the financing piece of it? So this construct that we developed with our partners at Fluidstack and Google really addressed both of those things. And so I think the long and short of it is that there's demand out there and the contracting for that demand will take a number of different structures, including -- what we announced this morning.

Yes. And do you think that new potential clients will see this as an endorsement from a marquee player in the space?

I hope so. I would expect -- I mean, Fluidstack and Google probably looked at a ton of sites. And I think they were -- they found it compelling in terms of the attractiveness of the site, our installed energy infrastructure, redundant grid connections, the land, the water, fiber latency, 89% 0 carbon power source, availability power.

Second, they were able to see the progress we've made on the Core42 buildings, the partnership that we've developed there, the way we work with our customer. I think these high-power density, all liquid cooled data centers are pretty early in terms of industry development and Fluidstack, Google saw the value in partnering with a team that's already navigating that complexity.

And finally, our team. I mean, I would encourage folks to work out when they can get up to the site, but we've got an amazing team up there led by Sean Farrell, Chief Operating Officer. Our management team, together with Sean, we've been together over 15 years. We've got decades of experience executing complex energy infrastructure projects. And I don't think that's something you can replicate overnight. So I think it was all of that. And I think new customers or potential new customers will look at our appreciation for electricity and electrical infrastructure and their needs and realize that it's a good match.

And just to underscore what Paul said, we had very detailed discussions with our customer around the design and configuration of what we're building for them. So we had a base design. And it's through that, that they got to see [ our ] chops and see what we do, how we work. And so I do think to your question, it should absolutely be an endorsement of our team and our ability to really work with some of the largest companies in the world in building the next generation of data centers.

Yes. Congratulations. It's a big win.

Thank you.

Our next questions come from the line of Chris Brendler with Rosenblatt Securities.

I have to add my congratulations as well. This is fantastic. I just wanted to see if we could potentially get some color. I would think you -- on the financing -- project financing side, I would think you had made significant progress down that path before these transactions. And so I just was wondering, is there any way to quantify sort of the benefit and the lift you may receive in the financing terms or the rate from the significant changes in your credit profile that are taking place? And how quickly can those transactions come together?

And then the second question I had was on the expansion. I think you were targeting 150 megawatts per year, and I think [ you announced ] 150 to 200. I just want to make sure that's true that you're going to probably maybe accelerate your expansion plans in this space now that you have these transactions under your belt as well as what's the future look like. Could that go up even higher as you expand your portfolio?

Yes. Chris, it's Patrick. I'll take the first part and then maybe Paul and Nazar, Kerri can take the second. So Chris, I think it's pretty obvious. You look at Google as a financial partner here working with us, a multitrillion-dollar market cap, one of the largest companies in the U.S. and with a measly $30 billion, I think, of debt or so. So absolutely pristine AA+ company partnering with us, now one of the largest shareholders. So I think you can imagine that whatever course we were on before, right, with regard to our financing for Core42, this is a game changer for that. And so I think we will look to finance the site in totality under this new credit regime and new support regime. So I think that's the obvious and clear takeaway.

And I'm not prepared on this call to precisely quantify that for you other than to say like it's -- we view it and I think our financing partners also view it as a game changer. And so I think now, as Paul mentioned, we are positioned, I think, perfectly to go get the most cost-effective and efficient capital structure in place.

And look, before, we were talking to you about project financing. That can create siloed entities, which is not the best long-term financing outcome. And so I think here, we now have the time and the support to figure out what is absolutely right for the company long term.

In terms of your second question, Chris, around growth and future capacity, we put out this target of 150 to 200 megawatts, and that was by design. And that really is a function of both the capacity we have available, but also what we think we can efficiently finance and pay for. And so with this transaction, one would think we should be able to maybe increase the pace of that. We'll -- as Patrick said, we're going to work through that here very shortly. And so at the back end of that, maybe that could be a possibility. But I think that 150 to 200 megawatts of guidance that we have really is within both the capacity we have available as well as our ability to finance it and pay for that [ construction ] in a timely way.

Okay. Great. Just one more on the financing side. Are you still anticipating a 70% LTV? Or does that potentially have the ability to increase with these transactions?

Yes. Chris, I think all of that is out the door. It's no longer relevant. So I think you'll, again, see us attack the market, I think, in a very significant position of strength now with our partners. And I think all of those prior discussions and targets are off the table.

Our next questions come from the line of Martin Toner with ATB Capital.

Congrats, folks. The EBITDA per megawatt of this deal is a little bit higher than what you've given us for Core42. Can you just kind of talk us through like what's different?

Yes. Martin, I think as you'll see in our disclosures -- and I think this is just natural as we get bigger, right? So before, when we just had Core42 as our customer, we were telling you EBITDA because that was our only customer. I think now what you'll see us do going forward is for Core42 and for this deal, right, we're kind of looking at it from a net operating income at the project level, which I think is a more appropriate way to look at it. And you'll see that in our guidance for bitcoin mining. We're looking at that as segment operating income as well, right?

So I think the way we think about it is there's profitability at the project entities and then there's SG&A that we've had to take on right at the parent, which we gave you updated guidance on today. So I think when you combine the specific project entities, right, and the bitcoin mining, that would then kind of roll up into a total and then you can subtract the SG&A from the top to get down to your EBITDA. Does that make sense?

Yes, I think so. Does Google have a significant contract with Fluidstack kind of like Microsoft and OpenAI have with CoreWeave? And just -- I'm just wondering, does Fluidstack have like a CoreWeave-like backlog?

I don't think we're in a position to disclose Fluidstack's customer base. I think as we mentioned earlier, Fluidstack works with some of the largest companies in the world in delivering customized compute solutions.

I appreciate that. And then last one for me. Can you kind of talk us through why like 95 million shares was the right price for Cayuga? Were there some competing bids? And any detail there you can give us there would be appreciated.

Sure. Martin, it's Kerri. So I think we probably first brought up the Cayuga process back in our first call in February, or at least that's when Paul had begun a process -- a competitive process for Cayuga. That process quickly attracted multiple third-party bids. Because it was a related party transaction, the Board then formed an independent committee, which brought in its own financial and legal advisers to sort of evaluate the opportunity for TeraWulf.

Ultimately, in terms of pricing, the committee approved moving forward at a discount to those third-party bids. And we agreed that equity rather than cash was the right way to structure the deal to keep everybody aligned.

So we're thrilled now to have Cayuga as our second site. Much like Lake Mariner, it's a former site of a coal plant and has all that amazing infrastructure: redundant grid, water, fiber and mostly zero-carbon power. So that's the process that we ran. It was a competitive process. We got third-party bids and then, ultimately, the transaction was done at a discount to where those bids came in.

Our next questions come from the line of Stephen Glagola with JonesTrading.

Just wanted to clarify on that prior Q&A. When you include the extra nonproject level SG&A, is the margin then more comparable then to that 75% level at Core42?

No. No, Stephen. Again, I think I was just trying to make the point that we've kind of -- we've changed how we're guiding you, right, because before we literally just had 1 customer. So it was easy to kind of just give you an EBITDA margin. Now what I'm saying is I expect the net project operating income margin for each project, right, to be its own. I mean I think we're saying, I think, 85% on that. So I think if you were to apply that to, let's say, each of the projects and then take our SG&A guidance, that would get you to EBITDA. Is that helpful?

Yes. No, that's really helpful, Patrick. Appreciate it. And then just had one more. So it seems like the successful delivery of CB-1 and CB-2 are pivotal for showcasing your capabilities and establishing credibility for Tier 3 or 4 data center development. So I was hoping maybe you could share any key learnings your team has gained during the build-out process. Any processes that you maybe expect will be more efficient or streamlined with the Fluidstack build-out or future projects, et cetera?

Naz, do you want to start and let maybe Sean talk a little bit as well?

Sure. Stephen, I think there's a couple of things that are important. One is we've been in construction at our site for the last 4 years straight. So we've been building bitcoin mining buildings. And so we had several electrical mechanical contractors on site. And those folks and some of those folks that showed up 4 years ago are still at the site because they're rolling from 1 building and 1 project to the next. So momentum in being able to deliver on a schedule is important. And part of the diligence process that all of our customers have is really, hey, if there's this time line of schedule out there, what's your ability to meet it?

And so given that we've got quite a bit of activity at the site and there's a lot of momentum, it's easier to see how the existing folks there can roll into the next. And then as we scale up and are delivering greater sizes and capacity, it's really a question of can we augment and add to that kind of labor force. And so that's one thing that I think is important and we've been able to convey and articulate to customers, everyone that comes on and being able to show them these schedules are really grounded in being able to point to who's going to do the work. That's one.

Second is the design of these things, as Paul mentioned earlier, we are in the very early stages here. And we are at our site deploying multiple different OEM GPU equipment. And each one of those comes with its own set of unique attributes and needs and desires. And so having the experience of going through that design process, understanding what those different configurations look like, what they need and being able to deliver solutions to our customers kind of comes through as well. We've been through this process now both with Core42 as well as with Fluidstack and Google now. And so I think people see that, as they're thinking through and figuring out how to deploy various types of technology, we are a very good partner to work with, and especially on the electrical side, delivering that capacity. And Sean probably has a few other things to add.

Yes. I would just add to what Nazar said. A lot of the contractors we've had on site have been there for 4 years. So maintaining the workforce at Lake Mariner has been a huge benefit for our deployment at this site. So rolling them from CB-1, CB-2, CB-3 and so on has really been the momentum that we've kept at the site.

We've also had huge lessons learned from CB-1 and CB-2. And also, as Nazar said, we've been in detailed technical discussions with Fluidstack for the past several months on CB-3. So instead of building a very wide envelope for Fluidstack deployments, we've really narrowed down on the design that meets their needs. So by doing that design and level-setting us upfront, it allows us to further accelerate our construction schedules.

And the only thing I'd want to add to that is the beauty of our bias regionally right now is that all those contractors and many of the same folks on the ground who are employees at Lake Mariner, they move right over to Cayuga when that's ready to roll so that we're taking advantage of all those lessons learned and working with the same people as we continue to refine our efforts here. So it made Cayuga far more compelling to us than just any other site.

Our next questions come from the line of Ed Engel with Compass Point.

Congrats on all the announcements. You just mentioned that you've been interacting with Fluid over the past couple of months. I was wondering what the involvement was with Google kind of during that whole process. Were they involved early in those negotiations or were they a bit more later in the process? And then just kind of curious whether, I guess, GCP themselves has expressed any interest in any other sites in your pipeline.

Yes. We've been involved with Google as part of the Fluidstack effort since very early on. It takes a while to get these things over the finish line. The deal took 3 to 4 months from beginning to end. And we've established great working relationships amongst the parties at every level of their respective organizations. Having a signed deal is more than simply having land and access to power. You have to finance the transaction. It's critical to the magnitude of spend. And that's why it was so important to have Google involved pretty much from the start.

I think Google taking 8% of our company is not simply about doing a deal, but they want to be a part of where we are going given their natural need for compute. And so again, it was complicated, and we've been talking about it for a long time.

Anyone who's listened to Google's Q2 call is aware they have substantial data center needs given $106 billion backlog. They have acquired substantial quantities of third-party data center space. And they're -- backstopping the transaction provides Google with a lower expected cost than a guaranteed value. And since in the unlikely event that Google steps in for Fluidstack due to the backstop, Google would likely add the facilities to their data center portfolio. They were instrumental in helping us get this deal done and very welcome partners.

That's great color. And then you also mentioned that there is early termination protections with Google. Can you talk about some of those conditions, just assuming that they're performance related?

And then just one quick follow-up. The 30-day option that's still being considered, would that also include more warrants to Google as well? Or is that just a onetime thing?

Maybe I'll start, and then we'll go to you, Naz, on the walkaway concerns, which we don't really have. But I would expect that -- a, I have high degree of hopes with respect to the option and the extension of additional capacity. And all parties have worked in good faith to come to the realization that, that's a good thing. And I would imagine that the terms will be very similar, as Kerri pointed out earlier, and that Google's backstop and equity participation would be part and parcel of that. So feel pretty good about that.

Naz, do you want to address the notion of SLA penalties or walkaway considerations?

Sure. I mean the terms or the specific terms are fairly confidential. That being said, the arrangement has what I would call usual SLAs around performance and delivery. We have worked very closely with the Fluidstack and Google teams to ensure that both of those items from -- on the performance side, the design, really incorporates the number of things that they're looking for. So the window of [ operational ] ranges is fairly tight.

And second, on the delivery, there is a fairly robust schedule kind of supporting that. Then again, if you look at what we've done with Core42, we had similar provisions within that agreement. We have an obligation to deliver capacity on a certain time line, and we've been doing that on schedule and on budget.

Our last questions will come from the line of Bill Papanastasiou with KBW.

Congrats on this deal. I think most of the analysts went through a lot of the questions. But with respect to the Cayuga site, it appears capacity will be coming online in the second half of '26. Curious to hear your thoughts on how management is thinking about diversifying the tenant base, if that's a consideration. Or now that you have the 2 partners in your pocket, will you continue to scale hand in hand?

Yes. It's Paul. I've been telling everybody for the last 3 to 6 months, I mean, we need to land a new customer and expand our capacity. We've done that. But in doing so, we've seen a tremendous amount of demand from multiple parties. I think we ended up in a great transaction here with world-class partners. And we want to grow with them and meet their needs. But we have had significant real interest expressed by other quality customers. And we're going to continue to engage with all those parties as we look to grow. We think that's prudent.

But at this point, we're really grateful for the confidence and trust Fluidstack and Google have demonstrated in us right now and really want to get going on that and then focus on CB-5. And then we could see who the right customer is for Cayuga. Demand is real. It's unbelievably powerful right now. So we're excited about the opportunity, but we're most grateful for the contracts that we've announced today and the future of this partnership.

Thank you. Ladies and gentlemen, that does conclude the question-and-answer session, and with that, does bring this call to a close. We appreciate your participation. You may disconnect your lines at this time. Enjoy the rest of your day.