CIFR - NASDAQ

Good morning, and thank you for joining us on this conference call to address Cipher Mining's business update for the second quarter of 2025. Joining me on this call today are Tyler Page, Chief Executive Officer; and Edward Farrell, Chief Financial Officer. Please note that our press release and presentation can be found on the Investor Relations section of the company's website, where this conference call will also be simultaneously webcast. Please also note that this conference call is the property of Cipher Mining and any taping or other reproduction is expressly prohibited without prior consent. Before we start, I'd like to remind you that the following discussion as well as our press release and presentation contain forward-looking statements. These statements include, but are not limited to, Cipher's financial outlook, business plans and objectives and other future events and developments, including statements about the market potential of our business operations, potential competition and our goals and strategies. Forward-looking statements and risks in this conference call, including responses to your questions, are based on current expectations as of today, and Cipher assumes no obligation to update or revise them, whether as a result of new developments or otherwise, except as required by law. Additionally, the following discussion may contain non-GAAP financial measures. We may use non-GAAP measures to describe the way in which we manage and operate our business. We reconcile non-GAAP measures to the most directly comparable GAAP measures, and you are encouraged to examine those reconciliations, which are filed at the end of our earnings release issued earlier this morning. I will now turn the call over to our CEO, Tyler Page. Tyler?

Thanks, Courtney. Good morning, everyone, and thank you for joining us today. I'm Tyler Page, CEO of Cipher Mining, and I'm pleased to welcome you to our second quarter 2025 business update call. In the second quarter, Cipher delivered on all previously outlined plans at Black Pearl Phase 1 and importantly, exceeded our prior growth guidance. Building on that momentum and now supported by even stronger cash flows, we've made several strategic decisions to continue to position the company for the evolution of the data center landscape. I'll provide more details on those decisions later in the call. But first, I would like to highlight a few key metrics that speak to today's themes: consistent execution, thoughtful investment and purposeful growth. In the second quarter, we grew our Bitcoin holdings from 1,034 to 1,063 Bitcoin. In addition to growing our inventory, we paid off all short-term borrowings, which Ed will discuss later on the call. We are incredibly proud to have energized and commenced Bitcoin mining at our Black Pearl data center ahead of schedule this quarter. As a reminder, we plugged in legacy rigs from our Odessa upgrade at our new site, while we await the latest generation rigs scheduled to arrive in batches before the end of the third quarter. Our initial projections targeted 16 exahash per second by the end of the second quarter and 23.1 exahash per second by the end of the third quarter. We exceeded that guidance, finishing the second quarter at 16.8 exahash per second. Today, we are pleased to report that we are on track to once again outpace expectations from previous guidance and reach 23.5 exahash per second by the end of the third quarter. All of our fully funded latest generation Bitmain rigs are scheduled for delivery to the site in batches by the end of the third quarter. Also, in the second quarter, we executed an order with Kenan to purchase a small batch of miners. This new order, coupled with our Bitmain order, will allow us to achieve this growth while also improving efficiency. At the end of the second quarter, with legacy rigs plugged in at Black Pearl, our fleet efficiency stood at 20.8 joules per terahash. Once our new fleet is fully deployed, we expect efficiency to improve to an impressive 16.8 joules per terahash, making us once again among the most efficient miners in the industry. While our growth is impressive, I'd like to take a moment to briefly revisit a few of our core metrics and foundational strengths. Our projected all-in weighted average power cost remains highly competitive at just $0.031 per kilowatt hour. The slight increase can be attributed to Black Pearl Phase 1, a 150-megawatt front-of-the-meter site coming online. Given our deep power expertise and the ability to dynamically curtail our data centers, including during the summer 4CP months, we expect to be able to maintain this low cost of power even at our front-of-the-meter sites. Our proprietary software has proven to be a critical advantage in optimizing for profitability, maintaining low power prices and monetizing older rigs. Across our sites that operated for the entire second quarter, including Odessa, Alborz, Bear and Chief, we paid an average all-in electricity cost of approximately $27,324 per Bitcoin produced at our data centers. This is a highly competitive figure as it reflects the total cost to deliver electricity to our mining rigs, including all taxes, transmission and other charges. To summarize, we will soon be running some of the industry's lowest cost power through an even more efficient fleet, a powerful combination that will continue to drive exceptional unit economics. This strong operational foundation will remain a key advantage as we continue to grow and scale. With the addition of Black Pearl Phase 1, Cipher's current operating capacity stands at 477 megawatts with potential pipeline capacity expansion of up to 2.6 gigawatts in the coming years. We'll discuss our impressive pipeline in further depth later on the call, so I'd like to shift focus to our second quarter growth highlights across key verticals and our outlook for the path ahead. During the second quarter, we delivered on Phase 1 of Black Pearl, the first 150 megawatts of our 300-megawatt data center, which is now mining at 6.9 exahash per second and on track to exceed prior guidance by the end of the third quarter. As previously discussed, we will fill Phase 1 of Black Pearl with a majority of Bitmain S21 XPs, supplemented by our new purchase of approximately 1.5 exahash per second of Canon A15 Pros. Importantly, both of these orders were fully funded through a combination of balance sheet cash and proceeds from our first convertible senior notes offering. As we've discussed in the past, we are focused on thoughtfully evaluating funding options that support growth while minimizing dilution. Through this successful convertible offering in the second quarter, we raised approximately $168 million in net proceeds. With this convertible offering, we were able to take advantage of favorable market conditions, expand our institutional investor base and lower our cost of capital while completing Black Pearl Phase 1 in a capital-efficient manner. We are also participating in ERCOT's ancillary services market at Black Pearl. This participation not only supports grid stability and reliability, but also creates an additional revenue stream for the site with minimal impact to uptime. Our participation is seamlessly integrated into our proprietary software, and we expect to scale our participation in the coming quarters. As we have made thoughtful investments to position our Bitcoin mining operations for the future, we have also been extremely focused on investing strategically to ensure the company is well prepared for the broader evolution of our industry on the HPC side. Given Barber Lake's 300 megawatts of energized capacity and ideal site characteristics, we believe Barber Lake is one of the most compelling HPC development opportunities in the U.S. today. Market dynamics and ongoing dialogue with potential tenants continue to confirm both a growing demand for power and a tightening supply. Against this backdrop, we're more confident than ever in the site's long-term value and are actively engaged in advanced discussions to secure the best possible deal for Cipher and our shareholders. We're encouraged by recent momentum and look forward to sharing further updates as progress continues. Let's now turn to Black Pearl Phase 2, which presents another compelling HPC opportunity. As a reminder, Black Pearl is our 300- megawatt data center in Wink, Texas that energized this quarter. Phase 1, the first 150 megawatts is currently hashing. Phase 2 is the second 150 megawatts of powered land. Over the past few quarters, we've thoughtfully evaluated various paths for Phase 2 with a focus on driving the best possible outcome for our shareholders. Today, we're excited to share a new strategic plan for the site, one that positions Cipher to capitalize on near-term Bitcoin economics if desired, while maintaining the flexibility to transition to HPC as demand for every available megawatt continues to rapidly accelerate. We're excited to announce today that we are moving forward with the construction of Black Pearl Phase 2, which envisions 150 megawatts of infrastructure that is designed to support both hydro Bitcoin mining and HPC compute applications simultaneously. This infrastructure will be built from inception to be able to seamlessly convert to Tier 1, 2 or 3 utilizations in response to tenant demand. Before I dive into specifics, I want to first outline a few guiding principles that shaped our decision to move forward with a flexible build-out for Bitcoin mining or HPC workloads. We firmly believe that energy availability is rapidly becoming the defining constraint for the future of HPC compute. This view is increasingly echoed by industry leaders. During a Senate hearing in May, Sam Altman remarked, the cost of AI will converge to the cost of energy. The abundance of it will be limited by the abundance of energy. In July, Anthropic released a policy report titled Build AI in America, highlighting the growing energy demands of advanced AI systems. The report stated, as AI systems grow more capable, the energy and computational requirements to train and deploy Frontier AI are surging and cited estimates suggesting that the U.S. AI sector is on track to require at least 50 gigawatts of electrical capacity by 2028. Notably, the report stated that training workloads alone are expected to account for 20 to 25 gigawatts of that demand. These statements and many others like them highlight a rapidly emerging paradigm driven by larger models, larger data sets and exponentially increasing compute requirements, all of which are fundamentally dependent on an ever-growing need for energy. Beyond simply articulating the need for energy, hyperscalers are reinforcing it through their capital plans. Just 2 weeks ago, Alphabet shared that it now expects to invest approximately $85 billion in CapEx in 2025 with a further increase in CapEx in 2026 due to the demand it is seen from customers. Similarly, Meta recently said it expects 2025 capital expenditures to land between $66 billion and $72 billion and anticipates another year of similarly significant CapEx dollar growth in 2026 as it aggressively pursues opportunities to bring additional capacity online to meet the needs of AI efforts. Against this backdrop, we reaffirm our thesis that near-term power is a scarce and strategic resource and that large-scale interconnections available in the next few years are exceedingly valuable assets. For these reasons, we are aggressively positioning the company to take advantage of anticipated demand and ensure we're fully prepared to support tenants when inevitably every available megawatt is absorbed by HPC workloads. Black Pearl Phase 2 will be built around a proprietary basis of design that bridges the needs of both hydro Bitcoin mining and AI compute. This approach will align with current industry standards while maintaining the flexibility to support evolving hardware requirements such as higher density racks, variable inlet temperatures and diverse cooling preferences to ensure seamless integration as technology continues to advance. Through upfront design and infrastructure prepositioning, we will be able to efficiently convert to Tier 1, 2 or 3 design specifications in response to tenant leasing preferences or Cipher's evolving proprietary compute applications. This conversion scope is designed to minimize disruption to multi-tenant operations, allowing us to convert easily in segments over time. It is also designed to maximize speed to market with current design and supply chain plans enabling a turnaround for conversions in less than 6 months. While we expect the site to be fully leased by HPC tenants over time, this infrastructure plan preserves our flexibility to utilize the space for mining near term, if preferred. In short, Black Pearl Phase 2's infrastructure will enable us to monetize power immediately and maintain seamless and expeditious optionality to pivot as the HPC market continues to scale into every available megawatt. This innovative and industry-first build-out aims to position Cipher ahead of the curve, anticipating where the industry is heading and aligning our strategy accordingly. Let's now turn to a review of current operations, which continue to serve as a strong foundation for our success. Slide 7 has a production summary for our Odessa facility. Odessa is still the most significant part of our portfolio, representing approximately 85% of our Bitcoin production in Q2, although this will likely decrease significantly as production at Black Pearl Phase 1 ramps up. As of June, the current operating hash rate at the site is approximately 11.3 exahash per second using approximately 207 megawatts. Odessa's fleet efficiency stands at approximately 17.6 joules per terahash. On this page, we also provide the observed all-in electricity cost per bitcoin at the site over the quarter, which was roughly $24,686. As a reminder, Odessa is a wholly owned facility operating under a 5-year fixed price power purchase agreement, securing some of the most competitive electricity rates in the industry and reinforcing our cost advantage and operational strength. On Slide 8, we provide a combined overview of our joint venture data centers of Alborz, Bear and Chief. The 3 sites have a total power capacity of 120 megawatts and can generate approximately 4.4 exahash per second. We own 49% of the JV sites, and our portion recently generated roughly 13% of our overall Bitcoin production in the second quarter. On this page, we also provide the combined all-in electricity cost per Bitcoin at the 3 sites in the second quarter, which was roughly $44,594. As a reminder, both Bear and Chief operate as front of the meter sites, so there are expected seasonal fluctuations with their electricity costs. It is worth noting that Black Pearl Phase 1 is not included in this section as it was not operational for the full quarter and therefore, contributed only 2% of total production in the second quarter. However, the site has since ramped meaningfully, accounting for approximately 24% of production in July and is expected to continue growing its contribution going forward. We look forward to showcasing Black Pearl's full impact in future updates as the site scales. Let's now shift to an update on our development portfolio. We've organized the pipeline into near-term growth opportunities at Black Pearl and Barber Lake and longer-term expansion in 2026 and beyond. At Black Pearl in Wink, Texas, the first 150-megawatt phase is now live and hashing and will continue to scale as new rigs are delivered and old rigs are swapped out. The second 150-megawatt development phase will soon be under construction and will be designed to offer future optionality for HPC hosting or Bitcoin mining. The photos here highlight the best-in-class infrastructure at Phase 1, purpose-built for Bitcoin mining and showcase the team's outstanding execution in delivering the site safely and rapidly in just 16 months. We're confident that this same level of excellence, innovation and dedication will drive the successful build-out of Phase 2. Slide 11 gives an overview of our Barber Lake site. With its immediately available capacity of 300 megawatts and 587 acres of surrounding land, along with the potential for future expansion, we are more confident than ever in the commercial potential at the site. Complex HPC lease deals take time to come together, but the team has been working diligently to evaluate the various proposals we have received and finalize the best possible deal for Cipher. Slide 12 outlines our expected growth in 2026 and highlights our latest site acquisition in Andrews County, Texas called Stingray. The site features 100 megawatts of front-of-the-meter capacity, all necessary regulatory approvals and 250 acres of land adjacent to the transmission assets. This quarter, we initiated development of the substation for the site and secured long lead time items, including transformers and high-voltage breakers. The site is on track to energize in the third quarter of 2026. Slide 13 outlines our expected future growth across 4 sites with 1.6 gigawatts of potential power capacity. Our Reveille site in Katula, Texas is on track to energize in Q2 2027. The site is fully approved for 70 megawatts, and we have initiated development of the substation. Our 3 Ms, Mikeska, Milsing and McLennan are currently undergoing final interconnection approval processes, including the completion of load studies at all 3 sites with decisions expected later this year. We are targeting up to 500 megawatts of capacity at each of these sites. In addition to interconnection rights, our purchase options also include significant land parcels at each location, all of which are well suited for HPC data center development. These sites are located further east than our current portfolio and are positioned closer to major metropolitan areas. We've already seen early interest from potential tenants and believe these sites will be in high demand as development progresses. Backed by our 2.6 gigawatt pipeline, a strong track record of consistent execution and a focused strategy to position the company for the future, we are well equipped to become a leading developer of HPC data centers while continuing to set the standard in Bitcoin mining. I will now turn it over to our CFO, Ed Farrell, for a review of our second quarter financials.

Thank you, Tyler, and hello to everyone on the call. I'd like to remind everyone that today, I will be discussing our performance for the second quarter of 2025, which ended on June 30. Before I dive into our financial results, I'd like to highlight this quarter was not only one of strong execution, but also a disciplined capital deployment to achieve that execution. In the second quarter, we successfully leveraged our first-ever convertible offering to complete Black Pearl Phase 1 in a thoughtful manner. We strategically sized this transaction in parallel with commercial discussions with our rig provider to ensure we had a direct use of the proceeds that would be accretive to the company. Of the $172.5 million of gross proceeds from the offering, $108 million was used primarily to purchase the latest generation of miners to be installed at Black Pearl. By paying upfront, we were able to negotiate for an expedited rig delivery schedule to avoid broader tariff impact. In addition, we received a 10% reduction in our outstanding obligation on the rigs as well as incremental value from Bitcoin linked call options. We now expect the full order to be delivered in hashing by the end of the third quarter. We look forward to seeing the full impact from these new rigs captured in production and revenue numbers next quarter, and we are proud to have achieved this growth in a capital- efficient manner. Let's now turn to a review of our financials, beginning with our sequential and year-over-year financial performance outlined on Slides 15 and 16. In the second quarter, we reported $44 million in revenue, down 10% from $49 million in the first quarter of this year. While the late June energization of Black Pearl contributed positively to revenue, rising network hash rate and rising summer power prices in Texas, which increased curtailment led to a slight dip in top line. Increased curtailment allowed the company to avoid CP penalties and maintain its positions of having some of the lowest power costs in the industry. In addition, insights from earlier summer months will inform further refinements to the curtailment model for the remainder of the summer. Moving down the slide. For the quarter, we reported a GAAP net loss of $46 million or a net loss of $0.12 per share. Bottom line results were largely impacted by a decrease in fair value of our power purchase agreement at Odessa. The expected but substantial fluctuations in fair value are influenced by prevailing forward power prices and the time value for the remaining term of the contract ending July 2027. As I've discussed previously, the true value of the contract lies in its provision of low-cost fixed price power at our Odessa site. Excluding that and other noncash expenses, such as the impact of depreciation and amortization, deferred income tax expense, the change in fair value of the warrant liability, share-based compensation and nonrecurring losses, we reported second quarter adjusted earnings of $30 million or $0.08 per share, up roughly 400% from $6 million last quarter. Moving to Slide 16. In the second quarter of 2025, we achieved $7 million increase in revenue compared to the second quarter of 2024. This result reflects the upgrade to our Odessa rigs in Q4 of last year as well as the production increase from bringing Black Pearl online at the end of the quarter, which partially offset the pre-halving portion of this quarter last year. We continue to take pride in our ability to navigate the post-halving environment driven by a low-cost power and operational efficiency. As noted earlier, due to the expected fluctuations in the value of our power purchase agreement at Odessa, we saw a decrease in GAAP earnings year-over-year. This quarter, we reported a GAAP net loss of $46 million or a loss of $0.12 per share compared to a net loss of $15 million or $0.05 per share in the second quarter of the prior year. In addition to PPA fluctuations, we also changed our depreciation schedule for the useful lives of miners from 5 years to 3 years in Q2 of last year. That, plus the increase in the number of miners we are now depreciating given the upgrade at Odessa meant this quarter's GAAP earnings are a significantly higher impact from depreciation than the same quarter of the prior year. Our adjusted earnings, which excludes these noncash expenses, increased significantly from a loss of $3 million or $0.01 per share to earnings of $30 million or $0.08 per share. Let's move on to Slide 17 and take a deeper look at the results of our operations. For the quarter, we mined 434 Bitcoin at Odessa and 10 Bitcoin at Black Pearl, bringing our production to 444 Bitcoin mined in total across our wholly owned sites. This production generated $44 million in revenue at an average price of roughly $99,700 per Bitcoin. This compares to 524 Bitcoin mined in Q1 2025 at an average price of $93,500 per Bitcoin, resulting in $49 million in revenue. While we began to see some contribution from Black Pearl this quarter, equivalent to about $1 million or 2% of quarterly revenue, it is important to note that the site came online near the end of the quarter. As a result, we expect significant greater impact next quarter when we'll benefit from a full quarter of production. Our consistent low-cost power remains a critical factor in maintaining attractive unit economics and our cost of revenue remained relatively flat quarter-over-quarter and year-over-year. Now let's shift our focus to operating expenses. Compensation and benefits decreased year-over-year, reflecting the efficiencies gained through scale without the need for significant changes to staffing. The lack of significant staffing changes aligns with our continued confidence that we have the right team in place to support our ongoing growth. General and administrative expenses, which include IT, corporate insurance, professional fees, occupancy and other public company costs remained flat quarter-over-quarter and year-over-year. For the quarter, depreciation and amortization expense totaled around $44 million, representing a 2% increase from the prior quarter and 120% increase year-over-year. Year-over-year increases are primarily driven by our Q4 upgrade at Odessa, which brought online over 36,000 new mining rigs. As discussed previously, we also changed our accounting policy in Q2 of last year for mining rig depreciation, decreasing the estimated useful life of miners from 5 years to 3 years. This change, coupled with the increased number of mining rigs in operation explains the significant year-over-year increase. The oldest rigs in our fleet will cease depreciating in the fourth quarter of this year, but I'd like to note that even when these older rigs are fully depreciated, they can still generate strong returns when deployed strategically. This was demonstrated at Black Pearl Phase 1, where we profitably deployed old rigs while awaiting new miner deliveries. Moving down the page, equity and losses of equity investees for the quarter was $2 million, down $5 million from the previous quarter. In Q2, we recognized a $17 million unrealized gain on the fair value of Bitcoin inventory compared to an unrealized loss of $20 million in Q1. Unrealized gains and losses on Bitcoin reflect the mark-to-market accounting of our holdings. The spot price of Bitcoin at quarter end increased compared to the end of Q1 as did our Bitcoin and treasury, which ended the second quarter at 1,046 Bitcoin held. Now let's turn to our non-GAAP measure slide, where we reconcile our adjusted earnings. When adjusting our second quarter GAAP net loss of $46 million, we added back $76 million to the items listed, resulted in adjusted net earnings of $30 million for the quarter. This compares to an adjusted net gain of $6 million in the previous quarter and a loss of $3 million in the previous year. Now let's turn our attention to the balance sheet. On Slide 19, our total current assets at quarter end were $220 million. A large portion of that balance was Bitcoin held totaling $112 million. In addition, our cash position increased from $23 million in March to $63 million in June, an increase of $40 million from the previous quarter. The increase in cash primarily reflects the remaining proceeds from our convertible offering, along with opportunistic Bitcoin sales. As we discussed in depth on our last earnings call, we actively manage our treasury, either selling or holding every Bitcoin mine and remain disciplined in our approach to capital management. I'll quickly cover some additional balance sheet line items as of June 30. Our prepaid expenses and other current assets amounted to $7 million. This balance primarily reflects call options granted to us by our rig provider as part of the negotiated transaction to lower the cost of the latest generation miners being installed at Black Pearl Phase 1 and to mitigate tariff exposure. As discussed, our Odessa power contract is reflected as a derivative asset on our balance sheet. Over the quarter, we had a mark-to- market loss of our PPA driven by decreases in time value of the contract as well as declines in the forward curve for power prices. The PPA ended the quarter valued at $78 million, down from $93 million prior quarter. While there are expected substantial fluctuations in the reported value quarter-over-quarter, the true value of this contract lies in its provision of low-cost fixed price power at our Odessa site. Other significant assets include property and equipment totaling $474 million. Deposits on equipment of $183 million primarily consist of deposits for mining rig purchases. As a reminder, our Bitmain and Canaan rig orders are fully funded, and we expect to have all new machines delivered to hashing by the end of this quarter. We look forward to seeing the production benefit of this investment. Additionally, we hold intangible assets totaling $10 million with $7 million attributable to ERCOT approval at Black Pearl and $3 million related to capitalized software. These balances are net of $1 million in accumulated amortization. At the end of the second quarter, our equity investee interest in Alborz, Bear and Chief JV stands at $46 million, and we had operating lease assets of $12 million. We had security deposits totaling $14 million, down from $20 million last quarter due to our receipt of the Barber Lake Encore deposits. The remaining balance primarily represents Encore deposits for construction of interconnects at various data centers in our pipeline. Current liabilities declined significantly quarter-over-quarter, decreasing from $139 million in Q1 to $53 million at the end of Q2, a 62% reduction. This was primarily driven by lower construction-related liabilities following the completion of Black Pearl as well as decrease in accrued expenses tied to the final payment submission for our previous rig order. Our accounts payable decreased to $15 million from $30 million, and our accrued expenses decreased to $30 million from $66 million. Lastly and importantly, I want to highlight this quarter, we reduced our short-term borrowings from $35 million to 0. We've always taken a disciplined approach to treasury management, ensuring we're well prepared for any capital needs. At times, that includes borrowing against our Bitcoin holdings to provide near-term liquidity while preserving inventory or selling production forward. However, given the strength of our current cash position, we made the decision this quarter to fully pay down our short-term debt and fulfill our deliverable forward obligation fully. We're proud to have no remaining near-term obligations on the balance sheet, and we will continue to manage our treasury with discipline and flexibility. Before we conclude, I'd like to thank everyone for joining today's call. We're proud of the progress we've made this quarter and the growth achieved, and we remain focused on disciplined execution and capital efficiency going forward. Thank you for your continued support, and we look forward to updating you on our progress next quarter. At this time, I will pause and Tyler and I would be pleased to take your questions.[ id="-1" name="Operator" /> Our first question comes from Bill Papanastasiou with KBW.

I was hoping we could first chat on the Phase 2 strategic plan that was mentioned earlier in the call. Are you able to quantify the time to convert from hydro Bitcoin mining to AI HPC compute? And maybe just for comparative purposes, how would that compare if you were to start from ground zero today going from nothing to essentially AI HPC compute?

Sure. Thanks, Bill. So I think of Phase 2 at Black Pearl as almost being like a Tier 1 half data center. So a little bit, let's call it, upscale from a typical Bitcoin mining build, but not a fully kitted out ready to do HPC data center day 1. To get the whole 150 megawatts there, we expect that will be ready back half of next year. The idea being here that we'll be in a position to have Bitcoin mining, and there could be megawatts that are available, that would be completion of it, but there could be megawatts available in advance of that, but call it, back half of next year. In our discussions with potential tenants, we have heard a few times that the challenge that some of them face, if I think of some of the neo clouds in particular, that they get a client that wants to deploy reasonably quickly. We're pretty confident that in the next 50 years, this data center will end up in HPC data center. The goal here is to have a flexible design that's ready to accommodate fast requests. And by fast, meaning in a couple of months, the ability to spin up HPC. We're also cognizant of the fact that we've got really good Bitcoin mining economics available to us there. And so either way, we want the second half to not be dirt with a long building schedule, but to be ready to go in either direction. I think the timing for the breakdown between HPC and Bitcoin will really come down to the level of tenant interest and how quickly it aligns with the completion of the construction. Based on the discussions we've been having, we do have interested clients in smaller chunks of megawatts. There is a stratification of that HPC market we've seen where obviously, we are showing Black Pearl to the types of clients that can take down 300 megawatts on a gross basis. There's a whole different segment of the market that's not prepared to take down that much, but is very interested in smaller chunks. And we are, let's say, very excited about the potential for Black Pearl. So the idea would be that we complete the construction of this flexible build in next year, and then it is ready to quickly jump to HPC from that Tier 1 half design. Now how quickly it goes will depend a little bit on whether or not the client wants to take it to Tier 1, Tier 2, Tier 3, how much redundancy do they need, et cetera. But the idea is to be ready to go quickly because as an overarching theme, we don't see demand going in the wrong direction, frankly, in the near term. And so we want to be able to capitalize on that quickly.

I appreciate that color. And then just digging deeper into your remarks about hyperscaler participation. Curious to hear what you think would trigger them to choose an operator in your peer set to be a dance partner. How far are we from seeing that happen? What are you seeing from boots on the ground perspective? And are you getting a sense that they're more open to dealing with players in the peer set following the recent deals by CoreWeave and COR42?

So yes, I do see them as more willing to take our peer set and us more seriously. I would say that interest waxes and wanes, and I know I'm not the only builder of data centers that thinks this. But if I had to generalize the market color, I'd say April, May, June was a little quieter than it had been earlier in the year. I'd say July, the level of interest across the industry really came back to life. We had a lot of inbounds in July from hyperscalers that we have had previous discussions with that had kind of been quiet for a while, come to life with requests for emergency Friday evening meetings to get ready for a big review they had Monday afternoon. As it relates to our own discussions, as you've seen with everyone across the industry, and we understand well also, you don't have a signed deal until you have a fully signed deal. And given that, if the phone rings, we'll pick it up and line up interest for the site. We are reasonably far along with someone we would like to be the tenant at the site. But until we have a deal, we don't have a deal. And I'd say just as general market color, the last month or so has been a very big step-up in the level of interest. So to answer your question directly, I do think that those hyperscaler names are coming to this sector for the people that can deliver what they need. [ id="-1" name="Operator" /> Our next question comes from Justin Pan with Clear Street.

This is Justin from Brian Dobson. Exciting hybrid strategy for Black Pro Phase 2. I was wondering if you could talk about any potential differences in costs affiliated with the hybrid model compared to a pure mining or a pure Bitcoin site? And then -- and is it something you would consider for future sites in the development pipeline?

Thanks, Justin. Those are actually very good questions. So the way we've looked at it, it will be about $1.5 million per megawatt for this, again, kind of Tier 1 half style build that we envision that is very modular and ready to move quickly. So on 150 megawatts, that will take us up to $230 million or so in infrastructure costs. From there, it will depend on the requirements of the tenant where if they're stepping all the way up to Tier 3, it will be in line with typical building costs for a Tier 3 data center. It might be another $8 million or so on top of that. It's going to depend on the levels of redundancy effectively that they're going to need. I think what we've heard, the reason for that thinking is that we've had some discussions with folks that, again, are looking for smaller chunks of megawatts than 300 megawatts at a single clip. And one of the concerns they have, and this is really kind of on that neo cloud side is that the GPU upgrade cycle is moving so quickly and construction time lines take a while that there's this fear that they build a data center that is -- by the time it's completed, is sort of built for older GPU technology. What happens if we get crazy increases in rack density and you've got 1 megawatt racks in a few years, how are you going to make your site ready for that? And so that's really informed the thinking behind what we're trying to do here, which is to say we will make this modular and very ready to accommodate upgrades that are industry-wide. I think we were sort of fascinated with Elon Musk building as the Colossus data center so quickly, and that really leveraged this modular concept. And so I think we're trying to echo that. Then as far as future sites, your second question, yes, I think it will depend a little bit on how our tenant negotiations go at Black Pearl. But I would anticipate a similar approach at Stingray at this point, which, again, we'll get a premium because it's going to be megawatts available in 2026. We're doing the substation work there now. And so it would be our plan to continue forward with this build spec that has optionality built into it. [ id="-1" name="Operator" /> Our next question comes from John Todaro with Needham.

Congrats on kind of the additional HPC pipeline. First question, when you think of a financing partner on board for the JV, could you just remind us of the split and expected economics? And also, would Phase 2 at Black Pearl include that same JV structure?

So John, I think you're referring to what we had previously announced our relationship with Fortress. We haven't come out with sort of every number in that because it's a little bit dependent on the terms of the lease we signed and how that flows through. The color I can give is that the structure envisions us owning 20% in the equity of the JV on day 1 within increasing levels of promote as the IRR increases on the total project. And so we have the ability to buy more of that equity up to 49%. I think what I've given in the past is that if we assume, let's call it, a market typical lease and we assume progress and successful refinancing and potential exit in a few years at holding cap rates consistent, when you collapse those layers of the payout waterfall, we would end up owning about 40% of the total economics if we didn't contribute any further dollars. But I want to give the caveat that that's all going to depend on the actual terms of the lease and refinancing and where the market is, et cetera. but that's the color I can give. As far as Phase 2, the discussions we have had and the backstop for the cost of the full build envisions just the Phase 1. That said, I would be very surprised if they didn't want to participate in a Phase 2 expansion down the line. And I would imagine there would -- if for some reason, they didn't, it would not be hard to find financing partners that would be very attractive for that Phase 2.

Got it. Okay. That's helpful. Yes, and that was what I was looking for. And then just going back to that question, the gentleman from Clear Street asked. So if I add it up, I'm still getting like $9.5 million per megawatt. If we just assume the same redundancy needs as, say, like CoreWeave needs with Applied Digital, are you kind of back into that $11 million to $13 million per megawatt for a greenfield build? Or is there cost savings on the flexible model?

I mean there should be some cost savings versus those numbers. I don't know if I call them savings, but keep in mind, we've already acquired the land, the interconnect, built the substation, right? So there's already some sunk costs if we were to spread them across all 300 megawatts. So I don't think that ballpark is off if you think about total cost to the same building spec and considered the sort of proportional cost of the shared infrastructure at the data center. It's in line. [ id="-1" name="Operator" /> Our next question comes from Brett Knoblauch with Cantor Fitzgerald.

On the Bitcoin mining side, 23.5% by the end of the third quarter, should we be expecting anything else? Or would that really -- or anything else after that? Or will that be really dependent on what you guys initially decided to do with the second phase of Black Pearl? And it sounds like that is a 2026 timing?

Yes, it will be dependent, Brett. I mean to give you a sense, we're excited to get the Phase 1 fully installed because you really see the benefits of scale, considering that we haven't seen any meaningful increase in headcount and kind of overhead. So if I look at our point-in-time hash cost right now, it's about $27 per petahash per day. I expect kind of post Black Pearl's Phase 1, we'll be at about $24. I'm talking cash -- total cash cost. So that's electricity, ops and SG&A cash after the completion of Black Pearl. So that's a really strong positioning from my perspective. I think beyond that, it's going to depend a little bit on the pace of tenant negotiations. Fair to say we're going to fill Black Pearl Phase 2 with HPC tenants if we can. I think the idea is depending on the pace of those negotiations and the attractiveness of mining, we'll be ready to -- hydro Bitcoin mining would fit in very easily at Phase 2, but we don't have any specific plans for that. I think the other thing to keep in mind there is given the kind of tariff landscape and how quickly things seem to move on that front, the way I think about that is we'd probably be more likely to buy hydro machines in the spot market that maybe aren't the latest and greatest generation because we can manage to a cheap -- we can manage curtailment to a cheap power cost there. And I think -- I don't know how other companies are thinking about it, but let's just say, long-term futures purchase contracts of the latest and greatest machines where you wait 6 months for machines to get delivered from Asia are just not as appealing to us right now. So I do think it'd be -- just trying to give some extra color, we'd probably look to do a cheaper build with hydro in Phase 2, depending on how those client negotiations go. And yes, that would be coming online back half of next year. [ id="-1" name="Operator" /> Our next question comes from Mike Grondahl with Northland.

Just trying to understand what sites would you say are being actively marketed for HPC right now? Is it basically Barber Lake and maybe Phase 2 Pearl? How would you characterize what you're actively marketing today for HPC?

Sure. So great question. I'd say everywhere -- so you're right. I mean, Black Pearl -- pardon me, Barber Lake is our primary focus. But again, what we're envisioning there is a single tenant. So that's a small pool of potential tenants. The sort of separate part of the market then that we're actively marketing to is Black Pearl Phase 2, but also Stingray and Reveille. We have one discussion going on where we have a potential tenant that's looking at Black Pearl, but also has San Antonio as a target market. And they are the type of tenant where 70 megawatts would be about the right amount. And candidly, as the calendar rolls forward, you start to get within the construction time line that is logical for Reveille. So I'd say that's where the primary marketing of like live tenant discussions are. I'll also say the hyperscalers ask questions about the 3 Ms. On Milsing, I do expect we'll be moving towards an interconnect this year. That has generated a fair amount of interest because it's close to the Houston market, and it's 500 megawatts. So we don't have the interconnect yet. I expect that we will this year. And I think at that point, some of those exploratory discussions with the larger end of the tenant spectrum will become more tangible.

Got it. And then maybe just as a follow-up, any specific comments on Stingray? How are you feeling there? What kind of color?

Yes. I mean I think, again, we've been encouraged by the discussions we've had with what I'll call this middle tier of the market, where there's a fair amount of interest from folks that may be at lower tier data center requirements with smaller megawatt interest. And so the truth is if you're a hyperscaler and you're making a massive investment, of course, you want the megawatts sooner, but you also want to be closer to a major metro because the use case for that data center then fits into all their different buckets. Is this cloud services? Is this training? Is this inference? If you're a high-frequency trading firm that wants 10 megawatts to train your model, you'd be happy to ship it to somewhere more remote. And even though, I mean, listen, the inference stats aren't bad at Stingray, we think it's sub-10 milliseconds to Dallas. I think what we've found is that there is more interest in that segment than we had appreciated. And I think we're -- while I don't think we're looking to have like 1 megawatt tenants, we'd be happy to run a multi-tenant colocation facility. And that's what I think it will end up being, but we'll know more as our tenant discussions progress at Black Pearl Phase 2. [ id="-1" name="Operator" /> Our next question comes from Chris Brendler with Rosenblatt Securities.

Tyler, I wanted to ask a question on another follow-up on Barber Lake and just sort of from a cadence standpoint and the time line here. It felt like last quarter with the Fortress announcement that things were making progress. And I know from other folks that I've talked to in the space that these things do ebb and flow. You've had -- I think certain companies have had serious discussions and then had them follow up had similar experiences where you've had made a lot of progress and think you're getting close and had folks walk away? Or has it been more of a steady progress towards a contract? And just maybe take your temperature on your optimism of getting something done before year-end for Barber Lake would be great.

Sure. I would say that, unfortunately, our experience has been very similar to others in the space that -- there's a lot of hurry up. We need to get a site visit next week. We're sending over terms because this is now a giant priority and then silence for a month. Without getting into the really gory details, we definitely have had some opportunities very close to being finished that have ebbed and flowed. That said, I am very confident that we will have a deal done. I anticipate it will be this year. The progress is not a steady drumbeat progress. It's more like a crab walks across a beach. It's sort of a few steps forward, a few steps sideways, a few steps back, a few steps forward. But we are only more confident based on what we've seen, especially over the last month. Lots of inbound interest, which just reaffirms that this market is there and the current discussions we have going on will get to the finish line. It's really hard to call the timing. I think the real message that I want to give to our shareholders is we're not necessarily going to do the fastest deal. There have been deals that we were uninterested in. We are trying to do the right deal that sets us on the path where we want the company to be. I know there's a lot of short-term trading mentality investors in this space that want any deal tomorrow. Just candidly, while I appreciate their support, that's not who we're managing the company for. We're going to build -- we're going to get the right deal. The right deal takes time. We're very confident in the pocket aces that we're holding, and we're going to play them appropriately.

Makes sense. I love the craft analogy, is there any time line on the Fortress agreement that you have to abide by and get something done by? Or is it pretty open-ended?

Yes. I mean there is a time line under which it's exclusive. I don't think of that as any kind of time line on which we need to get a deal done. I think they'd be happy to extend that, I'm assuming. And if not, we have other financiers that have been knocking on the door. So I don't think of that timing as necessarily anything that would impact our discussions. But yes, technically, we are exclusive with them as a financing partner for another month or so.

Okay. I'm going to ask a Bitcoin mining question.

We love both.

Yes, exactly. So this is a bit nuanced, but I know you mentioned it's clear in the slide deck that the average power cost ticked up this quarter. $0.07 to sort of $0.031-ish and it's also the impact of Black Pearl. But the cost of revenue only increased very slightly sequentially. So our model is like having trouble balancing that cost of power increase with the relatively stable sequential change in cost of revenue. So is there anything impacting those numbers that makes that math a little wonky.

No, let me explain that math. So the $0.031 headline number is an estimate over a large time sample of what we will pay. That's completely the impact of Black Pearl. So $0.027, which we have always had historically was really anchored by our PPA at Odessa, which is right around that level. And that's also when sites like Alborz, which has our cheapest power were a larger portion of our production. Going forward, Black Pearl will be a huge part of our production. That's merely an estimate because it's a front-of-the-meter site. I think what we're using in there for large sample size is about $0.035 on average. I actually hope we'll do better than that. But -- so that's the reason for the change is that if you're looking at numbers on a backwards-looking basis, that hasn't had much impact yet. The $0.031 is expectation across a cycle with everything running. [ id="-1" name="Operator" /> I would now like to turn the call back over to Tyler Page for any closing remarks.