Greetings and welcome to the TeraWulf 2025 First Quarter Earnings Conference Call. At this time, all participants are in a listen-only mode. A brief question-and-answer session will follow the formal presentation. 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, sir. You may begin.

Thank you, operator. Good morning and welcome to TeraWulf 2025 first 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 on sec.gov. And in the Investor section of our website at terawulf.com. We will also reference certain non-GAAP 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 this morning's call with prepared remarks from Paul and Patrick, followed by a Q&A session. I'll now turn the call over to our CEO, Paul Prager.

Thank you, John, and good morning, everyone. We appreciate you joining us to review our first quarter 2025 results. It was another active quarter for TeraWulf, as we continued to build on the strong momentum from 2024 across both our Bitcoin Mining and High Performance Compute or HPC businesses. We remain committed to executing our strategy, maximizing the value of our megawatts through scalable sustainable infrastructure. Before turning to the business highlights, I want to take a moment to thank you, our shareholders, especially those of you who have supported us since the very beginning. We are positioning TeraWulf to lead at the intersection of energy and compute. It's a long-term effort, but one that we believe will create substantial value over time. Starting with WULF Mining, our sustainable Bitcoin Mining platform at our Lake Mariner facility in Upstate New York. During the quarter, we successfully energized Miner Building 5, bringing total operational capacity to 245 megawatts. We exited the quarter with a self-mining hash rate of 12.2 exahash and fleet efficiency of 18 joules per terahash. As we mentioned on our February call, extreme weather in January and February temporarily impacted power pricing. However, by March and well into April, pricing normalized, and our mining operations returned to positive EBITDA in the month of April. I also want to confirm that we received and installed all of our S21 Pro miners before any potential tariff implications, ensuring uninterrupted deployment. Turning to WULF Compute, our High Power Compute hosting platform. Our mission here is to scale our purpose-built liquid-cooled infrastructure to meet the growing demand for AI and compute intensive workloads. Demand remains strong, especially from enterprises seeking secure high density infrastructure over the next 12 to 18 months. We are focused on working with prospective partners that have capital to deploy, secured GPU allocations and credit profiles that support project-level financing. Today, we are actively executing on three dedicated buildings for our HPC anchor tenant, Core42, the WULF Den, CB-1 and CB-2. These facilities are our top construction and operational priority, following five months of close collaboration with our partner and their integrator and many consultants. The WULF Den is operational and will begin generating revenues in Q2 and we expect CB-1 to go live in Q3 and CB-2 in Q4. Executing on these initial facilities will drive further demand for our site and further interest in partnering with the TeraWulf team for highly complex HPC deployment. Our partnership with Core42 is progressing exceptionally well. We collaborate daily to align on technical specifications and deployment timelines to give a sense of the scale and sophistication involved. Dell, Core42's integrator, expects to have over 180 personnel on site during the GPU tuning phase. We are optimizing design elements every day to ensure our infrastructure meets both current and future demand. These refinements may accelerate or delay timing and could affect overall infrastructure costs. Our goal is to design infrastructure that will support future generations of GPUs, so we are committed to getting it right the first time. A successful launch will not only position us for further potential expansion with Core42, but also establish TeraWulf as a leader in next-generation data center development. Success here will also accelerate our broader HPC hosting strategy. Additional prospective tenants are closely watching our progress, and we are actively engaged in discussions to secure new commitments as we build out our capacity. Let me touch on a few additional updates. First, capacity. In April, we received approval from NYISO for an additional 250 megawatts of capacity at Lake Mariner, bringing the current total to 500 megawatts. We intend to request an additional 250 megawatts that will bring our total power at Lake Mariner to 750 megawatts. We appreciate the support and collaboration of NYISO as we scale this site. Second, tariffs. We're monitoring the evolving tariff landscape. Based on current information, we estimate a 5% to 10% impact to build costs. We remain committed to underwriting projects to mid-teens unlevered returns and will adapt as needed to protect project economics. Third, project financing. We remain on track for mid-year execution of the project financing for our 72.5-megawatt Core42 buildout. We are scheduled to officially launch the process next week with our advisors at JPMorgan and Morgan Stanley, and early feedback from potential lenders has been positive. Finally, our growth pipeline. Beyond our 750-megawatt roadmap at Lake Mariner, we continue to pursue expansion opportunities. At the top of the list is the Cayuga Site, a sister facility to Lake Mariner located on Lake Cayuga in Upstate New York. It shares the same strategic advantages in land, power, and fiber access. That process is progressing with our Board, and we'll share more when we have firm updates. We're also evaluating additional sites with strong time-to-power potential and opportunities for on-site generation. This is an area where our energy infrastructure expertise truly differentiates us. Lastly, I want to address the Company Services agreement with Beowulf Electricity & Data, a private company owned by me that currently provides electricity and digital infrastructure services to TeraWulf. We believe the time is right to simplify the structure. We are currently pursuing a full integration of Beowulf and TeraWulf to eliminate related party disclosures, streamline operations, and better align incentives across the organization. The process is driven by our Board and guided by rigorous governance protocols and independent oversight to ensure transparency and shareholder alignment. To summarize, our key near-term priorities are: One, optimize our self-mining platform following the energization of MB-5. Two, deliver all three Core42 buildings on time and on budget. Three, lease additional HPC hosting capacity at Lake Mariner. And Four, close project financing for the Core42 buildout. With that, I will turn it over to our Chief Financial Officer, Patrick Fleury.

Thank you, Paul. The first quarter of 2025 presented challenging market conditions for our Bitcoin Mining operations with a temporary spike in power prices and increasing network difficulty, the impacts of which are reflected in our financial results for the quarter. While our EBITDA was slightly negative for the quarter, it's important to emphasize that we are carrying significant incremental costs related to our expansion into High Power Compute hosting without any current revenue contribution. As I'll discuss later in my remarks, this temporary burden will soon be addressed as our HPC hosting buildings come online in 2Q, 3Q, and 4Q 2025 as depicted on Page 15 of our May Investor Presentation. In the first quarter of 2025, we self-mined 372 bitcoin at Lake Mariner or approximately 4 bitcoin per day, a 12% decrease over the 423 bitcoin mined in 4Q '24. Our GAAP revenues were flat quarter-over-quarter at $34.4 million in 1Q '25 from $35 million in 4Q '24. Our value per bitcoin self-mined in 1Q '25, a non-GAAP metric, averaged $92,600 per bitcoin as compared to $82,739 in 4Q '24. Our GAAP cost of revenue exclusive of depreciation for 1Q '25 was $24.5 million, a 25% increase over $19.6 million in 4Q '24. The quarter-over-quarter increase was due to a 37% increase in realized power prices from $0.509 per kilowatt hour in 4Q '24 to $0.801 per kilowatt hour in 1Q '25, offset by demand response proceeds of $1.3 million in 4Q '24 versus $2.8 million in 1Q '25. Our power cost, or cost of energy per bitcoin mined, a non-GAAP metric, was $66,084 in 1Q '25 compared to $46,328 in 4Q '24. As mentioned previously, this temporary spike in power prices, which began in December and persisted through mid-February 2025, was historic as it resulted in a 1.76 standard deviation spike in New York ISO Zone A West average energy prices for January and February versus the average for this period over the last 10 years. We expect the remainder of 2025 to be in line with historical power pricing at Lake Mariner, with guidance of $0.05 per kilowatt hour for 2Q through 4Q '25, which is in line with the New York ISO Zone A forward power curve as of May 5th, 2025. Operating expenses increased 6% quarter-over-quarter from $2.7 million in 4Q '24 to $2.9 million in 1Q '25, following a 69% increase from 3Q '24 to 4Q '24. This trend higher is primarily the result of increased staffing levels at Lake Mariner necessary to support our mining expansion as well as our entry into HPC hosting activities. SG&A expenses increased quarter-over-quarter from $32.3 million in 4Q '24 to $50.1 million in 1Q '25, primarily due to stock-based compensation in 1Q '25. Adjusting for stock-based compensation, SG&A decreased quarter-over-quarter from $15.5 million in 4Q '24 to $11.5 million in 1Q '25. Depreciation increased slightly quarter-over-quarter from $14.9 million in 4Q '24 to $15.6 million in 1Q '25. Loss on fair value digital currency in 1Q '25 was $0.9 million compared to a gain of $0.6 million in 4Q '24. GAAP interest expense in 1Q '25 was $4 million compared to $3 million in 4Q '24, and we recognized interest income of $2.2 million in 1Q '25 compared to $2.6 million in 4Q '24. Cash interest paid during 1Q '25 and 4Q '24 was negligible as the 2.75% interest on our $500 million convertible notes is accrued and payable on May 1st and November 1st. Our GAAP net loss in 1Q '25 was $61.4 million compared to a net loss of $29.2 million in 4Q '24. Our non-GAAP adjusted EBITDA for 1Q '25 was negative $4.7 million, down from positive $2.5 million in 4Q '24. Turning our attention to the balance sheet. As of March 31st, we held $218 million in cash with total assets amounting to $841 million and total liabilities of $670 million. Through March 31st, 2025, we spent approximately $130 million of capital expenditures on WULF Den and CB-1 and CB-2. As disclosed on Page 17 of our May Investor Presentation, we achieved a BTC segment cost of production of approximately $72,000 in 1Q '25, and our guidance for 2Q through 4Q '25 is anticipated to be well below this at approximately $47,500, primarily due to the aforementioned decrease in forecasted power prices. Regarding fixed operating expense guidance for 2025. Page 18 details our anticipated SG&A, operating expense, and interest expense provisions. These costs reflect significant increases in the number of employees at both TeraWulf and Lake Mariner as we grow our business and expand in High Power Compute hosting. On Page 12 of the May Investor presentation, you'll find our updated total cost to build and net yield on cost analysis. Since signing the Core42 leases in December, we've worked hand in glove with Core42 and our respective partners including Dell and other consultants to refine our preliminary data center designs into a final high-quality product. The result of this tightly knit partnership and design process is an increase in total capital expenditures of approximately $65 million from $365 million to $430 million, along with a commensurate increase in initial rent from approximately $1.5 million per megawatt to approximately $1.6 million per megawatt. The net result of these changes is a slightly lower net yield on cost but a 14% increase in year one EBITDA. The incremental spend takes our CapEx per critical megawatt from $6.1 million to $7.2 million, well within our long-standing guidance range of $6 million to $8 million and well below industry peers. This page highlights the unique value of the existing infrastructure and site-specific advantages of building and operating High Power Compute loads at Lake Mariner. Regarding our capital position and growth plans for the remainder of 2025. Page 14 provides a capital sources and uses bridge. Our data center financing led by JPMorgan and Morgan Stanley will officially launch next week. We have a high degree of confidence in executing an approximately $300 million debt raise in the middle of 2025. In anticipation of a sizable unallocated 2025 cash balance, the Board has authorized a new $200 million share repurchase program for an incremental approximately $150 million to the remaining approximately $50 million on our pre-existing program. We also intend to file an updated ATM prospectus supplement of $200 million, which is a housekeeping exercise as the current ATM with approximately $87 million remaining has lapsed at year-end. While some may find it unorthodox to have active buyback and ATM programs, given TeraWulf's historical realized stock volatility of approximately 130 and current macroeconomic conditions, the management team and Board find it prudent to have every tool in the tool shed available and at our disposal. Finally, as a management team, we are repeatedly asked how we value a megawatt of long-term contracted High Power Compute capacity. One simple analysis we regularly reference internally is depicted on Page 16. In summary, the Core42 leases for 60 megawatts of critical load are worth approximately $2 per fully diluted share outstanding. And every incremental 50 megawatts of capacity contracted at similar economics is worth an additional $1.30 or so per fully diluted share. We hope you find this slide useful as a simple reference tool to measure our success as we announce further High Power Compute deals. With that, I'll turn it back to the operator and we look forward to answering your questions.

Thank you. We will now be conducting a question-and-answer session. Our first question comes from Nick Giles with B. Riley Securities. Please proceed with your question.

Hey, thank you, operator and good morning, everyone. My first question, it's good to see the integration of TeraWulf and Beowulf. And so how should we think about any potential cost savings for the parent? And what does the timeline look like there? Thank you very much.

From a timeline perspective, this is a very rigorous process because it's a related party transaction. So the Independent Directors of the Board have hired Independent Counsel, Independent Financial Advisors, and have negotiated an arrangement with the ownership of Beowulf. Until those Independent Directors approve the transaction, and the audit committee approves the transaction as required by the operating company's articles as well as by insurance, I can't comment on that. That's a process they run independently. We think it's something that will be very enabling to TeraWulf's shareholders long-term to just have everybody in one place. Ultimately, it will drive real value in terms of development and of additional sites as well as construction of High Power Compute on our existing site. So we're pretty excited about this opportunity and we think it's near-term. But again it's really up to the Independent Directors of the Board of the company.

Paul, it's very helpful. I appreciate those comments. My second question, one thing that's always stood out about your economics is really the lower capital intensity and you mentioned the tariff impact of 5% to 10%. So my question is, what are your expectations for build costs more broadly, particularly as you look to Cayuga or other potential sites? Could some of these other sites have the same capital intensity that Mariners afforded?

Patrick, do you want to field that, or would you like me to?

Go ahead. I'll clean up.

I think we have guided the markets to $8 million a megawatt. Our range was basically $5 million to $8 million. We've tightened it up to $6 million to $8 million. The answer is, we're working closely with Core42 and the integrated Dell, and we've had several design changes that have had some cost implications. But that is, there’s a mechanism within our agreement to capture that so nobody should be concerned there. But we think it enables us to get to a better standard design for a 50 megawatt gross building, 42 net megawatt building that we could just replicate over and over again for both the existing customer, Core42, and other customers that come to our site. So, by the way I look at that, the learnings, if you will, of the slightly higher price of the finalized design is pretty much optimal leading edge data center design that gives us a competitive advantage going forward. But we're pretty much at that range of $6 million to $8 million and it could be at the high end of that range. But I'm good with that.

Yes, Nick, if I may add a little bit to that. So, as you can see at our new Page 13 in our deck, I think this page really highlights A) the uniqueness of Lake Mariner and the other power sites that we have and also our ability to take those sites, remediate them. I mean, I think this really encompasses all of our strategic advantages. And I think what Paul just said about this close-knit partnership with Core42 and all the consultants we're working with puts us in a competitive advantage vis-à-vis our peers because now we have what I would call a perfected design at a cost that is very attractive. So, yes, I think it's - again, even if you take tariffs on top of that sort of $7.2 million critical build cost that we talked about of 5% to 10%, you're still within the range that we've been guiding to for a long time of $6 million to $8 million.

Great. Well, Patrick, Paul, I really appreciate all the color and continued best of luck.

Our next question comes from Darren Aftahi with ROTH Capital Partners. Please proceed with your question.

Yes, good morning. Thanks for taking my questions. Just two if I may. On Page 10, there's commentary I think around EBITDA margins being incrementally higher on future capacity. So two questions around that. I guess the first one is in terms of an additional tenant or an existing tenant expanding into future capacity, what sort of bogeys need to be seen in order for that to happen? Is it getting operational buildings at scale and then people can take the next step? And then I guess the question around incremental adjusted EBITDA. What kind of gives you confidence on that? Is it just lease rates and you think that CapEx can stay the same so yields will be higher? Just any kind of color would be helpful. Thanks.

Hey, Patrick, if it's okay, I'll answer the first part of the question. You could get the second.

Sounds good.

Yes. So Darren I don't think there's a bogey out there that we've got to tap in order to sign up additional capacity. I think those conversations are ongoing. We've been having them since we signed up Core42 or in advance of that even and we're talking to both Core42 and additional customers. So I think what we've tried to share with people is that, we sense, by the way, our stock is priced in the market that some people still don't sort of believe that we're very much a show me story. And we're really comfortable with that. I think if you want to bet on anything, you want to bet on the execution capabilities of TeraWulf. And so I think that it is natural for customers who are trying to whether it's Neoclouds or hyperscalers or enterprise folks. I think it's natural for them to, while they are organizing what their needs are, want to be able to look at something. It's the difference between, it's being able to go in and touch it and see it and hear it and you have all those machines humming. And it's just a pretty fantastic thing when you get into one of our data halls. So while I agree with you that we built WULF Den in order to be able to achieve that, and in fact we did, we landed the preeminent customer in the space in Core42. I think when we energize CB-1 it will be even more profound in terms of how people want to react to us and get to the place where they're finally willing to contract. So my view is that no bogeys are required. We're open for business now. We're negotiating with a bunch of folks and we're going to sign up deals if they make sense, if they're top quality credits, if they're the right duration today. But I believe, once we energize, CB-1 we'll be even busier, if that's possible, in sort of discreetly determining who the right customers are and signing them up to the appropriate contract. So I think that's where we are in terms of landing our next expansion. Patrick?

Yes, Darren, with regard to the margins, I think we've had this page in our deck since last year, last May. And there was some confusion because we initially had Fed margin 65% to 75%. And then I think I had spoken publicly, Paul, Nazar and others about how really that was EBITDA margin. And the midpoint of that, the 70% was actually, we thought about it as effectively cash available for debt service, so meaning EBITDA less maintenance CapEx. And so what I've done here is just to reduce that confusion and put in like this is an EBITDA margin. We expect it after lots of work with our partner, our operating partners, and other consultants, our hiring plans, actually executing on the ground that, that EBITDA margin will be at about 75% on the first 72.5 megawatts. We do expect we'll have going forward maintenance of about 3%. And then the callout on the right of Page 10 to your point of higher incremental EBITDA margins. This is the benefit of a big site, okay, versus having multiple sites all over the country. I only have to hire one or two security guards. I don't have to hire 10 because I don't have five different sites. So there's really significant. Most of the costs that go into that EBITDA margin for the High Power Compute business are labor. It's just that simple. We got to have a bunch of people on the site. And so once you have that base cost, which we already have, as I mentioned in my remarks, everything beyond that is incremental and comes in at a much higher margin. So I'm not going to tell you what that margin is today. I'll tell you when we announce additional capacity, but expect it will be much greater than 75%.

That's helpful. Appreciate all the detail. Thanks.

Our next question comes from Mike Grondahl with Northland Securities. Please proceed with your question.

Hey, guys, thanks. I wanted to ask, you've spent a lot of time, many months with Core42. What have been two or three of the biggest learnings for WULF through all those discussions?

Nazar Khan, are you on, and do you want to address that? I don't know if the operator is allowing him to enter. Why don't I field it, and if Nazar can get on, I'm happy to. Listen, it's been tremendous. The team that Core42 has brought to bear would indicate that we're in the very, very early stages of these data centers. Each generation of chips has its own particular inherent requirements in terms of the design of the rack. So what I like about Core42 is that they've just shared everything with us and our design team and construction team so that we've been able to appreciate why some parties focus on cooling in the building, some focus more on cooling on the racks, and some on the weight of some of these water-cooled racks. It's a constantly fluid situation because this is new technology and it's sensitive, and people are trying to figure it out in real-time, at the same time that there's this massive demand for High Power Compute. So I think the biggest lesson learned we have had is that we were right to pick the right partner. Because in our power experience, when we were developing power plants, if we had this level of fluidity as we worked towards the final design, which is where we're at now, because as I indicated, we're going to be energizing CB-1 and CB-2 very shortly, it would have been a far more confrontational kind of affair. But here, it's been a real collaboration and very cooperative, and these design changes we had weren't changes as much as an evolution of design, if you will. We had mechanisms within our agreement, and Core42 has been entirely stand-up about it. So I think the lesson learned is pick the right partner, and you'll get to the dance in the right place and have a great time. And so that's been the lesson for me. Patrick, anything you want to add?

I think Nazar is unmuted. Nazar is certainly better suited to answer this question than I am. Go ahead, Nazar.

Sure. Good morning, Mike. Nazar here. As Paul said, I think, when we signed the agreement with Core42 in December, they'd gone through 40% or so of all the design specs. Since that time, we've had the chance to go through every single component of the design, the size of every single valve, the location of every single pipe, the location of every single kind of busway. So now we've had a chance to go through 100% of the design with them. And through that, we've come up with a whole host of things that we're working with them on incorporating into future buildings as well. And so, as Paul mentioned earlier, we are in the early stages of the evolution of this business. What we build in CB-1 and CB-2 likely will not be what we build for CB-5 and CB-6. That process of really working with our customer, partner Core42, and being able to go through every single component of the design and thinking through what the trade-offs are, is important. For a lot of these decisions, there's no perfect answer, it's always a trade-off. Having that discussion with them and being able to take those learnings and reflect that on what we want to do next has been a very important part of the process. Those learnings are going to be invaluable because both the business is evolving and the underlying hardware equipment is evolving, all of those trade-offs and the decisions we've made are going to have to be revisited as future generations of those GPUs roll out as well. So that for me has been the biggest lesson—importance of having a partner to work with and not just simply a customer who signs a piece of paper.

Great. Hey, thanks for the insight.

Our next question comes from Brett Knoblauch with Cantor Fitzgerald. Please proceed with your question.

Hi guys. Thanks for taking my question. Congrats on getting the additional 250 approved for Lake Mariner. It seems like this year is definitely focused on energizing the three buildings for Core42. You now have 250 capacity and the charts show, next year you can kind of really allocate additional capacity. Would you look to start breaking ground on say CB-3 or CB-4 before signing an additional tenant? And is that something that you would look to start doing this year or would that be a 2026 thing?

Paul, do you want me to take that one?

Yes.

Yes, Brett. So I think that the short answer is we do some spend mostly preparing, so pads, but it's relatively minor, getting us whether it's site prep or site electrical. It gets us into a position so that when the customer is ready to execute, we can go right away. So I would say there's some site-level expense that we're constantly incurring. If you look in our presentations, you can see that. Like we have a site electrical and infrastructure column, and we've had that in every single one of our decks. There are some site-specific things that we constantly are doing. For example, you got to get electricity, high voltage electricity, around the site to the various buildings from the substations. So there's prep for that, and then, like I said, there are preparing buildings and pouring pads and things in the appropriate weather. There are those types of things. But I think anything that's significant and major, we would not expand without a signed agreement.

Perfect. That's helpful. And then Patrick, maybe just a follow-up on capital allocation priorities to re-up or increase the buyback. Obviously, there's a need for a lot of CapEx to be spent over the next few years. How do you weigh buying back shares versus spending that on infrastructure to sign additional tenants? Is there a point where one is more attractive to you? What is that point, or is that just kind of like an ad hoc ongoing discussion that you guys look at?

Yes, look, I'll answer that, and then Paul can chime in. But it's a management, board level decision that we constantly evaluate, and as you know, the purpose of having the buyback. Again, I think, as you'll see, as you heard in my remarks, we also had about $87 million left on our ATM. And so, as part of that process, we increased it from $87 million to $200 million. So we don't intend to use that. But I want every tool in the tool shed just like I want a buyback. As you know, Bitcoin mining in particular is an incredibly dynamic business—profitability can change dramatically every day. We can go from making very small amounts of free cash flow to very significant amounts of free cash flow very quickly. Other things can change also, for example, project financing. We've been targeting 70% loan-to-cost. Many of our peers are out there doing 80%, 90% right. So that's a lot of money that would come back to us if that changes. It's really my job particularly to advise the management team and the board to make sure we have every single tool available. Given the volatility in our stock, I want them all at my disposal, Paul's disposal, and the board's disposal.

Awesome. Really appreciate it, Patrick. Thank you, guys.

Our next question comes from Brian Dobson with Clear Street. Please proceed with your question.

Hey, good morning. So as you're out in the market speaking with other potential clients, what are you hearing from enterprises and hyperscalers? What do you feel about the near-term demand environment, but also the medium-term?

This is Paul. Thank you for your question. From a near-term demand perspective, I think people are very keen to sort of get power now. So we've seen a lot of energized land deals in the market. And so I think that demand is real. I think from a data center perspective, the hyperscalers are still out there. They're refining what they're looking for. I think there was initially this sort of global let's tie up everything we can. Now I think they're focused on, okay, they have a better sense of their needs and a better sense of what makes a site a great site. Lake Mariner is certainly that. So we have a lot of inquiries all the time and we have inquiries on development sites as well. I think near-term demand is real if you have proximate energy. If you could demonstrate that you have a few hundred megawatts near-term with the ability to scale after that, the demand is there. You have to understand these customers. A lot of them don't understand energy. Most of the world is out there advertising 1 gig and 2 gig sites, which sounds really great; they need to wade through a lot of that craft to appreciate that bringing the right level of quality in terms of electricity to a data hall is tough on the grid operators. You got to really study how robust is the grid? So I think the hyperscalers are getting smarter every day and starting to lean towards higher quality sites with higher quality owners who understand energy and infrastructure. Enterprise customers, the demand has been constant. Neoclouds, I think they're running around trying to allocate their book and then figure out what their needs are and how that scales not only in terms of capacity but in terms of calendar scheduling. From a mid-term perspective, we have said all along, we're less sanguine on the likelihood of increased generation hitting the market at the scale that it needs to within the next few years. It just takes a long time to develop a power plant. That enables higher values for us because we have what we have, and we're energy infrastructure folks that can develop things in a cost- and time-efficient way, more so than most of our peers. We see a lot of demand, so much so that we are starting to categorize the customers coming in based on—we want to sell data hall by data hall and we're seeing interest now from people that want to come in and take half of the data hall. So our enterprise customers are showing up in the marketplace. We're pretty constructive, and we just need to continue to execute.

Yes, thanks. That's really helpful. Just shifting over to the mining business. We've seen elevated global hash rates throughout the last quarter and continuing now. What's your outlook there and how do you plan to combat that as we head into the next halving over the next couple of years?

So we can't do a whole lot to combat the hash rate. We have to live with it. I think we are happy that the higher price of Bitcoin and hopefully want it to go higher benefits those who have Bitcoin and guys like us who have really low costs in generating Bitcoin. We've brought on the most recent generation of miners, and we're back to really low costs from an energy perspective, and we operate about as efficiently as anybody ever could. It’s a tough business being in mining, and we've guided the market to the belief that at this point in time, especially given what we see as the demand for our product in HPC and AI, we want to really continue to focus on that. We're certainly enjoying the benefit of having a mining operation now, but that in time, prior to the next halving, we'll probably look to take those megawatts and deploy them in the highest value, and we think it's likely to be HPC/AI. There's an additional benefit to that as Patrick has spoken about, which is predictable revenues over a long period of time and you get good project financing terms on that. We're excited about Bitcoin going higher. We have a real Bitcoin zealot, Nazar Khan, on our management team, and we'll continue to mine as long as we can make lots of money doing it.

Hey, Paul. If I may, Brian, just to address your question a little bit more specifically from the finance side. We just came through a really tough quarter where we had for us a one in ten-year type event on power prices. Even with that, we went through our fleet upgrade. We didn't bring miner building 5 on until the very end of March, early April, fighting with a hand tied behind our back. When you consider that the entire business is burdened with the cost structure of High Power Compute and we don't have any revenue yet. The results, actually, were not what we wanted but pretty darn good when you consider again we're running with that whole cost structure. We had really high power prices. Now as Paul mentioned in his remarks, we’ve seen in April that's all subsided and price has kind of ticked up, and we're making money. A quarter a year in Bitcoin Mining can be a lifetime. Our asset efficiency of our fleet is among the lowest of our peers. We're about 18 joules per terahash now. I think we are positioned well to reap the cash flow rewards of that business should Bitcoin move higher.

Yes, agreed. Thanks.

Our next question comes from Stephen Glagola with JonesTrading. Please proceed with your question.

Hi. Thanks for the question. Paul and Nazar, thanks for your color on sort of the Core42 design discussions. I was hoping you could elaborate on—if the current delay for additional capacity is primarily related to Core42's customer visibility or are they waiting to see how the first 72.5 megawatt buildout goes? And then I had a follow-up.

Yes, I don't like the word delay because I don't think that's the real word. The discussions with Core42 are going exactly as they're supposed to go because they're really focused, as we are, on delivering what we have contracted, right. It's pretty intense and collaborative. The end result is we're hitting the ball out of the park. The reason for that is that when CB-1 and CB-2 are energized and they're up and running, that enables Core42's ability to aggregate customers and enables our ability to talk to other parties interested in the site because there's something they could see and touch, and they can get excited about. A lot of people have said they were building data halls and turned out to be big buildings that cost a lot of money that didn't achieve their intended purpose. That's not what we're doing. We're building something that we can replicate, continue to improve upon, evolve over time, and that fulfills our customer's contractual requirements. I think having CB-1 and CB-2 energized enables us to get better terms, better pricing. I just don't feel we had a deadline to land another customer. We had an option within an agreement for them to have more because they wanted the option. But quite rightly they never talked about it; we've just talked about how do we grow together. I remain very constructive on our ability to sign up an additional 100 to 150 megawatts and put it in the ground each year. That's where we are on our next expansion.

Thanks, Paul. Appreciate that. And then, Patrick, on Slide 15 of the presentation, the hosting timeline. Do you anticipate the 170 gross megawatt capacity planned for 2026 to be brought online gradually throughout the year, or will the full 178 come online towards the end of the year 2026? Thanks.

Yes, so there's no customer for that right now that we're announcing today. That's what we have, sort of the arrow in the quiver. As we've mentioned, I mean, I think we're trying to perfect a 50 megawatt design so that capacity is effectively three different buildings. As we move forward in time, Stephen, I think you'll see us from signing a lease to bringing a building online, a reasonable timeframe is 12 months. Over the next six to seven months here before year-end, obviously, to bring that capacity online by fiscal year end 2026, we would have to announce contracts for that capacity and start on it by year-end.

Thanks, Patrick.

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

Hey, guys. Thanks for taking my question. First one, as you start to deliver for Core42, do you think that gives quite a bit more comfort to either enterprises or major hyperscalers for some of that additional capacity? Is that kind of the right way to be thinking about it? And then I have a follow-up question.

Yes, I've been saying that this entire call. It's one thing to think about what you want to have for breakfast if you're me, right? I'm a short fat guy and I like breakfast. It's another thing when I walk into a bakery and I look at the cinnamon rolls and I get really, really worked out and I'm a buyer. The answer is when we build and energize this data hall, it moves from the stage of talking about something and talking about design to actually delivering to a customer who is also delivering onwards to their customer base. It's a huge deal and it makes a tremendous difference in terms of our ability to sell our capacity at the terms and price that we require. So I think again Patrick likes to say that we’re bummed about our stock price. The management team here are serious holders of stock and we're bummed about it. But we got to manage the company in the best interest of the shareholders. We must recognize that we're a little bit of a show-me story because we're building a data center for the top customer in the business. This is new technology, new business for everybody. There will be doubts until we energize CB-1, and that’s okay. We care about execution. Once we deliver on an energized data hall, I think we're ready to roll. It’s an important element that we must energize CB-1 and have a happy customer because that will change the entire profile of the market.

Yes, fair enough. That makes sense. And then maybe one for Patrick.

No, think about it. There are only two customers right now in this space. Two, not customers, companies that have made this transition out of mining into HPC/AI, and that's Core42 and us. The market is tough in the current economic environment. People tend to view the glass as half empty instead of half full. They’re being skeptical. That’s okay—they’re doubters. But once you deliver an energized CB-1, that all goes away. Our shareholders will benefit from seeing greater value outright, and we'll all benefit because customers will say, 'Wow, that’s great. Core42 is so lucky. I want that too,’ or customers will go to Core42 and say, 'Wow, that’s great. I want to be one of your customers. Can you get some additional capacity from TeraWulf, and we'll be ready to go?' That’s what's happening.

Got it. Thank you, both.

Our next question comes from Joe Flynn with Compass Point Research and Trading. Please proceed with your question.

Hi, thanks for the question. You guys kind of ultimately answered it. But based on the project financing side, it seems like the corporate guarantee there will be a lot of interest in that and pretty tight spreads. But maybe just talk about like what the appetite is, and where you ultimately hope to partner with, and any color would be great.

Yes, look, I think there's a lot of chatter out there in the marketplace. You can see where deals are pricing. Generally, hyperscaler risk is SOFR plus 200. CoreWeave risk is kind of SOFR plus 400 to 500 but is tightening. I would expect our customer, in my opinion, to be a better credit quality than CoreWeave but not as well known. I would expect we're targeting somewhere around that same pricing of SOFR kind of 400 to 500 and hope to do better than that. But I think for our first financing, that's a good target range.

There are no further questions at this time. I would now like to turn the floor back over to Paul Prager for closing comments.

I want to thank all of you again for joining us. TeraWulf is very well positioned at the convergence of energy and compute. Our scalable sustainable infrastructure, attractive cost profile, and strong project pipeline provide a foundation for long-term value creation. As we move through 2025, our focus remains on execution. The deployment of our initial HPC buildings will mark a major inflection point, shifting us from promise to proof and unlocking new revenue streams for the company. As a significant shareholder myself, I want to reaffirm that our actions are aligned with long-term shareholder interests. We appreciate your continued support and confidence in TeraWulf. Thank you.

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