Core Scientific, Inc. Common Stock Q2 2026 Earnings Call

Greetings, welcome to the Core Scientific Fiscal second quarter 2026 earnings conference 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 during the conference, please press star zero on your telephone keypad. Please note this conference is being recorded. I would now like to turn the conference over to your host, Jon Charbonneau, SVP of Investor Relations. Please go ahead. Good morning, welcome to Core Scientific's second quarter 2026 earnings call.

Before we begin, I need to remind you that statements made on this call, other than historical facts, are forward-looking statements within the meaning of the Private Securities Litigation Reform Act of 1995 and are based on our current expectations. Words such as anticipates, expects, intends, believes, and similar words and expressions are intended to identify forward-looking statements. These statements are subject to risks and uncertainties that could cause actual results to differ substantially. For further information on these risks and uncertainties, we encourage you to review the risk factors discussed in the company's reports on Form 10-Q and 8-K filed today with the Securities and Exchange Commission, and the press release and slide presentation contained therein.

The forward-looking statements we make today speak as of today, we do not undertake any obligation to update any such statement to reflect events or circumstances occurring after today. Today's presentation is available on our website at investors.corescientific.com. The content of this conference call contains information that is accurate only as of today, July 28th, 2026. Joining me today from Core Scientific are our CEO, Adam Sullivan, our Chief Financial Officer, Jim Nygaard, and our Chief Operating Officer, Matt Brown. We will conduct a question and answer session after management's remarks. We will now begin with remarks from Adam.

Good morning, everyone, thank you for joining us. This morning, we announced a commercial partnership with AMD for up to 2.5 gigawatts of data center capacity, a clear validation of our deliberate strategy to begin development and construction across multiple locations before customer contracts were in place. Beyond their scale, contracted value, long-term growth potential, the agreements underlying this relationship reflect the strength of the company we've built, the discipline and conviction behind our approach, and the significant opportunity still out of us. Throughout this process, our priority has been to form the right relationships that recognize the value of our portfolio. The most valuable arrangements in this market are not one-off transactions. They are the ones with the potential for significant expansion over time.

Our agreement with CoreWeave began as a 16-megawatt lease at our Auburn campus in 2024 and has since expanded to 590 megawatts of total contracted capacity. That progression is important. Our partnership announcement today with AMD reflects this potential. The initial agreement represents more than $14 billion of base contracted revenue across the 15-year agreements with 2.5% annual escalators. Core Scientific will deliver 530 megawatts across five sites, which is one of the largest single deals announced among our peers. With this announcement, we will have two customers that have each committed to over 500 megawatts each across five campuses. Approximately 380 megawatts will be delivered directly to AMD under a triple-net lease across Pecos, Hunt, and Muskogee.

The remaining approximately 150 megawatts across Auburn and Dalton will support a Neocloud through a modified gross lease for which AMD will provide full credit support throughout the full 15-year lease term. Through this initial agreement, Dalton and Auburn will be fully leased. Importantly, the long-term opportunity at the remaining campuses extend well beyond the capacity included in these agreements. As we previously disclosed, both Pecos and Muskogee have the potential to support up to one gigawatt of leasable capacity through a combination of additional grid-connected power and behind-the-meter solutions. Notably, the credit support agreements do not include any equity step-in rights like those included in certain other transactions announced in the market, protecting our equity investment in these projects. As a sign of this partnership, we have issued a warrant to AMD with a strike price reflecting current market levels, investing subject to certain commercial conditions.

The structure with AMD provides meaningful potential long-term revenue, durable contracted cash flows, greater customer diversification, and substantial utilization of our leasable campus portfolio. The initial 530 megawatts of contracted capacity represents only the first phase of what we believe can become a much larger strategic relationship. AMD, at specific times and under specific conditions, has the exclusive reservation right to lease as much as two additional gigawatts. We believe we can make this power available to AMD through a combination of incremental grid-connected power, capacity progressing through load studies, and behind-the-meter solutions across Pecos, Hunt, and Muskogee. This structure positions us to grow alongside AMD as its infrastructure requirements continue to expand over time. Our decision to provide AMD with expansion options across our portfolio reflects both our confidence in its position within the AI ecosystem and our belief in the long-term growth potential of the relationship.

AMD is building significant momentum in a rapidly expanding market as hyperscalers, AI labs, cloud providers, and enterprise customers increasingly adopt its advanced computing platforms, making it a highly attractive and strategic counterparty for Core Scientific. The scale and structure of the relationship are important. AMD's decision to work with Core Scientific also reflects confidence in our ability to deliver. Discipline explains why we continue to seek only the right commercial agreement. Execution explains why we want it. AMD had the opportunity to evaluate not only the quality of our power and real estate portfolio, but also our demonstrated ability to develop and operate highly complex AI infrastructure across multiple campuses. Over the last year, we have shown that we can move from contract execution to construction, energization, and revenue generation at significant scale.

That experience is also enabling us to collaborate closely with AMD on co-design initiatives shaping the future of our campuses to drive greater efficiency and speed across their GPU and CPU products. Our execution capability is not theoretical. Today, we are pleased to announce we are ahead of schedule and currently billing for 437 megawatts of capacity, tangible evidence of our ability to move from signed agreements to delivered operational infrastructure. We believe this distinction will become increasingly important as the market shifts from evaluating companies primarily on the deals they announce to also assessing their ability to execute. Value is not announced, it is delivered. Delivering hundreds of megawatts of high density infrastructure requires far more than access to power. It requires an integrated development and operating platform capable of designing, building, and operating complex infrastructure reliably and consistently at scale.

We've built those capabilities. They position us to deliver against our existing commitments while continuing to establish and expand our capacity agreements with leading companies across the AI ecosystem. The agreements announced today materially increase the scale and diversification of our contracted portfolio. Core Scientific now has approximately 1.1 gigawatts of total contracted billable capacity, representing more than $24 billion of base contracted revenue. Just as important, we achieve this growth without compromising the principles that have guided our strategy. We have remained disciplined in how we value our power, allocate our campuses, assess customer credit, and evaluate the risk-adjusted returns of each opportunity. The result is a stronger, more diversified platform with greater revenue visibility, substantial embedded growth opportunities, and contracts with some of the most important companies in the AI ecosystem. Our focus now is clear. Finish the 150 megawatts remaining in the CoreWeave build-out.

Successfully build and deliver the capacity leased today. Position ourselves to expand our existing customers over time. Continue growing our site portfolio for additional new customers. Over the past year, our priority has been converting existing power capacity from Bitcoin mining to high density colocation. As we enter the next phase of our growth, we will complement that strategy by expanding our power portfolio through the selective acquisition of powered land and the development of new sites. Our acquisition in Hunt County, Texas, earlier this year is an example of how we are beginning to build this next generation of capacity. We have now identified a new site pipeline of more than two gigawatts of potential incremental power, with initial capacity potentially available from late 2028 through 2030.

This pipeline meaningfully expands our opportunity set. We will apply the same disciplined approach to advancing these projects that has guided the development of our existing platforms. We believe the late 2028 to 2030 timelines are well aligned with our construction schedules. Over the next several years, our primary focus will remain on executing against our contracted commitments and advancing the broader pathway towards 2.5 gigawatts with AMD. This longer-dated pipeline provides additional runway for growth beyond that opportunity. Our position is unique. We continue to have the balance sheet and operating experience to invest ahead of customer demand when the economics are compelling. The strategy we have outlined is a repeatable model that will guide our growth, secure power early, invest with discipline, deliver capacity at scale. Expand successful customer relationships. Today's AMD announcement is an important validation of the strategy we have pursued.

It is also a foundation for what comes next. We remain confident in the opportunity ahead and in our ability to continue building one of the most valuable infrastructure platforms serving the growth of AI. Before turning the call over to Matt, I would like to thank the entire Core Scientific team. Their expertise, commitment, and collective effort have brought us to this important inflection point and positioned us to continue delivering for our customers and shareholders. I could not be more excited about the next phase of Core Scientific and the opportunity that is ahead of us. With that, I will turn the call over to our Chief Operating Officer, Matt Brown, to discuss operations. Matt? Thank you, Adam. Today's announcement marks an exciting next phase of growth for Core Scientific and reflects the strong execution of our team.

I will begin with a major milestone achieved during the second quarter, then provide an overview of the AMD build-out and delivery plan. As we stated in our last earnings call, we expected to substantially complete four of the five CoreWeave campuses before the end of the summer. We achieved that milestone ahead of schedule, reaching 437 billable megawatts and demonstrating our ability to deliver complex infrastructure safely, efficiently, and at scale. Dalton Phase Two, the fifth and final campus, remains on track for full completion in early 2027. With the majority of the CoreWeave program now delivered, our focus is shifting to the next phase of development, led by our approximately 530-megawatt AMD commitment across five campuses.

Importantly, this program represents more than a collection of individual data center projects. It is an integrated colocation platform engineered to support AMD Helios rack scale systems optimized for the most demanding AI workloads. Our close-coupled AMD design framework aligns Core Scientific's infrastructure with AMD's technology roadmap through a repeatable, scalable design that accelerates speed to compute, optimizes capital deployment, and reduces execution risk as contracted capacity scales. Pecos remains our lead AMD site and is on track for initial megawatt delivery in the first half of 2027. Vertical construction is underway. Major infrastructure equipment is beginning to arrive on site, and the project continues to advance in line with delivery schedule. The remaining AMD campuses, Hunt, Aldern, Muskogee, and Dalton Phase Three, are also progressing through design, procurement, site preparation, and construction, with deliveries beginning in 2027 and ramping through the end of 2028.

As we move from one major customer program to the next, our operating priorities remain clear: deliver capacity efficiently, maintain schedule discipline, and deploy capital responsibly. We also want to give investors clear visibility in the true all-in cost of delivering high-density AI infrastructure. The cost per megawatt goes well beyond acquiring land and putting up the building. It reflects the total capital required to take a site from development planning through construction, utility energization, integrated systems testing, commissioning, customer acceptance, and ultimately, rent commencement. We organize that investment into three categories. The first is construction labor and on-site execution. Skilled electrical and mechanical technicians, pipe fitters, equipment operators, project supervisors, and safety personnel required to assemble, integrate, and commission the facility.

Second is the critical infrastructure equipment commonly referred to as OFE, owner furnished equipment, including transformers, switch gears, generators, chillers, pumps, liquid cooling systems, power distribution systems, control systems, and the other major components that high-density computing requires. The third is soft cost and general conditions, design engineering, permitting, utility interconnect, insurance, on-site construction offices, warehouses, temporary power, fuel, lighting, program management, testing, and contingency. Together, these categories capture the full cost of delivering a commissioned customer-ready billable megawatt. Based on current market conditions and site-specific factors, we expect build cost range from approximately $11 million-$12 million per megawatt. The key takeaway is simple. We have demonstrated that we can deliver AI infrastructure at scale and are now applying that experience to a standardized multi-site deployment platform for AMD with determined focus on schedule, capital discipline, and repeatable execution.

With that, I'll turn the call over to Chief Financial Officer Jim Nygaard to discuss Q2 financials.

Thanks, Matt. I'll begin with our second quarter results, which reflect continued momentum in scaling our high-density colocation business. We began billing for 437 megawatts in mid-July, nearly 200 megawatts more than at the end of the first quarter and ahead of expectations. This drove a significant sequential increase in GAAP colocation revenue to $137 million, and we expect another meaningful step-up in the third quarter. For context, under GAAP, revenue from the CoreWeave contracts is recognized on a straight line basis over the 12-year lease terms, effectively pulling future contractual escalators forward. Within Bitcoin mining, our strategy remains unchanged. We continue to optimize the business and operate it primarily to offset contractual power costs during the wind down. We ended June with nearly 30% fewer miners online than at the end of the first quarter and are now self-mining at only two sites.

We expect Bitcoin mining activity to continue winding down over the remainder of the year. On the expense side, second quarter cash SG&A was approximately $36 million. The $4 million sequential increase was primarily driven by one-time professional fees associated with our recent debt financing. While we are not providing explicit SG&A guidance, we continue to view the low $30 million range as a reasonable quarterly baseline, with the potential for some variability as we make targeted investments to support growth. With that overview of the quarter, let me turn to capital formation and our plans to fund the next phase of growth. The AMD announcement is a significant commercial achievement and an important validation of the investment and financing strategy we have pursued We ended the second quarter with approximately $1.8 billion of liquidity, giving us a strong foundation as we prepare to fund the AMD build-out.

At our current cost estimate of $11 million to $12 million per megawatt in CapEx, the initial 530 megawatts will require approximately $6 billion of capital, which we expect to finance through project-level bonds. Beyond our contracted commitments, we intend to continue to selectively advance capacity ahead of customer contracts following the same playbook that helped us position for the AMD opportunity. At a high level, we are prepared to invest up to approximately $1 billion to advance roughly 500 megawatts of initial build-outs for future capacity. This capital would be deployed to advance development, secure long lead equipment, and provide greater certainty around ready for service dates.

The AMD announcement demonstrates the value of this approach and reinforces our ability to convert customer demand into additional contracted capacity. We believe our balance sheet and financing strategy give us the flexibility to execute our contracted commitments while continuing to invest in the next phase of Core Scientific's growth. With that, I'll hand the call back to the operator for Q&A.

Thank you. We will now be conducting a question and answer session. If you would 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. Please limit yourself to one question and one follow-up. You may press star two if you would like to remove your question from the queue. For participants using speaker equipment, it may be necessary to pick up your handset before pressing the star keys. Our first question will come from John Todaro with Needham & Company.

Hey, thanks for taking my question and congrats on all the progress and the lease here. Two, if I may. First one, just on the potential expansion with AMD, looks like 1.5 gigawatts of that is, as stated, behind-the-meter and in the load study. I guess just wondering if we can get a little bit more color on kind of timeline there, and if it is a little bit more lengthy, does AMD then have an exclusivity period for a significant chunk of time or just maybe frame that up a little bit more for us?

Happy to, John, and thanks for the question. To start off, this is obviously a transformational deal for Core Scientific, and we could not be more excited about our partnership with AMD on this. It is not appropriate for us to speculate necessarily on what AMD will do as it relates to the additional megawatts. As you mentioned, it is up to nearly 2 gigawatts of additional capacity under the agreement, and that is an exclusive reservation agreement that we have with AMD. The important part here is, as you look at and as you mentioned, behind-the-meter. Behind-the-meter is becoming much more common in the marketplace amongst hyperscalers and labs.

That has continued to be a growing segment of this market, and I believe the comfort level amongst all of the peers are continuing to increase as it relates to behind-the-meter deployments. We have great solutions as it relates to our Pecos and Muskogee campuses. We are looking forward to growing those campuses alongside AMD. The important part here is the market demand for GPUs only continues to expand, obviously highlighted by the most recent AMD announcements. Data center capacity is still in very short supply. We feel like we are very well positioned for the continued growth here, and we are going to be able to deliver a significant amount of capacity for AMD.

Great. That is very helpful. Just a quick one on the Neocloud lease as well. Is there going to be a backstop or guarantee for them? Is AMD associated with that one as well or completely separate?

No, that is correct. They are providing a full credit support for the full 15-year lease term. As we noted in prepared remarks, that full credit support does not include any equity step-in rights as you have seen included in some other deals, which just protects our equity investments in these projects.

Our next question will come from Brett Knoblauch with Cantor Fitzgerald.

Hi, guys. Thank you for taking my question and congrats on the deal. Curious about the economics between maybe the two deals within the larger deal. Should we view that maybe the Neocloud plus backstop economics similar as maybe the straight AMD economics?

Yeah, I think that's right. I think what you could assume, given the modified lease or modified gross lease structure with the Neocloud, that those economics are in line with market. I would say similar for the direct leases with AMD. Those are also in line with market.

Awesome. Maybe just on the CapEx front. I know you guys have kind of been pre-spending already across multiple sites. Could you give me a ballpark maybe how much of the $11 million-$12 million you guys have already spent repairing some of these sites?

Similar to what we had mentioned in previous earnings call, we were looking at deploying about 2 million per megawatt across the portfolio. We had just about $1 billion in commitments across these projects. We're very far into our capital commitment as it relates to what the equity investment will be required for these projects.

Moving next to Darren Aftahi with Lucid Capital.

Hi, good morning. Congrats, guys. Two, if I may. Can you just talk to the thought process of if the deal is exclusive with AMD, just committing sort of one customer of that amount of capacity, just kind of the thought process that went through that, and then the timeframe you laid out in the presentation, just your level of confidence in delivering that capacity on time. Thanks. Thanks, Darren. I appreciate it.

As we look at the commitment that we made with AMD today, this is truly a one of one type partnership in this market. We feel very strongly that our execution capabilities over the course of the CoreWeave contracts, it was one of the main reasons why we are able to get such a unique partnership agreement with AMD. As we look at the product roadmap and the growth, the reservation rights that AMD has on additional capacity at sites where they have direct leases, that's Pecos, Muskogee, and Hunt. In terms of those delivery timelines and their reservation rights timelines, they pair up very well.

We feel very strongly that we're going to be able to continue to grow alongside of AMD at those three sites in particular, just given the fact that there's significant demand in the market. As I mentioned earlier, market drivers are all in our favor as it relates to both supply and demand here in this industry. We believe AMD is going to continue to expand into the future. Darren, to your other question, talking about our confidence and our ability to deliver, we have the equipment secured, we have contractors on site across these five campuses. This is a unique situation compared to others who are announcing deals with greenfield. We have bodies moving on site, walls going up. At Pecos, we mentioned we have the building fully complete at this point.

We feel like we're in a very strong position, and that confidence in our ability to deliver and the progress that we've made, we believe is really why AMD chose Core Scientific to partner with.

We'll go next to John Petrides with Jefferies.

Oh, great. Thank you. Congratulations on the AMD deal. That's really exciting. Looking at your slide on load study, additional power, Pecos, Texas. I think on August seventh, we're going to get an update from ERCOT. I guess it's 815 megawatts. Is there a potential that that's unlocked for you guys, or at least you have a timeline on it within the next few weeks?

It's really hard for us to judge. Part of that is about 300 megawatts of that is the load study for Pecos. The rest is behind-the-meter. I think in terms of our execution on the next megawatts at Pecos in particular, is going to be driven by the behind-the-meter strategy, just given the uncertainty related to timing of that next 300 megawatts from on-grid power.

Okay. All right. That's helpful. Then on the developments, maybe can you talk about the build-up to the lease, like the different pieces that you've already put in place, like deposits down on the various supply chain stuff that you need for the development, lining up General Contractors and subcontractors? I'd just be curious to hear just a little more about what you guys have been up to the past few months and just the pieces that you had to put together to get to the point of lease signing this morning.

Yeah, I can take that question. I think as we stated it in our previous earnings, we outlined a strategy of sort of leaning into development ahead of demand. Over the past year, we started developing Pecos, we started pre-construction in Hunt and the second building in Muskogee. We were already getting through pre-construction, through engineering. We had secured, actually placed orders for equipment for the initial delivery phases of each of those projects. What does that mean? Does that mean that we'd already secured long lead equipment for a large quantum of those megawatts, and then we had already secured the labor and the GCs on site, and sort of progressing through GMPs, and now we're at a stage where we've already completed, like in Pecos, the precast in the building is almost complete here in a number of weeks.

The full shell for the first 185 megawatts will be done. We're at pad-ready utility energization across the other sites. We released capital for substation construction earlier this year across multiple sites. All of that work that we've been going through over the course of 2026 has put us in position to really land a monumental deal with AMD, and we're really excited about our ability to execute through 2028.

Moving on to Nick Giles with B. Riley Securities. Yeah. Thanks very much.

Congrats, guys. You mentioned, I think, 2 gigawatts of new sites. Can you just break that down across how many sites, how advanced is due diligence, and then would you expect to spend capital at those sites ahead of any lease similar to your current footprint? Thanks. Yeah, I appreciate the question, Nick.

Yeah. Across the 2 gigawatts, we're not giving a site breakdown, number of sites. I would say the Hunt acquisition that we made earlier this year is extraordinarily representative of the opportunities that we're pursuing today. I would say that's a great strike zone in terms of total amount of power that's available at the site. To your last part of your question, absolutely. As we look at new sites and we look at acquisition costs, what we include in those calculations is being able to bring that site to really a pad-ready status at the very least. That includes putting the substation in place and releasing that capital. As we evaluate these sites, we think about how does this fit in terms of when the power is available versus our construction schedules.

There are a number of sites in that pipeline at varying stages of due diligence that we have confidence that we'll be able to bring a new site to market, hopefully by year-end here.

Thanks for that, Adam. Maybe a question for Jim. Should we expect to see you raise project debt at the site level? How do these AMD direct sites versus the Neoclouds with a wrapper influence the overall financing strategy?

Our primary financing strategy is going to be utilizing the project bond structure that is, I would say, fairly consistent in the market today, very similar to what we did with CoreWeave. Functionally speaking, the SPVs are very similar. In the direct case, the tenant has direct responsibility for fulfilling the lease payment, that's what effectively fuels that vehicle. On the credit support dynamic, the debt is fully supported. That dynamic still exists in a similar structure. They function and operate in a very similar fashion, but they do have distinctions of having the direct relationship in one and having a credit support feature in the other. They are project bond structures in the SPVs that we've now used in the CoreWeave example.

Our next question comes from Ben Sommers with BTIG.

Hey, good morning, guys, thanks for taking my question. As we think about the behind-the-meter opportunities, just kind of curious, what is the current kind of status of securing potentially long lead time items for this? I guess just, I know you can't give a direct timeline estimate, but just kind of curious how that development is progressing as we think about expanding with AMD or beyond.

Yeah. Thanks for the question. What I can say, our visibility into the behind-the-meter development across Pecos and Muskogee, we've already been in advanced conversations with natural gas suppliers. We've done preliminary planning around lateral development to those sites. We've had some really integrated conversations with equipment providers that would provide the generation on site for that. I would say where we stand today is that we have really clear visibility into the execution timelines, cost, and all the delivery partners required to pull that together.

Awesome. Thank you. Just one more quick one from me. For the expansion capacity, does AMD have the ability to potentially grant that to, let's say, like another Neocloud similar to how we're doing here and then backstop that contract? Does this all have to be direct with AMD?

On the direct leases, it has to be direct with AMD.

Our next question comes from Joseph Vafi with Canaccord.

Hey, guys. Adding my congratulations here as well. Great to see the AMD news. If we rewind about a quarter, I know you were winding down some exclusive negotiations for, I believe, some of these sites, with an investment-grade tenant. Now we have the AMD announcement. Be interesting if you could provide any color on if AMD was the exclusive negotiating partner there or if they arose after those exclusives ended. I think it'd be valuable and insightful relative to the cadence of negotiations out there broadly in the marketplace.

Yeah. I appreciate the question, Joe. Not going to comment on who the customer was in previous discussions. I think the key here is, this is a long-term relationship that was formed over a long period of time with AMD. They were evaluating our execution capabilities across the existing contracts that we have in place today. They were evaluating what we had on order in terms of long lead equipment, and we're evaluating the sites over a period of time as construction continued across the five campuses that they've signed up for today. These conversations are long, and I think that's expected across the market. What we signed today is truly unique and we couldn't be more excited about partnering with AMD on a project of this scale.

Sure. Great. That's helpful, Adam. Jim, you're kind of ahead of the pack here on generating revenue, and operating cash flow, I think, at this point. How does the revenue and more of a maturing P&L kind of shape strategy here versus where you were maybe six or nine months ago? Thanks. Yeah. I appreciate the question.

It is quite a transformation from our history of Bitcoin mining, which is certainly characteristic of a lot of volatility and lack of transparency to essentially a financial profile that is essentially opposite of that. That's what makes this business, in terms of its financing capability, its visibility, and its ultimately stability, allows us to really make investments with a lot of confidence. We're excited about that transformation. We've had a lot of noise in our historical financials, and what's exciting about next year is we're going to be starting with a clean sheet of paper, and you're going to start to see a much more mature financial profile emerge that is much more consistent with the new business model.

That's an exciting transformation for us and one that we think is going to serve us well. I appreciate your comments about us being ahead of the pack. That's an important observation for our differentiation in the market that often I don't think we get a lot of credit for. Thank you for calling that out.

We'll hear next from Stephen Glagola with KBW.

Hey, congrats on the deal. Thanks for the question. Adam, I'm curious to get your broader thoughts on what you're seeing in the funding markets today on the debt side. Has anything changed in terms of project financing availability over the last few months? Thank you. Thanks, Stephen. Obviously, we're excited about our announcement this morning.

It came at a very apropos time in terms of just the broader market, and recent reports that have been released. I think, as it relates to funding market, as you mentioned, what we're seeing in the market today is definitely a backup in rates. I think that's a broad digestion period as it relates to AI. What we've seen across all of the bonds that are in the market today are just a significant amount of digestion by investors that are speaking with their trades and where the market has been going to in terms of rates. It doesn't concern us with such a strong investment-grade counterparty here.

There will still and continue to be appetite for those types of bonds in the market. That's really what gave us confidence here in executing this contract. I think if you were signing a contract with anyone outside of this tier of credit, there's question marks about capital raising. Given where we sit today with our partnership with AMD, there's incredibly high confidence as it relates to financing this deal.

Great. Thank you. Our next question comes from Jon Hickman with Ladenburg Thalmann.

Hello. Could you just reiterate your timeline for the first delivery of power for the AMD deal? Is it early 2028? The initial AMD deal, the first megawatts to come online will be at our Pecos location.

We've said that that will come online in the first half of 2027. Broadly speaking, we've said about half the contract will be delivered in 2027, and the other half will be delivered in 2028.

Thank you for that. I appreciate it.

Of course. Yep. Moving next to Tim Horan with Oppenheimer.

Thanks, guys. Is there an optimal amount of megawatts you'd like to build per year, and is there kind of an upper limit on that? Just on the behind-the-meter power, what's really on the critical path there? Is it the pipelines? Is it the turbines? Anything else? Thanks. Yeah, absolutely.

I'll take the first part of the question, and I'll let Matt Brown take the second part. In terms of optimal megawatts per year, in terms of what we're looking at in 2028, we believe we could have a target of about 600 megawatts in 2028 based on current labor constraints and long-lead equipment constraints in the market. Obviously, that is dependent on timelines of signing additional capacity under the AMD agreement. I think that's a great target for us. Plus or minus 600 megawatts per year is a great cadence for the business. That's not to say, though, if the opportunity arises to deliver more than that in a single year, doesn't mean we won't continue to scale up our capabilities and team internally, to really meet that requirement. From where we sit today, 2028, we have a target of about 600 megawatts of delivery.

Matt, would you like to take the question?

Yeah. The constraints vary by site. I would say largely, delivery timelines, the constraints are either going to be the time to build the lateral pipeline, and/or the equipment delivery, and it just depends on the various sites. Some of our sites, the delivery timelines will be more tied to equip manufacturing and production capacity. In other locations, it's going to be more tied to pipeline development and delivery. Those are the two main things that sort of drive schedules with that.

Thanks so much. We'll go next to Paul Golding with Macquarie Capital.

Thanks so much for taking the question. Congrats on the deal. I wanted to ask on the CapEx, as noted in the slide, estimating $11 million-$12 million per megawatt, also as implied by the $6 billion across the 530 megawatts. I was wondering what's driving the higher CapEx versus the CoreWeave deal at the outset. Is that the greenfield versus brownfield? Is there a difference in basis of design? What we should expect going forward around cost per megawatt, if you do incremental deals with other counterparties. Secondly, just wanted to ask around the acceleration of the 437 megawatts delivered for CoreWeave. What unlocked that acceleration? Is that a lever that you can pull additionally with the AMD engagements? Is there breathing room there? Thanks so much. I'll take the first part of this question, sort of relating to cost and as it relates to where we're at today with those current $11 million-$12 million estimates on a portfolio basis.

When we think about the CoreWeave sites, we started that engagement in early 2024, so we started securing labor and equipment in 2024. What we've seen pretty much year-over-year is pretty much increases across the board, both increases in equipment cost, some of that driven by tariffs over the past year, some of it just driven by supply chain constraints and availability. The thing that's primarily driving, I would say, cost of construction today is labor. Labor is very scarce in a number of markets across the U.S.

A lot of the GC, a lot of the electrical subcontractors, the large ones, and a lot of the mechanical trades are just pretty saturated with work right now. What you're seeing is just that labor constraint is naturally sort of driving up the cost of that labor in some of the very, very competitive geographies across the U.S. I think that's one of the biggest differences of just the increase of cost over time. As it relates to some of the CoreWeave sites, there are just some fundamental design differences across a number of those sites that are just different to what we're doing today. Not to sort of drive into all the details, but there are some fundamental differences between what we've done with CoreWeave and the current product set.

Our next question will come from Brian Dobson with Clear Street.

Hey, good morning. Thanks so much for taking my question. Congratulations on the deal. Now that you have that signed and announced, do you think you could give us a little bit of additional color on what the demand environment looks like? I'm sure you were speaking to a lot of interested parties and maybe what we could come to expect from, call it sector deal signings over the next 6 months. What's your viewpoint here and how is demand evolving?

Yeah, Brian, appreciate the question. I think as we look at the demand picture really for developers that are having direct conversations with the counterparties that can sign contracts right now, it's really starkly contrasted against what you're seeing in the media headlines. What we're seeing on the ground is still a significant amount of demand coming out of the hyperscale channel and the AI labs. I think in terms of what we're going to see over the coming months and through the remainder of this year are continued new deal announcements as it relates to large scale infrastructure commitments. I think the big part here is a lot of these large scale GPU contracts are just getting signed today, and those GPUs do not have a home yet.

I think in terms of what we've seen over the course of the past 12 months and late deliveries across this industry, there's a lot of GPUs sitting on the ground, and those GPUs still need to be plugged in. As that backup continues to build, I think what we're going to end up seeing is more constraints across data center supply. Obviously given, I would say some of the headwinds as it relates to new developments of data center capacity across this industry, having available rack space within the next few years is going to be in high demand for all data center developers. This is a great tailwind for the DCs.

I feel very strongly that we're going to continue to see new contracts being signed over the remainder of this year just because this demand picture and supply are shaping up for continued new leases to be signed.

Great, thanks. Appreciate your candor.

We'll go next to George Sutton with Craig-Hallum.

Thank you, and congratulations. I'm curious, when we're looking at the reservation opportunity with AMD, how are we continuing to have discussions with additional parties? How do you keep that mix live?

Yeah, I appreciate the question, George. Under the agreement that we have with AMD, this is a partnership, we won't be speaking to customers while the additional capacity is under the reservation agreement. We're going to work very closely with AMD on their future demand needs, we're going to look to continue to grow alongside of them. This is a true and unique partnership in this industry, I think it's going to take time for people to understand truly how transformational this is and the unique position that Core Scientific sits in within the broader landscape. For us, we just look forward to continuing to grow alongside of AMD and look forward to developments across Pecos, Muskogee, and Hunt.

Just real quickly on the Neocloud customer, obviously Neoclouds vary greatly. Can you give us any sense on the size and scope of this Neocloud? I assume the contracts that support this are already in place.

Yeah, George, it's unfortunately something that we can't comment on at this time. I think the key here to the contract that we signed with the Neocloud, is that we do a full 15-year credit support agreement with AMD, with them standing behind the credit here. We feel very good about that transaction and those two leases that were signed across Dalton and Auburn.

Moving next to Michael Donovan with Compass Point.

Hi, guys. Thanks for taking my questions in. Congrats on progress. Following up on behind-the-meter questions, are you solely looking at turbines or also evaluating fuel cells?

Yeah, we're keeping our options open. The selection of what technology we might go with will likely vary from one location to the next. Everything from REPS to linear generators and fuel cells, we're considering all of those. Partially what will drive that decision will be kind of the local geography itself, and the environmental conditions at a site that may drive that. The second piece will just be supply chain availability, and that. We certainly have talked to a number of vendors across each of those equipment types.

Understood. On the modified gross lease associated with the Neocloud, how should we think about NOI margins?

You should think about those NOI margins as being relatively consistent with the market standards that have been disclosed across other deals that have been announced. We think we are right in line with that.

Moving on to Andrew Beal with ARETE Research.

Oh, hi. Could you just talk about the reservation right for incremental capacity? Does AMD have the right to sign up the phase one prices plus the 2.5% in escalator for whenever that right happens? Or is there another mechanism to get to a market price for the next signings?

The one thing we can comment on as it relates to the reservation capacity, is these would be direct leases with AMD as it relates to any future capacity that's signed under that agreement.

Right. Are they negotiated at the time or are they preset in terms of price?

They would be under substantially similar terms as the existing leases that are signed today.

Moving next to Paul Meeks with Freedom Capital.

Good morning, everybody. A lot of good news here today. Just so I'm crystal clear, when will you deliver and bill that last 150 MW for CoreWeave?

Thanks for the question. The last 150 MW is our Dalton phase 2 campus. That will begin delivering at the end of this year and be completed in early 2027.

Okay. My follow-on is, you talk about the maturity of your model. Now you have a contract visibility, more diverse customer base, Neocloud. If I take a look at your adjusted EBITDA margins, because on your P&L, you do show that metric, what will it look like with your business mix and how it develops at the end of 2028?

Thank you for the question. We have not provided specific EBITDA targets for the business at this point. We will certainly evolve that thinking as we get closer to a cleaner set of financial statements. We have been working our way through the mining dynamics of our business, and we expect that to be an even smaller portion as we approach year-end to start 2027 with a clean year. We have not provided specific SG&A guidance, although in my commentary today, we talked about that quarterly baseline of about $30 million a year, and we have, of course, disclosed the revenue numbers behind the CoreWeave contract. EBITDA margins in this business, we track very closely a cash perspective.

There's a GAAP dynamic in lease accounting that has escalators in the top line a bit early that inflate the GAAP margins on that metric, but we track it on a cash basis, and we feel very confident that this is a much higher EBITDA margin that's quite attractive and candidly has quite a bit of leverage on the operating expense line. The business that we've announced here today will only add to that financial profile and there's only upside to that margin profile going forward.

That's all the time we have for questions today. Ladies and gentlemen, thank you for your participation. This does conclude today's teleconference. You may disconnect your lines and have a wonderful day.