Greetings and welcome to the Core Scientific Investor Update call. At this time, all participants are in a listen-only mode. A question-and-answer session will follow the formal presentation. If anyone should require operator assistance, please press star zero on your telephone keypad. As a reminder, this conference is being recorded. It is now my pleasure to introduce Jon Charbonneau, Vice President of Investor Relations. Please go ahead.
Great. Good morning, ladies and gentlemen, and welcome to Core Scientific Investor Update call. At this time, all participants are in a listen-only mode. We will conduct a question-and-answer session after management's remarks. Please note on this call, certain information presented contains forward-looking statements within the meaning of the safe harbor provisions of the Private Securities Litigation Reform Act of 1995. Forward-looking statements include, without limitation, any statement other than historical or current facts that predict or indicate future events or trends, forecasts, performance, or achievements, and many of which may contain words such as believe, anticipate, expect, estimate, intend, project, plan, or words or phrases with similar meaning. Forward-looking statements are based on current expectations, forecasts, and assumptions that involve risks and uncertainties that may cause actual results to differ significantly.
For further information on these risks and uncertainties, we encourage you to review the risk factors discussed in the company's annual report on 10-K, the company's quarterly report on Form 10-Q filed with the Securities and Exchange Commission, and the company's current reports on Form 8-K filed today and slide presentation posted there. We also have posted an investor presentation to our website at corescientific.com in the presentation section. The content of this call contains information that is accurate only as of today, October 30th, 2025. The company undertakes no obligation to update statements made today to reflect events or circumstances occurring after today. Joining me from Core Scientific are our CEO, Adam Sullivan, our Chief Operating Officer, Matt Brown, and our CFO, Jim Nygaard. We will now begin with remarks from Adam.
Thanks, Jon. Earlier today, we held the shareholder vote on our proposed merger with CoreWeave. Our shareholders voted decisively not to move forward with the transaction. In line with the procedure set out in the merger agreement, we've delivered formal notice to CoreWeave to terminate the deal. Since announcing the transaction, we haven't had the opportunity to engage with investors in a public setting given the clear terms written in the merger agreement, and I'm excited to have the opportunity to do so today. Leading Core Scientific is a privilege, and I'm thankful for that opportunity every day. I love being CEO of this company, and I'm continually inspired by the talent, the drive, and the resilience of this team. This team is exceptional, and the opportunity ahead of us is the strongest it has ever been in our history.
We are the foundation of the AI revolution, building the critical infrastructure that powers the technologies of tomorrow. I want to thank our shareholders for their engagement and candid feedback throughout this process. The questions and dialogue have been extremely valuable, and we appreciate everyone's thoughtful input. Importantly, it was clear from our discussions throughout this process that shareholders overwhelmingly understood the industrial logic behind a combination with CoreWeave, including how it created a stronger combined company with significant financial synergies. That rationale was always the foundation of our discussions and the reason we pursued the transaction. And while there was criticism around how the deal was structured, it was designed specifically to preserve upside value potential for our shareholders and CoreWeave stock while ensuring the ultimate decision rested squarely in the hands of our shareholders.
CoreWeave's decision not to raise their offer confirms that our board negotiated the best deal possible for our shareholders. That said, the outcome of the vote ultimately reflects broader market dynamics, given demand and valuation changes over the past four months, as well as the recent performance of CoreWeave stock. Now that we can speak freely, I want to take this opportunity to clear the air on some of the rumors and speculation that have been circulating in the market. First, let's be very clear. We are executing on our objectives. We have one of the strongest development teams in the industry and a portfolio of strategically located sites that position us exceptionally well for the next generation of high-density AI workloads.
We are delivering significant buildouts across several locations, including the anticipated 2025 buildouts we've previously spoken about for CoreWeave, and have been making consistent progress under our existing agreements. I'll let Matt share some of those details, but let's not forget, we are the first company to sign a large-scale AI colocation agreement last year. Since then, we've been executing at a speed and a scale that no one else in the industry is matching, including delivering more liquid-cooled MW this year than any other publicly traded company, and we're on track to do it again in 2026. These are extremely complex, high-density buildouts, and the reality is very few companies have the experience, the people, or even the committed power to deliver them. We do.
We are the first in building an all-star operations team with the technical expertise and discipline to execute, and you're seeing those investments pay off today. Others trying to replicate this model will likely face real challenges, and that's no surprise. This is hard work, and it takes time, coordination, and a level of operational maturity that does not happen overnight, and with that said, we've been upfront with our shareholders about some of the short-term challenges we face, including permitting delays, weather impacts, and design modifications along the way. That is to be expected, though, of a program of this size and complexity, but the reality is that we're still going to deliver more high-density data center capacity in a single year than some operators have built in a decade. The facts are simple.
We're delivering, and we continue to execute against our commitments, and we're remaining on track to deliver 250 billable MW by year-end and the full 590 MW to CoreWeave by early 2027. Second is our power pipeline. Some are quoting total or gross MW, while others, including us, focus on billable or revenue-generating MW, a measure we've consistently used in discussing our pipeline in the past. We've also taken what appears to be a much more conservative approach in how we present our pipeline compared to what you might be hearing from others in the industry. So let me walk you through the numbers and how to think about them. To start, we have a total organic pipeline today of roughly 2.3 GW of gross capacity or 1.5 GW on a billable capacity basis.
Separately, we've also paid for targeted load studies based on thoughtful capacity feedback from our utilities, totaling an additional one gigawatt of gross power capacity. In total, this would imply we have a power pipeline of roughly 3.3 GW of gross capacity. Note historically, we've applied probability weightings for publicly disclosed pipeline metrics. We are taking this approach because we want to give you a realistic execution-based view of what we can deliver, not an inflated one. To that point, on our first quarter earnings call in May, we spoke about roughly 700 MW of incremental billable capacity existing in new locations. This only represented a subset of our actual pipeline at the time, specifically the portion we had a high degree of confidence we could secure by the end of 2027, which is consistent with the projections we showed in the proxy statement.
The reality is we now have more power opportunities in front of us than we can likely build over the next several years. To put it plainly, our challenge is not in the size of the pipeline. It's about prioritization and disciplined execution. Finally, I want to discuss CoreWeave. Our partnership with CoreWeave remains exceptionally strong. As you know, they are currently our only customer and will continue to be a very important partner. Together, we're executing under a pay-for-pay joint execution risk contract, a structure that is very strong for us and we believe among the best in the industry. It ensures both companies are fully aligned on meeting key milestones and delivery commitments. But let me be clear. Going forward, our relationship with CoreWeave is purely a commercial one. Our contracts with them are straightforward.
These contracts represent more than $10 billion in total revenue potential to us over 12 years, supported by strong margins. And unlike other deals in the industry announced this year, CoreWeave is funding the overwhelming majority of the capital expenditures tied to these buildouts. At the same time, the contracts do not bind us to additional business with CoreWeave, do not contain change of control provisions that limit our flexibility, and we will do business with other players in the industry. Continuing to execute on our contracts with CoreWeave remains a major focus for our business, and our teams work very closely together every day to deliver on multiple large-scale buildouts that are actively underway across several sites. Next, I will turn the call over to our Chief Operating Officer, Matt Brown, who will provide a detailed construction update.
He'll be followed by our CFO, Jim Nygaard, who will provide his views on financing projects going forward. I will then return to this call with our go-forward plan and what you can expect to hear from us in the coming months. Matt.
Thanks, Adam. And just to clarify, all references to MW are for billable capacity unless otherwise noted. Through 2025, our teams have remained focused on disciplined execution of a clear mission: build and deliver the infrastructure that powers large-scale AI and high-performance computing. In Texas, our 250 MW Denton campus continues to advance. On the Denton North campus, we have completed energization across three buildings, totaling 90 MW and 132,000 sq ft of a 460,000 sq ft campus, a key milestone. At Denton South, construction momentum remains strong, with the next building phase rapidly progressing towards energization in the coming months.
At our 65 MW Marble North Crown Electric facility, two of three data towers are energized, with the third moving through final energization later this quarter. In total, we have energized more than 120 MW a day of the total 590 MW of the CoreWeave contract. In the fourth quarter, we advanced two major builds from structural to interior phases. Dalton phase I, 30 MW, and Muskogee, 70 MW, have both completed work construction with new precast buildings in place. Trade partners are on site setting critical equipment and progressing interior electrical and mechanical system work that will continue over the coming months, bringing both sites closer to energization as we enter into 2026. During the fourth quarter, we broke ground on the first 145MW building at phase II, launching another AI factory for CoreWeave, with initial delivery anticipated in the second half of 2026.
Alongside scheduled projects, we introduced our first generation two AI factory design, which debuts at Muskogee. This next generation design is optimized for NVIDIA GB200s and GB300s, supporting up to 200 kW per cabinet via direct-to-chip liquid cooling and is engineered for greater flexibility and resiliency, enabling efficient adaptation between current and future GPU architectures. Our progress this year reflects a repeatable delivery playbook and a design philosophy that stays ahead of what AI customers need. With our Denton and Marble sites beginning energization, as well as our Dalton and Muskogee sites moving through interior buildout, we're executing the plan while bringing on our Gen Two architecture to market. We reaffirm our 2025 energization target of 250 MW, underpinned by recent progress at Denton and continued energization activities at Marble.
Looking ahead, we're executing towards a goal of delivering an additional 350 MW by year-end 2026, of which 280 MW are dedicated towards the current CoreWeave contract and the remainder for new customer contracts. Outside of the CoreWeave contract, we have begun pre-development on approximately 500 MW of billable capacity slated for 2027 energization. We're proactively investing in future builds by securing long lead equipment, strategically acquiring additional land, and positioning the company for sustainable long-term growth. Our development plan is anchored by three pillars: scaled repeatable delivery, next-generation density and efficiency, and customer-aligned capacity. Taken together, these initiatives position us to expand high-density AI infrastructure while maintaining discipline around schedule, cost, and technical performance. Our outlook reflects current assumptions and is subject to risk and dependencies, including supply chain timing, permitting, and utility interconnects. With that, I will now turn it over to our CFO, Jim Nygaard.
Thanks, Matt. I want to take a moment to address another misconception that we were not able to raise capital for buildouts beyond CoreWeave if we remained an independent company. Simply put, that is not the case. We believe we're in a unique moment in the market, one where customers, investors, and capital providers recognize the strength of our position, a position that includes a strategically located and highly sought-after data center footprint, a very strong 12-year foundational contract with CoreWeave, a deep pipeline of power assets, as Adam outlined earlier, and an operations team that I believe is the best in the industry. More broadly, demand remains exceptionally strong, with multiple gigawatts of data center capacity already leased by hyperscalers this year, either directly or through NeoCloud providers. Leasing momentum in the market is expected to continue as AI infrastructure buildouts accelerate.
To help fund these buildouts, we've seen more than $10 billion in debt and equity-linked capital raised across the industry in just the past few months. Many of those deals have included Bitcoin miners who have secured financing through a wide range of creative structures and new strategic partnerships as they transition from mining to high-performance compute infrastructure. More specifically, capital providers have already approached us looking to put money to work. Now, with the merger behind us, we believe we're in an excellent position to raise financing with attractive terms to fund a significant amount of growth opportunities beyond the existing CoreWeave contract. We don't believe this is a question of whether we can obtain financing.
It's a question of how we do it most efficiently, balancing the overall long-term cost of capital with structures that provide flexibility in how and where we deploy it, which hasn't necessarily been the case in some of the deals announced more recently. Our goal is to build scale over time, creating a virtuous cycle that ultimately drives down our cost of capital and strengthens our competitive position in the market. To that end, we plan to be opportunistic, including pursuing a mix of corporate-level financing solutions combined with more targeted project-level debt for individual sites or even multiple sites which share common financing attributes. Importantly, our CoreWeave contract, where CoreWeave is funding the overwhelming majority of the more than $5 billion of infrastructure buildout, also represents a significant financing opportunity and, quite frankly, a tremendous competitive advantage for us.
As we've shared publicly, we believe we can raise up to $4 billion of capital against that contract. This structure would be highly flexible, enabling us to fund multiple non-dilutive equity investments at the project level. These projects generally support 60%-80% advance rates on build costs, depending on the customer credit and site characteristics. As you know, we laid out a five-year plan in the merger proxy, which projected that we would double our billable MW portfolio with an estimated build cost of more than $7 billion. However, today, given the opportunities we're seeing, we believe that number could be significantly higher. Finally, from a balance sheet perspective, we're targeting a long-term leverage ratio of roughly five times stabilized adjusted EBITDA, which we believe provides the right mix of growth capacity, capital discipline, and flexibility to capture the opportunities in front of us.
Now, back to Adam for commentary on our go-forward plan.
Thank you, Jim. This management team and board is extremely excited about the road ahead and the opportunity to create significant long-term value for our shareholders. With that in mind, we expect to make some key announcements before our next earnings call. First, we anticipate signing at least one new customer, an important step towards diversifying our customer base beyond our existing CoreWeave contracts. Second, we plan to sign at least one new power contract at an existing site that we could not sign while operating under the operating restrictions of the merger agreement with CoreWeave. Third, we also expect to sign a new major power contract at an entirely new site location. Fourth, we anticipate making a financing announcement that will help bring more clarity on how we intend to fund future buildouts.
And finally, I want to make it clear that we plan to convert every MW in our portfolio into high-density colocation sites over the next three years as we continue to wind down our mining portfolio. Before opening the call for questions, I want to take a moment to thank all of our employees for their hard work and dedication through what has undoubtedly been an uncertain few months. Our team's focus, professionalism, and commitment have been extraordinary, and we wouldn't be in this position with a stronger business and significant opportunities ahead without them. With that, operator, we can open the call up for questions.
We'll now be conducting a question-and-answer session. If you'd like to ask a question, please press star one on your telephone keypad. A confirmation tone will indicate your line is in the question queue. You may press star two to remove your question from the queue. For participants using speaker equipment, it may be necessary to pick up the handset before pressing the Star keys. One moment, please, while we pull up for questions. Thank you. Our first question is from Darren Aftahi with Roth.
Hey, guys. Thanks for my questions. Good to talk to you again. Just to, if I may, I guess on that one gig of incremental capacity, what's the realism in terms of timeframe, what feasibility studies that could kind of come in-house? And then I guess in light of maybe some of the restrictions you had during the M&A process, I guess how much sort of engagement could you have with any prospective customers that you kind of have spoken about in terms of objectives in the next couple of months? I guess said another way, were you able to engage with any prospectives while the M&A process was going on, or are we kind of clean slate as of today? Thanks.
Thanks, Darren. Yeah, on the first point related to power availability, the power pipeline that we outlined is specifically power that would be available, and it's available either now or we have line of sight being available over the course of the next few years. So it's an important distinction. That's not power that has to go through load studies. It does not have to go through feasibility studies. That's power that either exists in existing today or is power that is currently under negotiation with utilities. So it does not have to go through that entire process related to the items that you outlined. Related to the customers, as permitted under the merger agreement, to continue to operate business as usual, we were able to have conversations with customers during that time period.
I think the one important thing to note is there were some customers that did not want to speak to us while we were under the merger agreement. However, later in the process, as things started to trend publicly towards looking like the shareholders were going to vote against the deal, that did begin to open up some doors, so we've been able to work through with a number of different clients through their engineering process, and we are currently working through terms with a number of clients today.
That's helpful. Thanks, Adam.
Our next question is from Jon Peterson with Jefferies.
Oh, great. Thanks, Adam, Jim, Matt. Welcome back. Good to hear your voices again. Curious if you could maybe pull the curtain back a little bit for us on your discussions with CoreWeave in recent weeks. Adam, it kind of seems like the door is maybe closed at this point on potentially them coming back with a higher offer. I mean, I just want to make sure that I heard that correctly, that you feel like that was the best and final offer, and at this point, we're just moving forward as a standalone company.
Yeah, John, I would just reference the merger agreement between our company and CoreWeave has been terminated, and I believe based on a number of public disclosures made by CoreWeave and Mike, that was their best and final offer.
Okay, and the financial impact, I believe a payment needs to be made to CoreWeave now. Can you remind us on what that is and the timing on when that payment happens?
Yeah, John, there is no payment due to the fact that our shareholders voted on the deal. So there is no termination payment related to this transaction given that our shareholders voted against the deal.
Oh, okay. Okay. That's good to know. Maybe then just on your business moving forward, just curious on the power pipeline, are you able to break down for us how much of that, the numbers you talked about, the gigawatts, the 3.3 GW, how much of that is getting additional power to existing sites, and how much of that requires you to go out and get new sites?
Right now, a significant amount of that power is at existing sites. That's really where our focus is today. We spent a significant amount of time acquiring land nearby our existing sites to give us the opportunity to continue to build out once that additional power is acquired. I noted earlier in the five items that were discussed at the end of this call that we are expecting to sign additional power at least one existing site before our next earnings call. That's a process that we're working through today, and we believe that is our fastest pathway forward given we have contractors either on the site or going to those sites, and we'll provide some opportunity for more straightforward growth plan going forward.
Okay. Sounds good. Thank you very much for the call. I appreciate it.
Thank you.
Our next question is from Brett Knoblauch with Cantor Fitzgerald.
Hi, guys. Thanks for taking my question. Adam, on maybe the CoreWeave contract, I feel like there's been some speculation about energization delays. I just want to make sure I kind of heard your prepared remarks correctly in that before you guys were expecting the full 590 to kind of be up and running, I believe by the first quarter of 2027, it seems like everything is still on track. Is that right?
That's correct. Right. Everything is still on track from our last update to the market.
Awesome. And then I would say a number of peers have maybe started to sign deals with maybe non-NVIDIA GPU, I don't know, compute sources going into the data centers. Is that something that your data center team has looked into or can also go down that route as well?
Yes. The short answer is absolutely. While we've certainly optimized our NVIDIA-based designs, we're not vendor agnostic. I mean, we are vendor agnostic, so we'll be able to have designs that are adaptable for a number of different hardware stacks and GPU architectures, both NVIDIA and non-NVIDIA systems.
Awesome. And then maybe just one last question on maybe just the speed at which you can build. What is a realistic, you guys said that you have more power than you could easily build out over the medium term. What if you get all these deals that you want. How much can you actually build in one year? Is it 300 MW at your upper limit? Can you push that further? Assuming demand isn't as insatiable as we think it is, what is the max that you guys think your team can go out and build?
Yeah. I think, as we sort of indicated in the prepared remarks, that we're certainly planning on the ability to deliver 500 MW to the market in 2027, so right now, that's kind of the new benchmark we're setting for ourselves.
Awesome. Thank you, guys.
Thanks, Brett.
Our next question is from Joe Flynn with Compass Point.
Hi, guys. Thanks for the question. Related to the existing CoreWeave partnership, as it relates to kind of as you guys are building out the sites and get increased power allocations, would you ultimately expect to see opportunities to CoreWeave expand their agreements? Or have you been able to talk to customers, maybe expand that, go multi-tenant at certain sites, given that a lot of CoreWeave and customers are now buying GPUs as of late?
Yeah. Thanks, Joe. I mean, from our perspective right now, it comes down to economics, and that's how we're evaluating each new deal that comes across our desk. There's a number of different players in the market today, and we're going to continue to evaluate each available MW that we have on a MW-by-MW basis. And so that's our plan moving forward, and we will continue to make decisions and continue to announce to the market as we make those decisions and sign new contracts.
All right. Thanks. And from the financing side, I know you guys are reiterating the $4 billion you can take out of the existing contracts. But ultimately, what is the realistic timeframe for that? Do you have to ultimately be closer to stabilize NOI, or does it have to be done in conjunction with CoreWeave? Any information there would be helpful.
Joe, thanks for the question. The availability will scale as we continue to build out. We think we're going to have a lot of upfront capacity and that type of lending arrangement on the CoreWeave contract directly. We also have existing liquidity on our balance sheet today. We finished the quarter very strong in that regard, and we also have very deep options at the corporate level to facilitate an expedited buildout, so all of the options are on the table, and we'll continue to optimize that equation.
Great. Thanks.
Our next question is from Joseph Vafi with Canaccord Genuity.
Hey, guys. Good morning, and welcome back from me as well here. Just wondering if, given where things are and a lot has actually happened in the market since the deal with CoreWeave was announced, if maybe you contemplate some GPU purchases of your own at scale and thinking about some GPU cloud. I think with a power pipeline that is emerging, maybe it could be a good use of some smaller uses of power. It's how you think about the GPU NeoCloud opportunity directly. Thanks.
Yeah. Thank you. I mean, what we're really seeing in the market today with a number of these players are really a data center value-add product. They're offering GPUs that oftentimes even being run by the end clients, and they're calling it cloud business. I would say there's a differentiation there, and that's something that we've had discussions with a number of different customers about. It really comes down to their want or desire to transfer CapEx to OpEx, and there's definitely optionality in that realm. I would say one of the things, though, that we've been evaluating more recently are there are a number of new players coming to market with extremely low cost of capital compared to existing providers of some of these NeoCloud platforms.
And we think there's going to be a dramatic shift in who are the winners among that market, and there are going to be a lot of losers in that market. And so if cost of capital is the name of the game, we're looking at long-term structures with a number of different counterparties that we believe will be extraordinarily competitive in that industry. And we're still leaving the door open on the other side where potentially some smaller customers may look to do some GPU-based deal with us as well.
Great. Thanks, Adam.
Our next question is from Kevin Dede with H.C. Wainwright.
Hi, Adam. Thanks for hosting this call, and thanks for having me on it. Congrats on teeing up a potential second customer. Can you give us any color on that? I mean, it's hard to imagine that contract would be anywhere as lucrative as the CoreWeave deal, and I'm wondering if there are any parameters that you might offer to help us sort of think it through.
Yeah. Thanks, Kevin. I completely agree with you. Looking back at the CoreWeave contract that we signed, it is the best deal in data center history and has not been replicated or has anyone ever come close to signing a deal of that magnitude and that level of profitability. So I completely understand your point, and one of the things that we've been evaluating with new customers is the fact that we do have frontier capacity available and utilizing that as a tool to work towards signing longer-term commitments on power, potentially anchor tenants to some of our new facilities that will help us on the financing for those buildouts.
And so we've considered a number of different factors here, but I'm going to hand it over to Jim to talk a little bit about how we're thinking about the new contracts that we're evaluating today and hopefully provide a little bit of color on how we're thinking about this long term.
Yeah. And Kevin, we're in a pretty unique position in the market in that because CoreWeave is funding the overwhelming majority of the capital expenditures on the buildout, we can leverage that contract and essentially use that money for the equity that is required at new greenfield projects. So when we look at our cost of capital and we look at our return target, essentially we're able to borrow much lower-cost equity dollars than other companies that are required to come up with other sources of equity that would be far more expensive.
We do not have hard and fast rules on target returns on any one site. We certainly analyze that, as you can appreciate, in great detail, but we think about it on a portfolio context. For the right customer and the right geography under the right circumstances, we may accept a lower return to anchor a site and be able to build upon that in the future in a synergistic fashion with higher return contracts to complement it. But overall, you're maximizing the return of the portfolio. We're going to continue to take that approach. We're excited about the opportunities that are ahead of us, and we candidly have a lot of wood to chop.
Yeah. I'll echo those sentiments, Jim. A lot of wood to chop. That sort of leads me into my next question. Congratulations on navigating all the things that you have. I think one of the biggest surprises for me is the fact that you had all these things running in the background with the CoreWeave deal on the precipice. And I guess what I'm wondering is how were you able to do that, number one? And number two, what do your new priorities become? Because Adam laid out a bunch of things, including two new power contracts, new financing, and this ability to deliver high-density compute.
Yeah. Thanks, Kevin. We were permitted to continue to operate under the merger agreement as business as usual, which allowed us to have conversations with clients. We continued to have conversations with our utility partners. And the reason is because it was advantageous also for CoreWeave for us to continue those conversations. We were looking at acquiring well in excess of a gigawatt of power, and we continued those negotiations, continued those discussions related to new PPAs for those utilities. And this was advantageous to both parties, whether we remained a standalone or we ended up merging with CoreWeave. We've put ourselves in a position today where we are looking to act upon a number of the things that we set up during the past few months.
And that's really the most important aspect here is we're starting day one as a standalone company, not under the merger agreement, ready to begin executing contracts. And that's the most important thing here. And so we're working towards a number of different things on the power side that we believe will be extraordinarily advantageous to this company, continuing to work towards finalized terms with a number of different clients across our entire site portfolio. Almost every MW in our portfolio is under discussion right now. And so that puts us in a position today where we feel very strongly about our standalone process moving forward, and we only would have been able to do that if we continued to work hard for the past four months. And so that's really how we're thinking about it today. We're in a great position, Kevin.
Okay. One last one, if I may, Adam, please. I understand that you invested pretty heavily in helping CoreWeave design the technology that they needed to support the stack on top of the GB200 and GB300s that they're rolling out at your facilities. I'm wondering how that intellectual property that's inherent now to Core Scientific can translate to the development that other customers might want to use.
Yeah. No, that's a great question. I think the advantage here is we're certainly able to take all these lessons learned, both in terms of the process of constructing at scale, the coordination with supply chain and what that's involved at large-scale projects, and what designs and what we should be optimizing for from a design standpoint. So we're able to sort of culminate all those lessons learned, all the best practices, all of the design iterations, and we'll be taking that up into future design improvements for future customers going forward.
Okay. Fair enough. Thank you very much, gentlemen, and congratulations. Kudos.
Thanks.
Our next question is from Brian Dobson with Clear Street.
Hey, thanks very much for taking my question. So I suppose that you've been on the receiving end of many, I guess, interested calls regarding demand for HPC utility, right? So as you're looking at potential clients, hyperscalers, what are they telling you about their demand needs over the next few years, and how do you feel about the broader market as you look forward?
Thanks, Brian. The broader market right now is robust. We're seeing today, oftentimes, many of the hyperscalers migrating their capacity needs through other providers. Then we're also seeing look-throughs up to the credit of those main hyperscalers, allowing those new providers to come to market, sign larger contracts at better rates, and also allow for the financing markets to open up. That's why we're seeing so many new providers coming to the market today. It's because the financing market has come around to looking through to the up all the way through to the hyperscaler credit. I think that's what we're going to see for some amount of time here. The market, from what we're seeing, there's a number of hyperscalers that are looking for gigawatts of capacity. Really, what we're seeing is the demand is almost insatiable.
And so the market's definitely changed dramatically since the beginning of this year, where we were seeing much fewer large-scale buildout demands. Today, that looks much different. We're seeing demands for anywhere from 100 to 500 MW with options to scale up to a gig at a single site. I think those opportunities are fewer farther between, I think, both from a demand perspective as well as an actual site availability perspective. But I think there are going to be more of a reckoning here in terms of who these folks sign contracts with as you start to see more slippage from newer entrants to the market who are looking to develop data centers. I think right now it's too early to see those cracks that are in the market, but you're going to start to see cracks among some of these other developers.
That's going to be through extended delays, the choice of the wrong general contractors, or there's a number of other issues that could present themselves as these companies start to work through that. I think that's really going to force a pretty thick funnel here down to a very few data center developers that are going to end up being the winners in this industry. That's how we're thinking about it today. We have a great opportunity ahead of us. We think we're in position one. We had a potential client, large-scale client, mention to us that after speaking with our design engineering team, they can see why we are one of the leaders in AI data centers globally. That's really something that we've taken to heart here.
It's really kudos to our entire construction and operations team that's really best in class and putting us in a position to win new contracts today.
Excellent. Thank you very much.
Our next question is from Paul Golding with Macquarie Capital.
Thanks so much for the question. Adam, you mentioned that every mrgawatt in your portfolio is going to be used for colocation or HPC. I wanted to ask, when it comes to PICOs, what's changed there in terms of the original plan to consolidate the Bitcoin mining into that facility? Is latency less of an issue now for prospective customers? Are you just seeing the pricing equation look more favorable for that type of site relative to latency and other resources? And then as a follow-up to that, where does that leave you with disposal of those ASICs, ProTo, and Block? How should we think about that whole ecosystem? Thanks.
Yeah. I'll take the first part of the question here around site suitability for data center conversion. I think it's going to be true of really all of our legacy HPC facilities, so all of our HPCs, including HPCs. I think what we've seen happen here in the market is a number of things. First off, we took a really fresh look at the infrastructure that I think we had some doubts. I think we had some reservations about how convertible and how adaptable those facilities would be. For HPCs, for instance, it was, how do we solve the latency and network connectivity issue in West Texas, which is, to be noted, while we think power is a constraint in the marketplace, when you go to a region like West Texas, connectivity is actually a much larger constraint than power in many locations.
And so what our team did was go back and go solve the connectivity and latency problem, which we have. And so we have a good plan of attack for how we're going to build massive amounts of connectivity to that PICO site that will improve the latency outlook and performance for that location, making it a pretty valuable asset for us for future conversion. The next thing is more sort of market-driven. And so as we continue to see these large-scale demand requests come to market, what's naturally happening is this demand is pushing out not only just from the tier-one markets, but it's pushed out to the tier-two markets in the more tertiary areas where the power is available.
So as long as we're solving for the connectivity equation, we have the power secured, we're finding that more suitable use cases and workloads to be able to target those facilities for. And the second part of the question, Adam, if you want to take that one.
Yeah. So we're really thinking about the wind-down of the Bitcoin mining business in either two ways. The first is we continue to take the new units that we're receiving today. We start to sell off some of the units that we do not have space for. That method will provide us the greatest flexibility for being able to maintain our existing power contracts, continuing to draw power at those sites while we are shifting that infrastructure to high-performance computing.
The important note on that strategy is that we've been acquiring land next to each of our sites to be able to run parallel paths to build out our data center capacity while still operating the Bitcoin mining business. The second path would be looking at potential to exit those rigs and enter into hosting contracts with a potential counterparty for the remaining time available at those sites. It's yet to be determined which path we're going to go down, but what we know right now is that we are going to continue to mine at those sites and continue to meet our utility obligations.
Thanks so much.
Our next question is from Greg Lewis with BTIG.
Hey, thank you and good morning or good afternoon, and thanks for taking my questions. I guess I had one question just following up on Paul's. I think you were alluding to it, Matt, as kind of maybe the kinda tier-two data center locations have become more interesting. I guess what I'm wondering is, and realizing you've been quiet for the last few quarters, have you started to notice any bifurcation between what pricing might look like for locations maybe that are more urban versus more rural? Or is there anything that's driving any kind of differentiation between pricing of locations that's kind of developed over the last couple of quarters?
Yeah. The way that I would think about that is it calls a couple of different vectors here, one of which is just short-term capacity. If you have near-term capacity, you have a little bit of pricing and leverage there from a pricing standpoint, right? If the power is sort of longer-term out, you certainly don't have the same sort of pricing power in the market that you would otherwise expect. So it really kind of depends on whether the capacity in these tertiary regions are short-term or long-term. And then just from a general market standpoint, the cost of construction, the cost of development in some markets is just really high, right?
And so that's certainly going to drive the economics of what sort of price point you're going to be able to come at into the market and still maintain a return profile that's attractive for making that investment or signing that contract. And so I would say it's like multiple different vectors here. We have to sort of triangulate to determine what we think the optimal pricing is. And so there is, to sum that up, yes, there are changes from market, but it's also changes not just by market, but also by timing of availability of power.
Okay. Perfect. Super helpful. Thank you very much.
Our next question is from Michael Donovan with Compass Point.
Hi, Adam. Thanks for taking my question. Just to follow up on HPC questioning, you have historically been able to push rack densities. With future buildouts, how are you thinking about builds to possibly handle Blackwell Ultra or Rubin?
Right. So as we said in our prepared remarks, we've brought our Gen 2 architecture to market, which is really sort of tailored around both the current and next-generation GPU architectures. The team is already working on our Gen 3 architecture, which is sort of solving for these higher density needs, as well as potential changes in delivery voltage as we've seen more popularized through the Open Compute Project and NVIDIA sort of pushing 800-volt DC. So we're actually sort of preparing for that additionality to be able to adapt to those future architectures. We're thinking about it. The team's already been working on it for a number of months.
Appreciate that, Adam, and then just to follow up on a pipeline question. So you mentioned demand in the range of 100 to 500 MW sites. Is there a size that your team favors for colo modular builds? And then with inference starting to pick up, should we expect smaller sites to come online?
I'll try to take this one. I think the way we think about it is not so much the size of the site itself necessarily. It's really what are the building blocks? And so as we sort of optimize the incremental building blocks of the site, that sort of allows us to do a couple of different things. One, we can sort of take a bigger campus, but we can build it in 20- and 30MW chunks, and then we can productize and build different data center products for different types of end users and workloads as we go along. So that could be lots of smaller customers in a shared building, each with their own data hall, as an example. So more of a large-scale wholesale type of configuration, which might be suitable for inference given the geography of wherever that site and that product might locate.
But it also sort of sets us up for more multi-tenant configurations and being able to set ourselves up for more enterprise-type workloads, in addition, while we're combining that with our large-scale agent tenants at a single site. So I think we look at that as a 100MW site or a 500MW site kind of matters less than the building blocks themselves.
Thanks, Adam.
Our next question is from Tim Horan with Oppenheimer & Co.
Thanks, guys. Obviously, you have world-class engineers both on the design and on the construction side. And I would assume that they're in incredible demand at the moment. Have you lost any of them through this process, and how are you retaining these engineers? Thanks.
Yeah. So we haven't lost any engineers in the process. In fact, we've actually been growing that team throughout this entire process. So I think the question is, how do we retain them? And there's one thing that's true about all engineers and about all high-performance teams. Everyone wants to work for a high-performance team, and engineers like working for other rock star engineers. So we've built a team of rock stars. And so that's one reason. And the second reason is these guys want a challenge, right? We're doing something that's pretty unprecedented in this space, right? Building AI factories at scale. The projects are incredibly challenging. They are engineering masterpieces of themselves. And I think that's what engineers want. They want to be challenged. And they want to not only be challenged, but they want to be a team on a team of other rock star engineers.
So that's what keeps the talent in place. We keep them challenged. We continue to build resources around them, give them the support they need so they can do what they do best. And that solves our retention problem.
Can you talk about from a critical path scheduling of these projects? What's the longest lead time for equipment? Are the bottlenecks improving there? Are they getting worse? Just any color around the whole construction process would be helpful.
Yeah. So there are kind of a multitude of bottlenecks that one has to traverse, particularly in the current state of the market. Certainly, long lead equipment is an issue. So supply chain is still constrained. The supply chain constraints are largely sort of targeted around medium voltage and standby generation predominantly, depending on data center design and configuration that sort of governs the availability and how long that equipment may be out. And the second piece of the question being the other constraint that we're kind of drawing in the marketplace is just the labor. There's only so many electricians. There's only so many GCs that have availability.
Our ability to sort of stay ahead of the curve, having a lot of sites to subcontractors, electricians, mechanical contractors that can do the work to fill our pipeline is just as much of an importance as staying ahead of the supply chain curve and making sure we have orders, continuous deliveries of long lead equipment. Those would be. I would frame our constraints as those two things.
Lastly, if you don't mind, a lot of utilities are running out of capacity, and some are requiring large customers like yourself to put up upfront capital to help them build out and ensure that you'll be around. How confident are you that you can get the capacity that the utilities are promising you? And are you now required to put up some capital yourselves beforehand?
We are seeing increased demands on utility partners for capital to secure the future power. That trend will likely continue. And we do expect to follow suit on that. Look, at the end of the day, we have long-standing relationships with our utility partners that really afford us a number of benefits in terms of giving them transparency on what our objectives are. And the fact that they've had experience working with us, I think, puts us in the best competitive position possible. So again, I put that in the category of optimization as we think about our buildout costs collectively. But even with those increased requirements from a capital perspective, we still are looking at projects with attractive returns.
Thank you very much.
Our next question is from Stephen Glagola with Jones.
Hey, thanks for the question. Welcome back. Adam, in your prepared remarks and some verbiage in the Q&A, you mentioned that the opportunity set for new power buildouts is robust and that the main challenge is execution rather than access to power pipelines. I was just hoping maybe you can elaborate on that, sort of given that the prevailing market view is that quality power pipelines are scarce. And then I have a quick follow-up.
Yeah. I think this goes back to the roots of this business. As you look at our history, our company has the strongest history in the industry of sourcing and locating and developing large-scale MW sites. That's in our DNA, and given our nimbleness with our team to be able to execute on sites that become available, our ability to due diligence them quickly, our ability to work through negotiations and discussions with utilities provides us an advantage in the market that we are capitalizing on and will be capitalizing on here very shortly as we continue to expand our power portfolio, and so that's really the advantage we have right now, and it really comes down to a disciplined approach to how many MW we'll be taking on each year to ensure that we continue to meet our customer expectations for delivery.
Because the more slippages you have in your schedules, the faster you're going to lose credibility with these large corporations who are signing up to get delivery on a timeline that they have their own plans for, whether it be moving in, selling capacity, whether they've sold that capacity potentially onward as well. But these are the customers that you're going to be working alongside for 10 years, 15 years, 20 years. And starting off the relationship by being late on execution, it's not a great place to start off. And so we want to ensure that we continue to meet all the expectations we set forth in our discussions with these counterparties. And we believe that's going to help us continue to lead this industry moving forward.
Appreciate that. And then last one from me. Are you considering the convertible market as part of the potential financing mix? Thank you.
All options are on the table as it relates to our financing alternatives, and we will continue to optimize that equation.
Thank you, guys.
Thank you. There are no further questions at this time. This does conclude today's conference. You may disconnect your lines. We thank you again for your participation.