Home / Transcripts / Cipher Mining Inc. (CIFR) · November 3, 2025

Cipher Mining Inc. (CIFR) Earnings Call Transcript

November 3, 2025

NASDAQ US Information Technology Software earnings 65 min

Earnings Call Speaker Segments

Courtney Knight executive
#1

Hello, everyone, and thanks for being here on this call as we discuss Cipher Mining's Q3 2025 business update. With me today are Tyler Page, our Chief Executive Officer; Greg Mumford, our Chief Financial Officer; and Edward Farrell, Senior Advisor and previous Chief Financial Officer. Our press release and presentation are available in the Investor Relations area of the company's website, and this call is being webcast there at the same time. I also want to note that this call belongs to Cipher Mining, and recording it or reproducing it in any way is strictly forbidden unless we give prior consent. Before we begin, I want to remind everyone that what we discuss now, along with our press release and presentation, includes forward-looking statements. Those statements cover, among other things: Cipher's financial outlook, business plans and objectives, and other future events and developments, including statements regarding the market potential of our business operations, potential competition, and our goals and strategies. The forward-looking statements and risks discussed on this call, including in response to your questions, reflect current expectations as of today, and Cipher has no obligation to update or revise them, whether because of new developments or otherwise, except as the law requires. In addition, we may discuss non-GAAP financial measures. We sometimes use non-GAAP measures to describe how we manage and operate our business. We reconcile non-GAAP measures to the most directly comparable GAAP measures, and we encourage you to review those reconciliations, which were filed at the end of our earnings release issued earlier this morning. I'll now hand the call to our CEO, Tyler Page. Tyler.

Tyler Page executive
#2

Appreciate that, Courtney. Morning to all of you, and thanks for being here with us. Tyler Page here, chief executive at Cipher Mining, and it's my pleasure to have you on our Q3 2025 Business Update call. Q3 proved genuinely transformative for Cipher, since we achieved enormous progress on our strategic shift toward high-performance computing and laid the groundwork for what is, no question, the most thrilling earnings update in the history of our company. During this quarter, we completed a pivotal deal with Fluidstack and Google, which solidly cemented our credibility within the HPC arena. After that groundbreaking transaction, and building on that achievement, we have now advanced another significant step. I'm delighted to share with you today that we've closed a second landmark HPC transaction, this time alongside Amazon Web Services. Working directly with one of the biggest and most innovative enterprises on earth highlights Cipher's rise as a trusted leader in next-generation compute infrastructure and affirms our complete transformation into an HPC data center developer. That initial HPC deal involving Fluidstack and Google proved not just Cipher's credibility as a data center developer serving the world's most demanding tenants, but likewise how desirable more remote parts of Texas are for next-generation data centers. For more than a year, we've been telling investors about this thesis and stating our belief that the market would move in our direction. Our second long-term lease, now with Amazon, demonstrates that neither we nor West Texas amount to one-hit wonders. Our second lease confronts the largest hyperscaler on the planet directly, on a 15-year lease with highly attractive terms. This isn't a fluke, and it won't be our final HPC deal. Per the agreement, we signed for 300 MW of gross capacity, and the project holds roughly $5.5 billion in contract revenue across the initial 15-year term. Delivery of that capacity comes in two phases, starting July 2026 and finishing in Q4 2026, with rent beginning August 2026. Because of how strong the lease we've locked in is, we anticipate using debt financing to cover most construction costs at the site, and whatever construction obligations remain will come from cash on hand, with no additional equity fundraising required. With these milestones achieved, Cipher has formally arrived as a leader in the HPC revolution, drawing on our sourcing expertise, energy assets, best-in-class team, and operational excellence to power the world's most advanced computing workloads. Carrying that momentum forward, we're excited to announce today that we've obtained ownership in a joint venture to develop a 1 GW site in West Texas. Assuming standard lease and development terms, we expect to hold about 95% of the JV once a turnkey HPC lease is signed. We're naming the site Colchis, a reference to the mythical birthplace of the Golden Fleece and a land of legendary wealth sitting at the edge of the known world. Over the last eighteen months, the accepted wisdom among traditional data center players held that hyperscalers would stay within big metro regions and that our locations sat at the far fringe of the map. Those established competitors have now been definitively shown to be mistaken. At Colchis, we will keep proving that point. This marks the biggest single addition ever made to our development pipeline; the property carries a fully signed 1 GW direct-connect arrangement with American Electric Power, which delivers dual interconnection capability along with targeted power availability in 2028. The deal further grants options to buy as much as 620 acres of land neighboring the current substation. Every criterion for a top-tier HPC development opportunity is met by the Colchis site: abundant acreage, power capacity at large scale, access to multiple diverse fiber routes, and the ability to interconnect in two ways; early-stage conversations with prospective tenants for this location are already underway. Once again, completing this transaction showcases our team's sourcing know-how and its capacity to lock down some of the most desirable large-scale sites anywhere in the world. Few firms globally besides Cipher can merge hands-on, on-the-ground expertise — collaborating directly with landowners to identify best-in-class sites — with the deep technical sophistication required to serve hyperscalers. Thanks to this distinctive and potent blend, Cipher is positioned exceptionally well to close the widening gap between scarce suitable sites and soaring demand from large-scale tenants. What we announced today reflects years of diligent effort along with the robust execution and momentum we have accumulated during the last quarter. Let me pause to look back at a few of our Q3 wins. Leading those highlights is our recent deal with Fluidstack and Google — a transformative 10-year, 168 critical IT megawatt AI hosting agreement that established Cipher, for the first time, as a major developer in the HPC arena. Through this contract, Cipher will provide 168 MW of critical IT load at our Barber Lake site in Colorado City, Texas, backed by as much as 244 MW of total capacity; across the initial 10-year term, the project accounts for approximately $3 billion in contracted revenue, and options could push the overall contract value to about $7 billion across 20 years. Significantly, construction at the site has already started, and we remain on schedule to hand over the complete 168 MW of critical IT capacity by September 30, 2026. Of note, Google is backstopping $1.4 billion of Fluidstack's obligations in support of project financing, and in return will get warrants amounting to an approximately 5.4% pro forma equity stake in Cypher. Full ownership of the site stays with Cipher, which is currently working to secure debt for construction funding. Additional specifics regarding that construction financing will be shared soon. We held the belief — and have since demonstrated — that Barber Lake was merely the start, the first among multiple projects that leverage our team's sourcing know-how, demonstrated development capabilities, solid industry relationships, and unmatched construction track record. We're eager to keep working alongside top technology firms to lock in HPC leases across our expanding set of sites. Underpinning this growth is the $1.3 billion convertible offering we wrapped up during the quarter. That deal stands as the biggest digital infrastructure convertible issuance ever, and demand ran about seven times the deal size, which speaks to investor belief in both our approach and our project pipeline. Thanks to that robust interest, we could capitalize on welcoming market conditions, landing a 0% coupon while making our balance sheet even stronger. Greg is going to go into more detail about the convertible offering afterward on this call. Between the Amazon deal, the Fluidstack and Google transaction at Barber Lake, the sizable new capacity brought online at Colchis, and the convertible offering we just closed, each one marks a significant step forward in our HPC strategy. Taken together, these wins broaden our business model, lock in considerable future capacity, and bolster our balance sheet, all of which put Cypher in position to seize the enormous demand before us and to serve a vital role in constructing the coming generation of AI infrastructure. Even as we grow and widen our business model, our Bitcoin mining operation keeps producing substantial cash flow. The company beat expectations during this quarter and now runs roughly 23.6 exahash per second in self-mining capacity. That same disciplined base we built in Bitcoin mining — bringing five data centers in on schedule and within budget — will drive our successful push into HPC.

Tyler Page executive
#3

At this point, I want to give a quick rundown of our energy portfolio, showing how we've executed across our business lines and the strength of the pipeline ahead. Turning to mining, during this quarter we completed the full ramp of Black Pearl, taking our operational mining capacity up from 423 MW to 477 MW spanning Odessa, Alborz, Bare, Chief, and Black Pearl. With that, we came in above our earlier hash rate targets and reached a total self-mining hash rate of roughly 23.6 exahash per second. Beyond that, fleet efficiency now sits at a very strong 16.8 joules per terahash, which places us among the industry's most efficient miners. The software we built ourselves allows our data centers to be scaled back on demand, and that capability has obviously given us a major advantage when it comes to maximizing profits, holding electricity costs down, and earning revenue from aging equipment. Looking ahead, we anticipate this know-how will remain an important differentiator, and in fact it may grow into a more and more valuable piece of the company as the HPC sector continues to change. One key point: our existing mining activities are completely funded, and we don't anticipate committing additional capital to that portion of the company as we steer our project pipeline toward HPC. To start, this quarter was huge for Cipher. We went from having no contracted AI hosting capacity at all last quarter to reaching 544 gross MW during this period, thanks to two deals with top-tier partners. On top of that, we hold a strong pipeline totaling 3.2 GW of potential future capacity, stretching from 2025 out past 2029. We're very proud of what our mining operation produces, but the way the market is behaving, combined with limited energy capacity and intense tenant interest, makes it obvious that putting our large site pipeline toward HPC workloads is the smartest move. Talks are continuing with major partners about that pipeline, and we're eager to make HPC development the priority across all those sites. Now let's look at how things stand in each part of the business. Over at Barber Lake, we're building a data center for Fluidstack and Google, partners who lead the industry. Work at the site is moving along nicely. We've broken ground, and both the engineering and procurement work are going well. We have the labor we need, and we've locked in most of the equipment with long lead times, so we're in a strong spot to hit every key construction milestone on schedule. We remain firmly on track to deliver all 168 MW of critical IT capacity by September 30, 2026. We expect the lease to begin the next month, in October 2026. Keep in mind that 56 MW of existing capacity is still ours at Barber Lake. With those extra megawatts, we could go after one more colocation deal, possibly weighting certain deal components differently, or we could put our own compute on-site. To get the most out of the 56 MW that remains in phase I, our team is thoroughly weighing every option available. On top of that, we have an MOU covering an additional 500 MW of expansion at the same location, which we expect to energize sometime between 2029 and 2030. Given the site's ongoing growth potential and the negotiations currently taking place, we look forward to sharing further updates as things progress. Now, regarding the mining business we currently run, Slide 9 lays out a production recap for our five active mining locations. Odessa remains the largest piece of our portfolio, accounting for about 56% of the Bitcoin we produced in Q3. As of September, the site's present operating hash rate sits at roughly 11.3 exahash per second, drawing about 207 MW. Odessa's fleet efficiency comes in at about 17.6 joules per terahash. On this same slide, we also show the observed all-in electricity cost per Bitcoin across our five locations. Further down, Black Pearl started delivering meaningful cash flow to the business during the third quarter. The initial 150 MW at the 300 MW site is presently mining at about 10.1 exahash per second, beating earlier guidance and delivering roughly 36% of production this quarter. Fleet efficiency there is an extremely impressive 13.9 joules per terahash. Finally, we gave a combined look at our joint venture data centers — Alborz, Bare, and Chief. Together, the three sites hold 120 MW of total power capacity and produce about 4.4 exahash per second. We hold a 49% stake in the JV sites, and our share recently accounted for roughly 9% of our total Bitcoin production in the third quarter. Let's move now to an update on our development pipeline. Slide 11 gives an overview of Stingray, our next site to energize, located in Andrews County, Texas. That site comes with 100 MW of front-of-the-meter capacity, every required regulatory approval, and 250 acres of land next to the transmission assets; during the third quarter, we kept advancing substation work there and locked in long lead-time items, such as transformers and high-voltage breakers. Energization remains on schedule for the fourth quarter of 2026. Slide 12 covers the rest of our capacity stretching across 2027 and beyond. Reveille, situated in Cotulla, Texas, stays on track to energize in Q2 2027. That site has full approval for 70 MW, and we've kicked off substation development. With interconnect approvals already secured for both Stingray and Reveille, plus energization schedules that are well established, we've been in discussions with multiple prospective tenants and continue negotiating in order to secure the best possible lease terms at those sites. Final interconnection approval is currently underway for each of our three Ms — McKeska, Millican, and McLennan — and load studies have wrapped up at every one of them. For McKeska and McLennan, the interim Oncor FEAs have been executed with Oncor, and the required deposits have been made. Each of these locations could support up to 500 MW, and our purchase agreements don't just include interconnection rights — they also take in large tracts of land at all of them, every one of which is a strong fit for HPC data center buildouts. As we move forward with development, we're confident demand at these sites will be robust. Colchis, the newest acquisition I mentioned earlier, closes out this part of the discussion, and it represents the largest single addition ever made to our pipeline. There, we have a 1 GW direct-connect contract fully executed with American Electric Power, providing us with dual interconnection capability and power that we expect to have available in 2028. Situated about 80 mi to the southwest of Abilene and some 80 mi southeast of Barber Lake, the site sits extremely well for HPC thanks to its abundant acreage, power capacity at large scale, access to multiple fiber routes, and that dual interconnection capability. During the prior quarter, we walked through how we intend to keep Cipher ahead of the curve as the AI data center landscape evolves. In the time since, two landmark HPC transactions have been completed by us, plus the most significant pipeline addition we have made to date. Because the industry is advancing even quicker than we expected, our confidence is stronger than ever that Cipher ranks among the best-positioned firms globally to capture near-term opportunities arising from the widening power shortfall. To put it simply, our work is only beginning. I will hand things over now to Greg Mumford, our new CFO, who will review the financials for our third quarter.

Greg Mumford executive
#4

Thanks, Tyler, and good morning to all of you. I'm delighted to join this call today, stepping into the position of Chief Financial Officer at Cipher. It's genuinely an honor to be affiliated with a forward-thinking organization that holds such a vital role in shaping where digital infrastructure and high-performance computing are headed. To begin, I want to voice my gratitude toward Ed Farrell for the way he has led and for everything he has contributed over the past five years. Ed has built a finance team of exceptional caliber and leaves behind a solid base that positions Cipher favorably for the growth chapter ahead. We're fortunate that he will remain available as a Senior Advisor during this handover period. In stepping into this position, I plan to stay rigorous about how we manage finances, broaden the avenues through which we can secure capital, and bring down what it costs us overall to fund the business. Our stance on deploying capital will remain deliberate, ensuring we bolster growth that can be sustained over the long haul and generate returns for those who hold our stock. It's exciting for me to collaborate with Tyler, the rest of the leadership group, and the skilled finance professionals here as we carry forward the considerable momentum Cipher has built. Before going further, a quick note that my remarks today center on results from Q3 of 2025, a period that concluded on the 30th of September. I want to point out that during this quarter we showed solid execution — we formally moved into HPC hosting and saw our pipeline grow — and we also raised capital in a disciplined way, which puts us in position to keep that momentum going and even speed it up down the road. In the quarter, we closed our second convertible offering, a private placement that was upsized, consisting of $1.3 billion in 0% Convertible Senior Notes maturing in 2031. Investor demand for this deal was strong, and it showed confidence in Cipher's long-term strategy. The notes carried an initial conversion premium of roughly $16.03 per share, which works out to a 37.5% premium over our stock price when they were issued. Additionally, we put in place capped call transactions, which raised the effective conversion price to roughly $23.32 per share—meaningfully cutting back the possible dilution our shareholders could face. The offering's net proceeds covered the expense of setting up those capped call transactions, and they'll go toward building at our two currently contracted HPC sites, pushing our HPC strategy forward across a development pipeline that now stands at 3.2 GW, as well as toward working capital and general corporate needs. The key point is that this financing strengthens our balance sheet and shows the disciplined way we pursue growth. The market's response has made us quite happy, and we think this deal puts Cipher in a strong position to seize the major opportunities that lie ahead in HPC and digital infrastructure. Now let's go over the financials, and we'll start with how things looked quarter over quarter, which you can find on Slide 14. Our hash rate for Q3 rose 40%, and that came from getting our Black Pearl site powered up and running — phase one of that 150 MW front-of-the-meter location went live in June. Early in the quarter, Black Pearl was producing around 3.4 exahash per second, and by the end of the period it had grown to roughly 10.1 exahash per second. As a result, output climbed 35%, and because Black Pearl is a front-of-the-meter site, our electricity cost per Bitcoin also went up. One more thing that pushed up what we paid per Bitcoin: network hash rate climbed over the course of the quarter. Moving down the slide, you'll see revenue of $72 million — that's up 65% from $44 million in the prior quarter. The growth was mostly due to a stronger Bitcoin price and more production out of Black Pearl. On a GAAP basis, we posted a net loss of $3 million for the quarter, or $0.01 per share, compared with a net loss of $46 million, or $0.12 per share, the quarter before. We're pleased with the sequential gains in our results, particularly since earnings were pressured by higher depreciation costs. That depreciation comes from assets we brought online at Black Pearl, including the newest rigs we put to work, along with the Odessa enhancement completed during the fourth quarter of 2024. Additionally, our earnings continue to be influenced by fluctuations in the fair value of the power purchase agreement at Odessa; these expected movements reflect shifts in forward power pricing and the declining time value remaining on that contract, which extends through July 2027. Ed has pointed this out previously: the true worth of this deal lies in securing fixed-price, low-cost power over the long haul for what we do in Odessa. During the current quarter, as we ran our HPC lease at Barber Lake, we handed Google warrants as payment for their commitment to guarantee the lease payments coming from Fluidstack, the tenant. Since those warrants are carried at fair value, we are booking a $32 million gain this quarter tied to the shift in the fair value of the warrant liability. If we set aside the non-cash items — including the fair value movement on our power purchase agreement, stock-based compensation, depreciation and amortization, deferred income taxes, the change in warrant liability fair value, and losses that don't recur — adjusted earnings for the third quarter came to $41 million, equating to $0.10 per share. That marks a roughly 34% increase over the $30 million recorded in the previous quarter. Our cash and cash equivalents climbed significantly, driven by $1.2 billion in net proceeds from the convertible financing we completed most recently. Let's now move to Slide 15 and take a closer look at our operational results. During this quarter, our Odessa facility produced 383 Bitcoin while Black Pearl contributed 246, giving us 629 Bitcoin in total from all wholly owned locations. At an average price of roughly $114,400 per Bitcoin, that production generated $72 million in revenue. For comparison, Q2 2025 saw 434 Bitcoin mined at an average of $99,700 per Bitcoin, yielding $44 million in revenue. When it comes to general and administrative costs — IT, corporate insurance, professional fees, and other public company expenses — we saw a modest decline both sequentially versus the previous quarter and year over year against the same period. Depreciation and amortization came to $60 million, higher than earlier periods because of new mining rigs brought online during the last 12 months. The oldest machines in our fleet will reach full depreciation in Q4, yet those rigs are still able to stay productive and keep delivering attractive returns when we deploy them strategically. This quarter we recorded a small unrealized gain on our Bitcoin holdings, versus a $17 million gain in Q2, which mirrors a modest rise in the spot price as of quarter end. We ended the quarter with roughly 1,500 Bitcoin held in treasury. Turning to our non-GAAP reconciliation, we posted a GAAP net loss of $3 million. Once you strip out $44 million tied to non-cash and one-time items, adjusted earnings for the quarter land at $41 million, up from $30 million in the preceding quarter. Turning next to the balance sheet: as shown on Slide 17, total current assets finished the quarter at $1.4 billion, compared with $220 million the quarter before, driven largely by $1.3 billion in net proceeds from our convertible offering. On top of that, we also held $170 million of Bitcoin. As we've discussed in detail on prior earnings calls, we take an active approach to treasury management — we don't sell every Bitcoin we mine, nor do we hold every one — and we remain disciplined in how we handle capital management. Let me quickly cover some additional balance sheet entries for September 30. Property, plant, and equipment reached $650 million, up 37% from $474 million. That growth largely relates to equipment put to use at Black Pearl. Equipment deposits stood at $8 million, compared with $183 million the quarter before, which mainly reflects the reclassification of Black Pearl rigs out of deposits and into in-use property and equipment. As of the end of Q3, our equity interests in the Alborz, Baer, and Chief JVs came to $42 million. Further down the balance sheet, derivative assets rose mainly because of the inclusion of $90 million of capped calls tied to the new convertible note, which lifts the effective conversion price on the convertible debt and effectively minimizes potential dilution for shareholders. Current liabilities were higher this quarter owing to the short-term classification of the Google warrants tied to the Fluidstack lease at Barber Lake. Finally, and importantly, I want to point out that short-term borrowings are still at zero. We keep managing the balance sheet conservatively, making sure we're well-positioned to meet any capital needs. Before we wrap up, I'd like to thank everyone for joining today's call. This quarter brought impressive strides forward, along with game-changing expansion, as we continue broadening what we offer, building up our development queue, and reinforcing our financial position to fuel that momentum. Our dedication stays unwavering when it comes to running the business with discipline, making the most of every dollar of capital, and creating lasting worth for those who hold our stock. Thank you for your continued support, and we look forward to updating you on how things go in the coming quarter. With that, I'll wrap up, and Tyler and I are happy to take any questions you may have.

Paul Golding analyst
#5

Thank you very much, and congratulations on the announcement and everything you've accomplished with HPC. I'd like to begin with a question about the deal itself. 300 gross MW, Stingray, which you have on schedule for energization, 100 MW in 2026, and Barber Lake, where you have 56 MW. Once the Fluidstack deal is done, how should we think about how the power will be distributed to deliver that 300 MW, and perhaps also pricing for liquid versus air-cooled, since you're providing both? Based on my rough calculations, the deal averages out to roughly $1.7 million per critical megawatt. If you could walk through some of those deal points regarding pricing, and also how you intend to deliver that capacity across your fleet. I have a follow-up. Thank you very much.

Tyler Page executive
#6

Sure. Thanks, Paul. Thanks for asking. Let me begin with the framework that the ink hasn't quite dried yet on the agreement we signed with AWS. There's still some work finalizing the basis of design and getting you the precise figures that will be shown. They're taking 300 gross. We're recutting an existing air-cooled 150 MW. The first phase of that build will have a quick time to market. On the second build, design is still being finalized, and among the debates taking place is speed to market, meaning how quickly the compute becomes available, versus optimizing for the highest critical IT load achievable; that hasn't been settled yet. I'd say that, broadly speaking, if the entire site winds up air-cooled, the PUE will match the design we have at Barber Lake, which comes out to approximately 1.4 to 1.45. Depending on how much might be handled with more of a liquid-cooled approach, we could see an improvement by giving the second phase a lower PUE, a more efficient PUE. We're still working out precisely what those figures will be. Regarding cost, which you mentioned, it would match, once again, Barber Lake, what we've done previously; we would anticipate the cost per critical IT megawatt to align with that estimate.

Paul Golding analyst
#7

Thank you very much, Tyler.

Tyler Page executive
#8

Then you — Answer from Tyler Page: If it isn't, if it isn't superior, since we already have certain infrastructure that we acquired in a less expensive market.

Paul Golding analyst
#9

Understood. Thanks for that detail. You brought up as well that most CapEx funding comes from debt financing, with cash on hand after that. Could you share more about the financing plans, given that you're already building out the Fluidstack capacity? Does that cash on hand come from prepayment deposits? Is there some form of backstop in place to help with project financing and the process of going to market for it?

Tyler Page executive
#10

I'll lay out a high-level framework on that, and then Greg can jump in if he'd like to address any specifics. There are two separate structures. Clearly, the first deal involves Fluidstack and Google. That structure resembles one that's already been seen in the market. It's fair to say that in the coming days and weeks, we'll be looking to pursue our debt financing options for it. Depending on where the market stands, I would expect those structures aren't too different. That's probably how we would envision it playing out; regarding the AWS lease, that's a direct-facing hyperscaler lease. I believe it's the first of its kind for anyone who has moved from Bitcoin mining to a long-term, 15-year direct-facing hyperscaler lease. That ought to be quite financeable. When it comes to the equity support for whatever the final terms end up being for that financing, remember that we upsized the convert we did recently by a substantial amount. The market was so favorable that it reached a $1.3 billion offering; we already hold a fair amount of excess cash on hand, and by all estimates, we've got that should support what we would anticipate to be the equity piece of the financing to build the structure related to the AWS lease. Greg, would you add any further color, or do you think that's enough?

Greg Mumford executive
#11

Right, Tyler — I believe you put it well earlier. The thing is, when it comes to the Google Fluidstack deal, we aren't ready to share details about the financing we're considering. That said, we are looking into possibilities, and hopefully we'll be bringing the market an update before long. On the AWS deal, we believe plenty of opportunities lie ahead for us to examine — various forms of project-level or construction-level financing — and we'll evaluate those choices and ensure we're landing on the right one.

Paul Golding analyst
#12

Thank you very much.

Greg Lewis analyst
#13

Yes. Thanks, and good morning, and I appreciate you fielding my questions. I suppose the initial question concerns the extra power sourcing. Nice work on that; it appears quite difficult. Tyler, when we consider, and you're mentioning items speeding up and sort of what can be done, might you roughly estimate how matters are advancing and what you're observing at ERCOT? You've got those various extremes you mentioned, 500 MW. At what point did those enter the queue? Clearly, we have some power becoming available in 2026. Just sort of a general update regarding how we ought to view power availability from that expansion pipeline you possess.

Tyler Page executive
#14

Absolutely. I'll provide some detail regarding the sites waiting on final ERCOT approval. Much of this comes down to — To begin with, they've sat in the queue for some time, and every load study and all items have been turned in. A bit regarding the expected timing depends on the sort of business operating model of the specific transmission distribution service provider you're dealing with; regarding Colchis, we expect a 2028 energization; we've already made a capex payment, so construction and advancing construction payment to help with the work the TDSP must perform. That involves AEP, and AEP feels confident proceeding with that construction given an expectation of getting that site energized by 2028. That's our status there. I mean, construction will move forward on the AEP side, and we're in active discussions while they wait for that final approval from ERCOT. At McKeska and McLennan, we've executed interim FEAs; that's a mandate of the TDSP there, which is Oncor. In those instances, once more, the deposit's been paid, but construction will probably start on the Oncor side once that final ERCOT approval is secured, which we're eagerly awaiting. At Millican , we haven't paid a deposit as of yet. Again, working with a separate TDSP there, their process functions somewhat differently. That's roughly the complete picture. Regarding ERCOT, it's difficult to offer any forecast with precise specificity. Given the advancement in anticipation of where we believe those sites will be available based on feedback from the TDSPs, we're confident in the timelines we've provided.

Greg Lewis analyst
#15

All right, that's really useful. Now, regarding the flexibility we have with that 56 MW — I believe you brought up the possibility of perhaps providing AI cloud services of our own. Could you expand a bit on how we're approaching that, in the sense of simply adding a different customer? Perhaps there's an option that might be extended. Just how we ought to be thinking about that 56 MW, and perhaps regarding when it might happen. Is this a matter we want to have wrapped up within the coming 12, 18, 24 months, or is it, hey, it's sitting out there and we've got time working in our favor?

Tyler Page executive
#16

It depends — that's the answer. There have been plenty of inquiries and a good deal of interest, I'd say, regarding the notion of us owning and running our own GPUs and then selling the compute to an off-taker. Broadly speaking, our progress on that front has been gradual, since we need to be certain we're achieving the best risk-adjusted returns on the megawatts we hold. Naturally, you can generate figures that look bigger on the revenue side when you're selling compute, but you're absorbing a whole set of risks — far greater financing risks, GPU lifecycle and obsolescence risk, and so forth. In my view, one key to making such a business really attractive would be securing a long-term off-take for the compute with a counterparty that's highly credible. Those deals are out there — we've seen them, and we're examining them. Honestly, I'd say the numbers we agreed to on our lease with AWS are superior. My sense is we'll likely earn more in profits on both counts, while taking on far, far, far less risk. From where we sit, it's still unclear which use of a megawatt is best for maximizing money. We're in very active talks and are exploring every business model available. Clearly, as we sign up additional 1 GW sites, a great deal of optionality will open up as things move forward. As for that particular 56 MW, I'm quite confident some sort of deal will come together there pretty soon. There's substantial interest both in using that capacity to run our own GPUs and sell compute, and in having it leased on a co-location basis. It's reasonable to say this market is literally becoming more frenzied with each passing week, if not each day. Lease rental rates are climbing quickly. The amount of interest is overwhelming. From our standpoint, we're spoiled for choice; we've placed ourselves in a highly advantageous position. We'll proceed based on which deals we believe will yield the best risk-adjusted returns. My view is that the 56 MW there, along with the 100 MW at Stingray and the 70 MW at Reveille, will all be taken up. Should this level of market interest persist, we won't have a single available megawatt. Multiple parties are interested in all those sites and in locking them up as soon as possible.

Greg Lewis analyst
#17

Alright, thanks a lot, and congratulations regarding that AWS news.

Tyler Page executive
#18

Appreciate it, Greg.

Andrew Beal analyst
#19

Hi there. One thing I'd like to understand is how you're approaching the design for Colchis. For a greenfield project like that, what do you figure the capex will likely run per megawatt? And when it comes to ERCOT approval, could you explain how securing the leases with Google, Fluidstack, and AWS helps you get the green light at your other locations, including the three Ms, and how significant an impact does working alongside AEP have when it comes to that approval process?

Tyler Page executive
#20

Yeah. Thanks so much for asking. It's somewhat difficult to forecast the budget numbers for Colchis, for the simple reason that, first of all, again, this is yet another one. We only just closed the acquisition — the ink hasn't even dried. Because of how large the site is, we've started having preliminary discussions with parties who might want to co-locate there. In the meantime, what I'd say we'd do is spend capex on the bare minimum needs at the site. That means fiber, substation, land, water sourcing, and so on. My expectation is that our construction cost would match what we've seen at our other sites, assuming we build the same kind of co-location access, which has typically come in at roughly $9 million-$11 million per critical IT megawatt. That being said, inflation is possible, pricing may shift, supply chains, and so forth. Other than the simple fact of time passing and those variables, I see no reason why the per-megawatt cost would differ; we'll be in a position to share more specifics over the next months and quarters. Honestly, with power expected to be available in 2028 and given how big Colchis is, we're hoping to line up a partner before long, simply because that construction schedule and commitment are enormous. We'll need to get started on it. I don't see any reason the cost would come out differently. Sorry, what was your second question again? I lost track there.

Andrew Beal analyst
#21

Pretty much, yeah. I mean, when we sign those leases with Google and AWS — I mean, how does that shift things for us with, sorry, three Amazon approvals.

Tyler Page executive
#22

Yeah, thanks. Yeah, these partnerships bring enormous advantages, in my view. Once more, until just a few months back, I can't count the number of times people told us, "Nobody will ever commit to those locations. Nobody will ever commit alongside a former Bitcoin miner — certainly not a mainstream hyperscaler." Well, that argument has ended now, clearly. It likely won't be only for us; it'll be for others too, as additional agreements get inked throughout the ecosystem. My sense is each agreement brings added credibility; we've earned plenty of credibility. Anybody who came to understand how good our team is, what they've experienced, what they've constructed before, and simply examining Cipher's record itself — if you removed the term Bitcoin and simply stated, "Our team has completed five data centers on schedule and within budget in this very same geographic area," nobody would have cause to question our word. It's purely the ingrained prejudice of established sectors toward that term and Bitcoin. Each agreement, I believe, builds credibility with everybody. Agreements lead to more agreements. I've spoken about this quite a bit — about landing our first agreements, prioritizing counterparty quality, and structuring our business as a franchise so we can capture maximum value from the whole pipeline we possess. I'm glad to report that's precisely what's happening. Each discussion becomes slightly simpler, and we've gained far more credibility on fresh leases with regulators, with transmission distribution service providers. Honestly, that kayak I referenced in the Colchis situation, which is in fact set to depart soon — that's what ERCOT cares about, correct? Greater credibility plus capital deployed in this sector and being a trustworthy counterparty motivates a transmission distribution service partner to advance your project and commit their own funds, since repayment becomes more probable. Same thing on ERCOT's end. All of this generates further success, and that's likely the main source of optimism around here currently.

Andrew Beal analyst
#23

All right, thank you.

Michael Donovan analyst
#24

Appreciate you fielding this, Tyler, and nice work on the momentum so far. I suppose when it comes to the supply chain, what kinds of bottlenecks are you encountering with long lead assets?

Tyler Page executive
#25

Sure. Look, I believe we've discussed this across the years — when building out a schedule, we frequently begin from the back end, asking what those long lead time items will be. Broadly speaking, the substation being in place is the key driver. Past that point, on the HPC side, the basis of design for each specific site — shaped by what the tenant needs — has some influence. Generally speaking, should they require backup gens on hand to deliver the uptime that's needed, those usually become the following gating item on the timeline. I'd say our track record is strong. Bear in mind our construction team has roots at firms such as Vantage, Whiting Turner, Google, and Meta. They bring deep experience securing every item these data centers require, and they've built relationships throughout the supply chain. To offer some perspective, by rough calculation for Barber Lake, I believe more than 85% of the equipment is locked in, encompassing every long lead time item; this happens within each build spec and will keep happening. Broadly, the risk we now face — and that anyone signing these deals now faces — is obviously executing the construction, funding whatever gets built, and then finishing that construction on schedule. Our team's record on that is outstanding. I see no reason to think we won't outperform everyone else in this space when it comes to on-time delivery. The supply chain shifts around somewhat, but I believe our positioning is quite strong. Across the builds we have underway, our confidence in the timelines — which are aggressive — is very high.

Michael Donovan analyst
#26

Thanks, that's useful, Tyler. My follow-up is probably a little more obscure. I keep hearing people talk about interconnecting locations — for instance, a 500 MW facility in one place and another 500 MW facility elsewhere, tied together so they can provide a 1 GW campus for a particular workload. Are you encountering these kinds of conversations more often? In theory, could we imagine the three Ms being combined into a single 1.5 GW campus?

Tyler Page executive
#27

How the market develops probably determines that. Without question, many hyperscalers are looking for redundancy among data centers located near one another within the same geographic region. What I'd say is, right now we've got a cluster of data center locations — essentially a dozen in West Texas. That said, I don't view the three Ms as situated near enough geographically, at least as things stand today, for us to consider connecting them. The fact that they aren't extremely distant is a positive. Whether people would actually look at them that way, I'm not certain. That pattern certainly is real, but I wouldn't necessarily put our sites together in that category. Having a workforce in that region, along with many other scale efficiencies, makes concentration really valuable. None of our sites sit only 10 mi apart or anything comparable. They're generally somewhat more spread out. Take Colchis — I believe it's roughly 80 mi from Barber Lake. At a broad level, I'd say those customers are interested in discussing the possibility of building their own availability zone. We aren't far enough along to state precisely that these particular sites would be set aside for that single tenant.

Michael Donovan analyst
#28

Excellent. Thanks for that, Tyler. This is my final question, I swear. Fantastic strides being made out at the edge of the earth. How ought we to be thinking regarding places beyond Texas?

Tyler Page executive
#29

Excellent inquiry. Opportunities are something we constantly evaluate. It just so happens that Texas is a place we adore, and it appears we consistently uncover the finest opportunities there. I genuinely believe that one factor is the sheer number of things going on. In Texas, business is fantastic. Conducting business there is wonderful. On top of that, it possesses a tradition of risk takers and entrepreneurs who are eager to speculate on early-stage opportunities. When I look at this, it reminds me a bit of the oil and gas business. In that world, some folks are willing to take a gamble on grid interconnections and bear the risk that comes with trying to lock something down. If I'm being straightforward with you, what might be Cipher's secret sauce — now that we've originated 12 sites down there — is that our team has shown it excels at sourcing these sites from what I'd call grid wildcatters. These are people who invest early in an interconnection opportunity but lack the capability to develop the site to the high standard needed to make it ready for an end user like a hyperscaler. My argument would be that Cipher is basically the only company doing this. Maybe there aren't many companies we're up against, but I'd say nobody matches what we bring. We can talk convincingly to the people who start these projects, and then sit down for a full day of technical discussions with a hyperscaler alongside our whole construction and operations group — and win them over too. Closing that distance, going from what you might call early-stage bets on grid possibilities all the way to handing things off to a top-tier end user, we handle that ourselves, internally. That's honestly the reason I see us as a huge growth opportunity for investors. What we have isn't merely a collection of assets. The key thing is, we have no plans to stop building out these sites. To give you a straighter answer, though — since I'd just been going on about how great Texas is — yeah, we're evaluating locations, especially within PJM. In the past, our site searches have spanned the entire globe. A lot of times the numbers just haven't reached a level we're comfortable with. Our approach here is a relentless focus on risk-adjusted returns. Frequently, either the risk is too great to warrant putting money in, or maybe the cost is too steep. They're too far along. There isn't sufficient risk that we believe we can measure better than anybody else. PJM is on our radar. That's a market we want to grow in. Stay tuned. My expectation is that there will be announcements down the road.

Mike Colonnese analyst
#30

Morning, Tyler, and to everyone on the team — congratulations on those two sizable HPC contracts. It's genuinely encouraging to see. I understand the anticipated delivery schedules that came along with both of those agreements. When it comes to 2026 and afterward, how should we model the revenue phasing from those two deals?

Tyler Page executive
#31

Sure. Delivery in full for the Fluidstack Google arrangement should wrap up by the close of September in the coming year. Rental payments kick off during October 2026. As for Amazon, that one is still being finalized, but it starts in August of next year. That said, it will come in phases. The subsequent phase should land nearer to the end of next year.

Mike Colonnese analyst
#32

Understood. Tyler, here's a question at a higher level. From where you sit, what's different now on the counterparty side that has sped up how quickly deals are being announced across this sector during the last month or so? The urgency coming from the hyperscalers, NeoCloud, and certain others really seems to have escalated compared with only a few months back. I'd love to hear what you think about that.

Tyler Page executive
#33

Yeah. I'd say it's reasonable to claim that across my 25-year professional career, nothing I've seen comes anywhere near what the market is doing at present. You asked why. I don't know. Just like everyone, I listen to the podcast and hear hyperscaler CEOs talking about a shortage. My feeling is that when you're a big, diversified cloud provider, it's easier to predict how much capacity the traditional cloud business will need several years ahead. What I believe surprised everyone was the truly meteoric rise in AI demand. And what's happening at the moment is that not only is that demand off the charts, there's also a scramble, since those folks underestimated how much they'd need quickly. Naturally, there's somewhat of a race among them. At this point, conversations go beyond. The way conversations begin has shifted. At first, people would say, "We want megawatts that are available right now." Then it turned into, "We want anything in 2026." And now it's become, "We want anything in 2027." Honestly, from one week to the next, the mood keeps shifting — more excitement, more demand. Obviously, in a market behaving like that, lease rates are climbing, which is exactly what you'd anticipate. I feel really good about the spots we chose to plant our flags, so that we could secure the strongest anchor tenants anywhere in the world for our company. My sense is that the pricing leverage has shifted to our side now, which lets us enhance the economics and get better terms on our deals. To repeat, this marks our first 15-year long-term lease signed directly with a hyperscaler in our sector — and not just any hyperscaler, but the largest one. It shows where the leverage sits now: with whoever controls the scarce assets, and we put ourselves in that position on purpose over the past several years. Whether this intensity lasts indefinitely, I can't say, but given what we're able to see, we do feel somewhat like the tip of the spear. I've said this for some time, and it remains true — demand is completely off the charts, and it keeps climbing further off the charts.

Mike Colonnese analyst
#34

That's genuinely interesting, Tyler. Tyler, I appreciate your perspective.

Tyler Page executive
#35

Thank you, Mike.

Joseph Vafi analyst
#36

Hi, everyone. Morning. Congratulations on the tremendous headway. Greg, glad to have you here, and Ed, best wishes on retiring. Just 2 quick things. Tyler, maybe you can give the latest thinking on that behind-the-meter arrangement and where things head next for Black Pearl, considering that location along with how its power is uniquely sourced and the fact that the deal has expired, and then factor in, clearly, all the activity happening in the HPC space. What's the path forward for that site?

Tyler Page executive
#37

Joe, are you referring to Odessa? You mentioned Black Pearl.

Joseph Vafi analyst
#38

I'm sorry.

Tyler Page executive
#39

Are behind-the-meter PPAs at Odessa.

Joseph Vafi analyst
#40

Exactly.

Tyler Page executive
#41

Yes. All right, I simply wanted to confirm that what I heard was accurate.

Joseph Vafi analyst
#42

A little earlier in California.

Tyler Page executive
#43

Yeah.

Joseph Vafi analyst
#44

Absolutely.

Tyler Page executive
#45

I apologize for this call happening before the market opens. Yes, regarding Odessa. As a quick refresher, we hold a power purchase agreement for extremely inexpensive electricity covering 207 MW at the Bitcoin mining site we operate in Odessa, and it extends until the close of July 2027. That deal carries enormous worth. It sits deeply in the money. On our balance sheet, we reflect it at a reasonable value, which is possible since the rate is locked in and remains so low for a period. It's accurate to say that amid these conversations, which are frantic regarding additional power that can be brought online right away, we receive considerable interest along the lines of, "Hey, would you ever consider repurposing that location?" My sense of where we stand today is that, because our electricity cost there is so remarkably low, Bitcoin mining represents a terrific business at that site. Down the road, HPC might become appealing there, but we feel no urgency whatsoever, considering how robust our agreement is and what that means for the economics of Bitcoin mining. It's reasonable to say it could be an excellent location. A natural gas generation facility owned by Vistra sits alongside it. As circumstances change, and again tying back to something I addressed previously, as our reputation in this sector strengthens, I believe it's reasonable to say that additional prominent players from across the board will turn to Cipher for their data center needs. There's a chance that something could materialize there; we would need to work alongside our electricity supplier, Vistra, and also work with a prospective tenant. We feel no urgency at all, simply because the economics remain fixed at highly advantageous levels there for a further year and three quarters.

Joseph Vafi analyst
#46

Absolutely. Appreciate it, Tyler. And then, just very briefly — perhaps I overlooked this — regarding this arrangement involving Microsoft, will it take place at a single location, or will it be spread out across several? I'm simply not certain whether that detail appeared in the press release. Thank you.

Tyler Page executive
#47

No agreement has been reached with Microsoft as of now. I realize things look muddled at the moment, since I believe there are a handful.

Joseph Vafi analyst
#48

My apologies. Yeah. Today I keep mixing all of them up. There was one more. Yeah. Amazon. My bad on that one.

Tyler Page executive
#49

No, no problem at all. With Amazon, there's a single large site that has to be switched over from a Bitcoin mining facility into HPC.

Joseph Vafi analyst
#50

Understood. However, I suppose you still haven't told us the specific site.

Tyler Page executive
#51

Correct — that would be Black Pearl.

Joseph Vafi analyst
#52

All right. Excellent. I appreciate it.

John Todaro analyst
#53

Excellent. Thanks, everyone. Nice work on the lease. The schedule for handing over that Black Pearl location for AWS looks quite fast. I'm just curious whether there's something I'm overlooking, or how much certainty there is about hitting that delivery. I have one brief follow-up.

Tyler Page executive
#54

Yes. So our level of certainty is extremely strong. Once more, an agreement of this kind comes about after numerous in-depth technical discussions alongside their staff. Remember that at this particular location, we constructed 150 MW at a remarkably elevated standard of build quality. That differs from our remaining locations, where the schedule was tighter and we possibly relied on a containerized approach. That one was consistently constructed with a long-term view toward being convertible. I'm glad to report that, once more, the bulk of that location can be reused right away in phase one for 150 MW. That is the reason behind the aggressive schedule for phase I. As for phase two, once more, that simply depends on the discussions we've conducted, working through a procurement process and mapping out a supply chain schedule. In my view, we can comfortably achieve it. That aggressive schedule mostly stems from the fact that we're reconfiguring a location that was recently constructed to a very high standard.

John Todaro analyst
#55

Understood. That's clear. For my final question, when you go about securing a location such as Colchis, which companies are you up against? Clearly, you're getting contracts signed with the big hyperscalers. At this stage, do those companies also want to develop certain of their own facilities, or would you say it's mainly perhaps other Bitcoin miners that you're battling?

Tyler Page executive
#56

Yes. Excellent question. As I mentioned before, I touched on this, but I believe this represents the growth equity element of our business that doesn't get enough appreciation, because pulling off transactions of that kind demands genuine familiarity with the area and comprehension; by way of comparison, this typically means working straight with a wildcatter, correct. Regarding the hyperscalers, they're far more accustomed to — for one thing, they're large organizations that operate; they lack the agility that we possess. For another thing, they expect Jones Lang LaSalle to hand them a nicely packaged deal presentation for a finished data center or location that's highly refined and prepared to show them. They won't go into a community to grasp what's needed locally and handle whatever messy circumstances may arise in some of these transactions. I'd say we're definitely the top, or perhaps the sole, firm with exceptionally strong trust among that group for closing transactions, plus the capacity to engage with hyperscalers. A traditional commodities production business attracts multiple tiers of capital that simply aren't present in this space; at that level, we don't encounter nearly as much rivalry from those players, and that genuinely forms a piece of our value.

John Todaro analyst
#57

Okay, that makes sense. I understand now. Thanks for explaining. I appreciate it. Congratulations once more.

Tyler Page executive
#58

Thanks.

Tyler Page executive
#59

I appreciate all of you being here today. There's someone I'd like to recognize, and that's Ed Farrell. From the very first day at Cipher, Ed Farrell served as my right-hand man. Internally, we've shared plenty of thank-yous and congratulations regarding his move into retirement and his shift from Chief Financial Officer to Senior Advisor. I felt it was worth using this moment to offer a particular thank-you aimed at investors. Speaking as one of Cipher's biggest shareholders, I'd like to express gratitude on behalf of all of us for the effort he's put in. To everyone who holds shares, I can say plainly that reaching this point would have been impossible without him. He's been outstanding, and it's genuinely thrilling to see the company arrive at where it stands now, built upon the work he contributed. It's difficult for me to accept that I won't be able to wander through the office trading those obscure Godfather lines any longer. I won't get to hear from him — or rather, I won't get to tell him — that the Don needs to learn about bad news immediately. I suspect that whenever I hit some frustrating obstacle, Ed won't be around to remind me, "Tyler, this is the business we have chosen." Our new CFO, Greg Mumford, means we're in wonderful shape. Considering every bit of capital raising and optimization still ahead of us, our hands are more than capable. On behalf of every shareholder, thank you, Ed, and we wish you a wonderful retirement. Thanks to all of you. We'll be speaking with you shortly.

Operator operator
#60

Thank you for your participation. You may now disconnect. Everyone, have a great day.

Read the full transcript via the API

You're viewing the first half of this call. Get the complete Cipher Mining Inc. transcript - plus 132,000+ transcripts from 7,000+ companies, speaker segments and full-text search - through the EarningsAPI REST API or hosted MCP server.

Get an API key View API docs →

For developers and AI pipelines

Programmatic access to Cipher Mining Inc. earnings transcripts and 132,000+ others is available through the EarningsAPI REST API and the hosted MCP server. Quarterly plans from $162 - full transcripts, speaker segments, full-text search, and the /api/v1/transcripts/recent polling endpoint for ETL pipelines.