Galaxy Digital - Earnings Call - Q4 2024
March 28, 2025
Transcript
Operator (participant)
Good morning and welcome to the Galaxy Digital Fourth Quarter 2024 Earnings Call. Today's call is being recorded. Later, you'll have the opportunity to ask questions. To register to ask a question, please press the star and one at any time. At this time, I would like to turn the conference over to Jonathan Goldowsky, Head of Investor Relations. Please go ahead.
Jonathan Goldowsky (Head of Investor Relations)
Good morning and welcome to Galaxy's fourth quarter and full year 2024 earnings call. Before we begin, please note that our remarks today may include forward-looking statements. Actual results may differ materially from those indicated or implied by our forward-looking statements as a result of various factors, including those identified in our filings with the Canadian securities regulatory authority on SEDAR+ and available on our website or in future filings we make with other securities regulators. Forward-looking statements speak only as of today and will not be updated. In addition, none of the information on this call constitutes a recommendation, solicitation, or offer by Galaxy or its affiliates to buy or sell any securities, including Galaxy Securities. With that, I'll turn it over to Mike Novogratz, Founder and CEO of Galaxy.
Mike Novogratz (Founder and CEO)
Good morning. Listen, we've got a lot to unpack today. I'm going to dive right in, but not before I tell you it's a beautiful day in New York City. I'm very pleased with our full year 2024 results. I certainly wish they were indented by our settlement with the Attorney General. Our team thought long and hard about settling, but when we weighed being tied up for two to three years in a lawsuit with the New York AG versus the opportunity ahead of us, it was pretty clear that the right thing for Galaxy to do was to put this behind us and move on. What I'm most pleased with in 2024 is that our operating business had both positive EBITDA and over $400 million of revenue. In 2018, when we started Galaxy, the idea was to build an institutional crypto business.
I really feel like 2024, we hit the first major milestone in accomplishing that. With the Trump administration's new, really welcome approach to crypto, we're seeing nothing but new opportunities with institutions getting in the space. I couldn't be more excited for the forward trajectory. Our balance sheet had a fine 2024, which should have been expected given crypto's excellent performance. Unfortunately, 2025 has not been kind to many of the crypto assets we hold, and our balance sheet has suffered some. I point this out because up until now, Galaxy's results really have been driven by balance sheet management and the balance sheet in general. I firmly believe that we're at the cusp of transitioning from a balance sheet company to a company with two great operating businesses, one in crypto and one in data centers.
Over time, the balance sheet will have less impact on results. That probably does not happen for at least one or two more years, but that is certainly the direction we are traveling in. The most exciting news of the day is certainly our announced deal with CoreWeave. This is a giant and great first step in what I believe will be building out one of the U.S.' largest AI data centers. I am going to give most of the details of this to Chris Ferraro, who has worked diligently and tirelessly to get this done with a great team. Finally, I want to leave you guys with some macro thoughts. We went into the year with risk assets mostly on the highs. Equities were expensive, and crypto had had a great run. The Donald Trump presidency has certainly added volatility into the mix.
If you think about it, we are taking nearly 40 years-60 years of what was a security apparatus built around NATO and a financial apparatus built around multilateral cooperation, trade agreements that existed since after World War II, and changing all of that. I think that is going to continue to provide a lot of volatility to the market. Overall, I think it is going to be a tough time for risk assets for a while. Equities probably do not trade nearly as well as they did last year. The dollar probably is sold off, and yield curves probably steepen. In that environment, some of that is bad for crypto and some of that is good for crypto. Crypto went into the year with this spectacular enthusiasm based on coming out of four years of purgatory and into this new regulatory environment that was very pro-crypto.
We can't really have a more pro-crypto administration than the one we have right now. A lot of people are asking, well, why are the prices lower? Every once in a while in the markets, lots of the good news get priced in. I think you saw probably a short-term market top around the inauguration. There was plenty of euphoria. There was a crypto ball in D.C. where everyone wanted to go. Crypto CEOs were everywhere. While that maybe marked a short-term top, I think for the medium and longer term, it's wildly bullish. What do I see for the rest of this year? I think we're going to have choppy to tough markets for the next quarter or so. By the second half of the year, the Fed should be cutting interest rates. We already have China and Europe increasing money supply.
When you have all three major forces in the world increasing money supply, that's usually good for Bitcoin and the rest of crypto assets. We also have a government that has pledged to make a lot of changes in crypto regulation. That does not happen really to the second half of the year. You talk about it early and then you get it, you put the sausage through the sausage machine. We should end up by third quarter, fourth quarter at latest with stablecoin legislation, with market structure legislation, and with potentially the federal government figuring out ways to buy crypto. When I look at that backdrop, while I'm cautious in the short run, I'm still very optimistic about crypto in the long run. Finally, I want to welcome Tony Paquette to Galaxy as our new Chief Financial Officer.
Tony joins us from Point72, where he was a CFO for the past four years, leading their global finance team and overseeing the implementation of their digital assets trading strategy. He brings extensive experience in banking, fintech, and asset management, and his strategic vision and leadership will be instrumental in driving our business forward. Tony, now over to you.
Tony Paquette (CFO)
Thanks, Mike, and thank you everyone for joining the call today. I'll start with a quick review of 2024, then talk about the details for Q4, and lastly, provide an update on Q1 2025 preliminary results and an update on our U.S. listing process. First, 2024 was a milestone year for Galaxy, as Mike mentioned, marked by record operating business performance and the decision to strategically expand into building and operating AI infrastructure. Throughout the year, we developed and deepened partnerships with leading institutions globally, expanded our trading capabilities and investment offerings to meet the evolving needs of our clients, and played a pivotal role in executing the largest bankruptcy mandate in the history of crypto while ensuring maximum value recovery for creditors.
In 2024, we delivered $365 million in net income, or $1.02 per diluted share, which includes the $166 million charge from the legal settlement with the New York Attorney General, which is the discounted value of the $200 million in total that will be paid out over three years. Excluding this settlement charge, in 2024, we generated $532 million in net income, or $1.49 per share, on operating revenues of more than $400 million, surpassing the total generated over the previous two years combined. Importantly, this marked the first year of our combined operating businesses achieving profitability, as Mike mentioned. Turning to expenses, excluding the one-time settlement charge, operating expenses totaled $782 million in 2024. Notably, this includes roughly $240 million in staking costs and interest expenses within our global markets business, which are directly tied to increased operating activity and higher revenue throughout the year.
Looking ahead, we expect certain operating expenses to decline in 2025, driven by the absence of one-time items and a meaningful reduction in mining-related expenses as we transition our Bitcoin mining operations over the coming quarters. We had over 520 employees globally at year-end, with roughly 70% in the U.S. and 30% internationally. We will continue to hire in 2025 in order to, one, invest in the growth of our core business, and two, advance key strategic initiatives, including the build-out of our data center. We ended 2024 with approximately $7.2 billion in total assets, $2.3 billion in equity capital, over $1 billion in cash and net stablecoins, and $10 billion in combined assets under management and assets under stake across the platform, which is a powerful testament to the trust we've earned from our clients.
Additionally, we have roughly $809 million in non-current investments on our balance sheet, consisting primarily of fund, private equity, and venture investments, which are marked at fair value as of the end of December 2024. Building and maintaining a strong balance sheet is a core principle for Galaxy, and we will continue to invest in order to fuel growth in our operating businesses and position our balance sheet to capitalize on opportunities we see in the digital asset ecosystem. Now turning to our Q4 and full year operating business results. In the fourth quarter, Galaxy delivered $174 million in net income and operating revenues on operating revenues of $117 million. Excluding the settlement charge, net income was $341 million in Q1. Within global markets, we observed a notable uptick in trading volumes in Q4 as the results of the U.S.
election drove anticipation for greater regulatory clarity and provided a strong tailwind for the digital asset space. Spot and derivatives trading saw sustained momentum with increased institutional participation and heightened demand for option strategies. Additionally, crypto's strong price action alongside with growing on-chain activity contributed to robust market conditions. Against this backdrop, our markets business generated counterparty trading and advisory revenues of $68 million, a 26% increase quarter-over-quarter. This growth was primarily fueled by strength in derivatives trading and lending activity, where we capitalized on rising institutional demand and expanding market depth. As discussed last quarter, we previously crossed the threshold for notional derivatives volume traded, which required us to register as a swap dealer in the U.S. This registration unlocked a massive opportunity for Galaxy, and in 2024, we traded over $20 billion in notional derivative volume, more than twice the total in the previous year.
This strong fourth quarter performance capped off a record year for our markets business, which saw counterparty trading and advisory revenue of $215 million, a 107% increase from 2023. This growth underscores the strength of our platform and the unique ability to serve our over 1,300 institutional clients. This business continues to gain momentum, and we feel we are well positioned to build on this success in 2025 and beyond. On the lending side, we delivered another quarter of meaningful growth with revenue, net of funding costs, up 19% quarter-over-quarter to $20 million, as we saw a notable increase in demand for a variety of capital solutions. We continued to originate loans to both new and existing clients, and our average loan book grew to over $860 million in the fourth quarter, a 29% increase from Q3.
For the full year 2024, the lending business delivered $60 million of net revenue, or 72% year-over-year growth. As a reminder, we take a conservative approach to underwriting risk in the lending space, with average over-collateralization of more than 130% across the entire portfolio and a focus on underlying asset liquidity. We will continue to dedicate capital to the lending business in order to capitalize on growth and maturation in this overall market. On the advisory side, M&A picked up in 2024 as capital markets began to reopen throughout the year. In Q4, our investment banking team successfully closed three deals, and while this business is still relatively nascent, we've seen a renewed appetite for strategic transactions. In 2025, we've seen a modest uptick in activity as traditional finance institutions and crypto-native firms are identifying strategic synergies, and both are looking for opportunities to scale and expand.
Lastly, on the topic of stablecoins, we're very excited about the upcoming launch of AllUnity, the fully collateralized euro-denominated stablecoin that we're backing in conjunction with DWS and Flow Traders. This initiative is pending regulatory approval from the BaFin, which we expect to receive in mid-2025 and has already attracted strong interest from traditional financial institutions and corporates. Longer term, we are very optimistic about this initiative and the broader stablecoin space as adoption has clearly taken hold for this increasingly important financial product. Now turning to the asset management business. In 2024, we made significant progress in expanding and enhancing our investment platform, offering a comprehensive suite of solutions to meet the rising demand for digital assets and emerging technologies. A few highlights. One, we successfully launched two crypto ETFs with Invesco in the U.S. and two ETPs with DWS in Europe.
Two, we forged a strategic partnership with State Street Global Advisors and launched three active equity ETFs, expanding our investment program into emerging technologies beyond digital assets. Three, we began fundraising for a new crypto venture fund, which has raised nearly $150 million so far, garnering strong interest from the institutional LP community. We ended the year with $5.7 billion in assets under management and generated a record $49 million in operating revenue throughout 2024. In addition, we launched the Galaxy Absolute Return Fund in January 2025, a fundamentally driven multi-asset hedge fund designed to provide investors with thematic exposure to digital assets while maintaining an equity-like volatility profile. In parallel to scaling the alternative segment of our platform, our team has been actively engaged in liquidity provisioning deals, deploying capital to new networks and applications in need of liquidity for their protocols.
With a strong balance sheet, deep market expertise, and extensive relationships across the ecosystem, Galaxy is well positioned to capitalize on these emerging investment opportunities. We're excited for the road ahead and look forward to sharing further updates as we scale our AM business throughout 2025. Now turning to our digital infrastructure business, let's start with staking. In Q4, we generated $87 million in gross blockchain rewards and distributed $61 million, resulting in net blockchain rewards of $26 million, a 147% increase quarter-over-quarter. For the full year, the staking business generated $58 million in net rewards, underscoring our rapid growth and expanding market presence. At the start of 2024, Galaxy had approximately $240 million in assets under stake, a figure that grew by nearly 20x to $4.2 billion at year-end.
Over the past few years, we've built a globally distributed multi-cloud staking infrastructure supported by a growing team of dedicated professionals. We established and maintained our position as one of the largest validators on Solana and expanded our capabilities to support an additional eight unique protocols. We have integrated with multiple custodians who collectively have more than $100 billion in assets across their platforms and whose clients can now access our staking services directly through their custodial accounts. We expect these integrations to drive meaningful organic growth for our staking business. For 2025, we are focused on developing new integrations and partnerships to help broaden the reach of Galaxy's staking services, and we are very excited about the opportunities to build, grow, and innovate in this space.
Now turning to mining, our mining business reported revenue of approximately $22 million in the fourth quarter for a total of $95 million in 2024. In Q4, we maintained a competitive average marginal cost to mine of less than $38,000 per Bitcoin, flat quarter-over-quarter, and reported a marginal cost to mine of approximately $26,000 for the full year. As discussed in our Q3 earnings call, and as Mike mentioned earlier, we've made progress on the strategic shift of transforming the Bitcoin mining operation at our Helios campus into an AI HPC data center business, which Chris will talk about in more detail shortly. As a result, we plan to discontinue mining activity at Helios in the second quarter and expect our remaining Bitcoin mining activity to be materially smaller going forward.
Before I turn it over to Chris, I want to touch on our Q1 2025 preliminary performance and provide an update on the U.S. listing process and some future enhancements to our financial reporting. As of March 27th, we expect to generate a pre-tax loss of between $275 million and $325 million in Q1. We expect our operating business net income, excluding balance sheet holdings, to be down modestly in Q1, and we anticipate ending the quarter with total equity capital of between $1.9 billion and $2 billion. As outlined on page five of our earnings release, we came into 2025 with approximately $1.3 billion in net digital asset exposure across Bitcoin, Ether, and Solana, which together comprised the majority of our net digital assets held on our balance sheet.
Crypto market volatility has increased in Q1, as Mike mentioned, with Bitcoin down 7% and Ether down 40% year-to-date, which has led to a slowdown in client activity, lower assets under management and assets under stake, and a reduced value of the net digital assets held on our balance sheet. Importantly, Galaxy has a strong track record of successfully navigating challenging market environments and leveraging volatile periods as opportunities to expand client relationships and grow market share. On the process for our U.S. listing, earlier today, we filed an amendment to our S-4 registration statement with the SEC, responding to the ninth round of comments. While we can't provide a specific timeline for re-domiciling and listing Galaxy publicly in the U.S., we are very encouraged by the progress we have made to date, and we're hopeful for a quick turnaround from the SEC on our latest submission.
Lastly, on financial reporting, in order to improve alignment of our disclosures with the changing nature of our business, we will be reviewing our financial reporting in the coming quarters and plan to update the operating business segment breakout and other financial disclosures in order to improve transparency and understanding of our company and operating performance. We look forward to sharing more with you on this in the coming quarters. With that, let me turn it over to Chris.
Chris Ferraro (President and CIO)
Thanks, Tony. All right, let's talk data centers. Across our data center footprint, our mission has always been very clear: to monetize our electrical capacity and infrastructure in the most profitable way possible for shareholders.
At the heart of this strategy is Helios, Galaxy's flagship data center campus in the panhandle region of West Texas, where we have spent the past two years developing and operating industrial-scale Bitcoin mining infrastructure and providing turnkey colocation services to other Bitcoin mining companies. Today, I'm excited to share the next phase of our digital infrastructure evolution, a strategic expansion into AI and high-performance computing to meet the increasing demand for highly reliable power and data center infrastructure to support accelerated AI growth within the United States. Galaxy has entered into a transformative 15-year lease agreement with CoreWeave to power the next wave of artificial intelligence innovation. As part of this agreement, Galaxy will initially deliver 133 MW of critical IT capacity to host CoreWeave's AI and high-performance computing infrastructure at our Helios campus.
Under peak design day conditions, we expect the facility to operate with a PUE of 1.5, utilizing roughly 200 MW of gross power capacity at our private substation. Upon energization of the full 133 MW of critical IT, we expect to generate approximately $240 million in revenue in the first 12 months. Over the 15-year initial term of the agreement, taking into account annual escalators, we expect to generate approximately $4.5 billion of total revenue or $300 million of average annual revenue. Note this all excludes two five-year extension options that exist in the lease. Operating expenses, apart from expenses associated with Galaxy on-site personnel, will be passed completely through to our tenant. We therefore expect EBITDA margins for the duration of the contract to meet or exceed 90%.
We expect go forward cash costs, inclusive of various financing fees and capitalized fees during the construction period, to total between $11 million and $13 million per MW of critical IT load. Projected CapEx is based on infrastructure design requirements to satisfy high-density direct liquid cooling infrastructure in a true Tier 3 N+1 redundant environment. As discussed in our last earnings call, Galaxy is responsible for financing this transition from Bitcoin mining to AI and HPC infrastructure. We are actively pursuing project-level debt financing and anticipate finalizing agreements within the next one to two months. Additionally, in November, we raised approximately $400 million through an exchangeable note offering, and we expect to allocate a significant portion of these proceeds to fund the equity component of this construction, which is already underway.
We have already commenced the selective demolition of our 126,000 sq ft data center shell and will begin unplugging our mining machines in the second quarter of this year while we continue to explore creative structures to monetize our remaining A6 fleet. The retrofit of Helios will be completed in phases, with the first full 133 MW of critical IT load expected to be delivered throughout the first half of 2026. Beyond the capacity that we currently have contracted with CoreWeave, today we have ERCOT approval for an additional 600 MW of gross power capacity at our campus. Given the immense power demand and low data center vacancy rates across the U.S., we have very high conviction in contracting this additional capacity for AI HPC.
In addition to the 800 MW of gross power capacity we're already approved for, we have an additional 1.7 GW under various stages of load study at Helios, which we anticipate some portion to be approved of in the next several months. CoreWeave is a key player in developing scalable and efficient computing solutions for AI and other HPC computing applications. Their tenants include major investment-grade hyperscalers who are deeply committed to leading in AI innovation, investing hundreds of billions of dollars in infrastructure research and strategic partnerships to drive the future of artificial intelligence and cloud-based computing. CoreWeave selected Galaxy as a partner due to Helios's unique attributes, as well as Galaxy's proven ability to develop, own, and operate digital infrastructure at scale. Our ability to serve CoreWeave underscores Galaxy's evolution into a trusted provider for one of the world's fastest-growing AI cloud service providers.
Now, back to the operator for questions. Thank you.
Operator (participant)
Thank you. As a reminder, at this time, if you would like to ask a question, it is the star and one on your touchtone telephone. If at any point you find your question has been answered, you may remove yourself from the queue by pressing star two. We'll go first to Patrick Moley with Piper Sandler. Please go ahead. Your line is open.
Patrick Moley (Director and Senior Research Analyst)
Thank you. And congrats on the deal announcement. On the last call, you said that the hyperscaler, the term sheet that you signed, which we now know is with CoreWeave, included the option to develop all 800 MW of the approved power at Helios. Just wondering how we should think about the remaining 600 MW. Is that still exclusive to CoreWeave, or can you kind of shop that to other folks?
How confident are you, or what are your thoughts on that additional 600 being added to your agreement with CoreWeave over time, which is somewhat similar to the deal I think they struck with Core Scientific? Thanks.
Chris Ferraro (President and CIO)
Yep. Good morning, Patrick. Appreciate it. You joining the question. Yes. In the term sheet and currently, the remaining capacity at Helios remains under option with CoreWeave, with a couple of different tranches over various time frames. We are between CoreWeave's expansion plans, which we know are very significant given the demand that they see on the client side, as well as continuing conversations with other market participants, which we are allowed to have and continue to have. We're highly confident that that space, as it gets delivered, is going to be contracted.
Patrick Moley (Director and Senior Research Analyst)
Okay. Great.
Then just looking at the deal specs of the deal announced today, how would you expect, I mean, those specs to kind of change in terms of revenue, EBITDA margins on that additional 600 MW if CoreWeave does, in fact, exercise the option to add that to their contract?
Chris Ferraro (President and CIO)
Yep. I think a couple of things, right? The actual development CapEx side, let's start there. At our site, phase I, which was the first contract at least with CoreWeave, was a retrofit of our existing shell for Bitcoin mining. Whereas the rest of the campus development is a greenfield development.
There is some extra civil work and structural work and design planning that has to go in at a baseline for the rest of the campus, all very doable, all not significant in the grand scheme of things, but would add to a development budget for the rest of the phases. The other thing that is at play is, as technology develops, the design specs for what future data halls for various customers would be is likely to change: higher rack density, etc. That part is still unknown. The thing I would point to, though, is the arrangement that we have with CoreWeave in phase I as well as for under the options, provides for Galaxy to earn a fixed deal based on CapEx costs.
It is really a collaborative partnership with these potential large customers, CoreWeave most significantly, to spec out what they ultimately need from a design requirements perspective. Whatever that resulting cost is creates a function of what Galaxy is going to earn from a yield perspective.
Patrick Moley (Director and Senior Research Analyst)
Okay. Great. That's it for me. Thanks, guys.
Operator (participant)
Thank you. We'll take our next question from Brett Knoblauch with Cantor Fitzgerald. Please go ahead.
Thomas Shinske (Associate)
Hi, guys. This is Thomas Shinske on for Brett Knoblauch. Thank you for taking my question and congrats on the quarter and deal announcement. I guess maybe to start on the 90% EBITDA margins, I know you highlighted that a lot of the operational and maintenance expenses will be passed through there. Is there any additional expenses to highlight maybe being passed through there, maybe power costs?
I know that's something that's been brought up in deals across the space. How confident you are in reaching those 90% EBITDA margins on a go forward basis?
Chris Ferraro (President and CIO)
Yeah. Good morning. Thank you, Thomas. Thanks for joining. The vast majority of all expenses are passed through to the tenant, in this case, CoreWeave. That includes electricity, power costs, other utility costs, operations and maintenance. The only costs, in addition to the CapEx, which we are funding and financing, and the ensuing ongoing carry costs of that financing, the only costs that Galaxy incurs are Galaxy personnel who are going to remain on site. Given that, we have pretty high visibility and pretty high confidence in the operating model and therefore the ability to achieve 90% plus EBITDA margins.
I hesitate to call it a full triple-net lease, but it's triple-net-ish in structure.
Thomas Shinske (Associate)
Perfect. Very helpful, color. Thank you. On the additional 600 MW options, I guess I know you highlighted CoreWeave has some options there. Are those on a ROFR basis, or is the shopping pretty free for other potential hyperscaler tenants?
Chris Ferraro (President and CIO)
I would say that we only have so much time and a team that's so large. We are spending a lot of our focus and time working with the great team at CoreWeave. Those agreements are not on a ROFR basis. However, we are not restricted from talking to folks in the market and understanding what their forward needs are going to be, particularly as we think about expansion beyond just our existing approved 800 MW.
When you look at our expectation for what our actual full campus availability is going to be, you start talking about power delivery not just in 2026 and 2027, but 2028, 2029, 2030, and beyond through the whole additional 1.7 . Those conversations, many people in the marketplace with very large aspirations and very large balance sheets are very interested in thinking about their forward planning for those out years. Our business now, in addition to executing on near-term power and data center delivery, is also about business development and building a smart pipeline with great customers for that very long-term build for the entire campus.
Thomas Shinske (Associate)
Perfect. Thank you. If I can just sneak in one more.
On the $11 million-$13 million per megawatt CapEx guide, are there any upfront investments that will need to be repeated for future deals, potentially lowering the per megawatt CapEx for the remaining 600 MW?
Chris Ferraro (President and CIO)
Yeah. There is going to be a good mix and balance there. For example, the fiber investment we are making is a long-haul fiber investment to deliver data streams back and forth to Dallas at 10 milliseconds sub. That, for example, is a project that is going to serve the entire site, not just the first quarter that we have contracted today that we are building. The other element that is going to be at play in the CapEx on the go forward is also speed to delivery, right?
Phase I, what we committed to in concert with CoreWeave was a pretty aggressive delivery schedule to meet their demand needs in 2026 for capacity. With that includes expediting equipment acquisition, which comes at a price. As we plan out forward, late 2026, 2027, 2028, delivery of data center halls, we have a lot more flexibility in terms of capital equipment acquisition from a price negotiating perspective, offset by the fact that, like I said, there are going to be technical design requirements that I think are going to change in terms of data center design for rack density as technology evolves. There is a lot that goes into there. Yes, there are components of what we are building here that are leverageable across the whole site.
Thomas Shinske (Associate)
Awesome. Thank you a lot. And congrats again.
Chris Ferraro (President and CIO)
Thank you.
Operator (participant)
Thank you.
We'll go next to the line of Mark Palmer with The Benchmark Company. Please go ahead.
Mark Palmer (Managing Director and Senior Equity Research Analyst)
Yes. Thank you. And I'll add my congratulations on the deal. Can you talk a bit about the process that led to the signing of the term sheet? How many other counterparties beyond CoreWeave were involved? Are those counterparties potentially in the mix for the remaining $600 million or 600 MW that would represent the next phase in the deal?
Chris Ferraro (President and CIO)
Sure. The process prior to and we signed the term sheet with CoreWeave. God, I don't even recall back when. In October of 2024 now. The process leading up to that, I think, was pretty robust. We hired an advisor, Newmark, who not only worked on us on the term sheet and the lease, but also helped us tap the market and talk to all the market participants.
It was a pretty robust process, all things considered, given that the large buyer base for scale power infrastructure, when you're talking hundreds of megawatts, is not in the hundreds of potential buyers, just in terms of actual size. All of those players we are currently in contact with. Like I said before, we're very focused short and medium term on working with CoreWeave because we think they're awesome. We think they have a—they're very smart about this. They're fast-moving. They're very diligent, and they have a pretty big pipeline of demand.
As I said, sort of as we think about unlocking the incremental capacity for the Helios campus that could make us a 2.5 GW mega data center campus over the long run, the conversations with folks who would take up that size of power over longer term periods, 2028, 2029, 2030, 2031, those are the conversations that we are also having today.
Mark Palmer (Managing Director and Senior Equity Research Analyst)
Just one more. If you could shed some light on what the time frames are for CoreWeave's options with regard to the remaining 600 MW, I think you said that that would be in different chunks. Thank you.
Chris Ferraro (President and CIO)
Yep. That is not something we are going to disclose right now, but I would say that demand for locking up our capacity is pretty high.
It is important to us to have clarity on the timelines that we are going to develop the sites because we have a lot of work on our end, not just planning and building, but raising financing and staging out our own balance sheet appropriately so that we can responsibly develop the site. We are very clear there are not long-term options past 2025 to have the capacity locked up. We will be pursuing it diligently and working with CoreWeave and, if not, other people in the marketplace to know that we have what the forward plan is for existing capacity.
Mark Palmer (Managing Director and Senior Equity Research Analyst)
Thank you.
Operator (participant)
Thank you. We will go next to the line of Joseph Vafi with Canaccord. Please go ahead.
Joseph Vafi (Managing Director of Equity Research)
Hey, guys. Good morning. Nice to see the Helios announcement and strong performance across the business. Just maybe one for Mike real quick.
Spot prices are down a little bit, but the medium-term outlook is looking pretty bright, I think, given maybe a better regulatory backdrop and other things. Just wondering how you're looking at the M&A market now for Galaxy. Is this an interesting time to be looking at things? Are there any areas of the business that you think that would benefit from some additions? I have a quick follow-up.
Mike Novogratz (Founder and CEO)
Yeah. Listen, it's a great question. I would tell you that for us, what has stopped us from M&A in the past is that our stock always felt very undervalued to us. I'm still of that belief when I do some of the parts balance sheet plus what I think the value of our crypto company is plus the value of what our data center company is becoming.
I just think our stock's too cheap to use it at this point to buy crypto companies. I'm hoping that once we get public in the United States and once this data center build-out has more clarity to the market, we trade at a value that will allow us to gobble up some really talented groups of people out there. I would tell you a lot of private crypto companies probably want to sell because they see TradeFi companies coming in with big balance sheets. Most crypto businesses, fortunately or unfortunately, are tied to the overall index of crypto prices. If you're a staking company, when the price of Solana and Ethereum goes down, your revenue goes down. I think there's probably a lot of sellers in the private company space.
We're not a buyer at this point of anything significant because of the level of our currency, our equity. That does not mean we're not looking and turning over a lot of stones and coming up with plans for when our currency gets to a level that we think it's worth trading.
Joseph Vafi (Managing Director of Equity Research)
Sure. Thanks for that, Mike. We just kind of drilled down a little bit on funding the CapEx for Helios. Sounds like it's coming along pretty nicely. Any other details there, kind of what the cost of capital might be, and are those kind of baked up financing arrangements yet, or you're still shopping it? Thanks a lot.
Chris Ferraro (President and CIO)
Yeah. We've had the benefit and luxury of a pretty robust financing market, in particular focused on this space, which is great.
We have spent a lot of time running a pretty robust process with dozens of lending groups who we know very well from past lives and current. We are already past the phase of sorting through them. We have selected a very strong lending group who have great reputations, who were highly confident in they understand the asset, they understand us, and we are moving forward together on finalizing the financing. I think we have talked about this a little bit in the past. The financing market for these assets is pretty well-known and pretty straightforward on the project side. We will see where it lands, but you should think about it as like 80/20 debt to equity financing, roughly. The cost of capital for project financing is sort of 10%-11% yield.
Now, that's really for the construction phase, which for us in phase I really is a 12-month expected process or less. The way we think about how Galaxy is going to finance this asset on a go forward basis is think about it 80/20 debt to equity is what we've currently circled for the project financing part. Once we reach COD, commercial operating date, and rent starts flowing, there is a whole new market open to companies like us that have stabilized assets that are paying rent, which our expectation is going to allow us to refinance the project debt for lower cost, likely higher quantum debt, which should unlock a lot of equity for Galaxy to then reinvest in building the campus.
Matching up timing and having capital markets be open and vibrant for all that to happen is what we are looking at and what we do really well and what we are focused on nailing. That is how I would think about the evolution over time of phase I, phase II, phase III, and sort of debt equity mixes and how we are going to finance it. Is that helpful?
Joseph Vafi (Managing Director of Equity Research)
Yes. That is great. Thanks a lot, Chris.
Operator (participant)
Thank you. We will take our next question from Mike Colonnese with H.C. Wainwright. Please go ahead.
Mike Colonnese (Managing Director and Senior Crypto and Blockchain Analyst)
Hi. Good morning, guys. Congrats on the CoreWeave deal today. Thank you for taking my questions.
As we look across the crypto ecosystem and the regulatory tailwinds that are expected to come from legislation in the U.S. this year based on Mike's comments, what major themes do you see playing out as we progress through the rest of this crypto cycle? Where will Galaxy be focusing its resources in 2025 to best capitalize on some of these things?
Mike Novogratz (Founder and CEO)
Yeah. That's a great question. I think the first obvious answer is stablecoins, right, both here and abroad. Stablecoins for payments, stablecoins for Internet of Things, stablecoins for cross-border transactions and remittances are only going to accelerate. There is a lot of activity in D.C. right now on this stablecoin bill that is being worked on. On our venture side, we're investing, and Chris could go into some of the companies, in plenty of businesses that operate in and around that area.
The second kind of obvious one is, and what everyone's kind of waiting for, stablecoin is really the first real-world asset that's been tokenized, if you want to think about RWA, but the tokenization of other stuff, if it's mortgages or collectibles, equities, is coming. I keep saying it's going to be like bankruptcy. It happens real slow, then real fast. It hasn't happened fast yet, but every single TradeFi participant is gearing up for that. I think the crypto world is working on that. What's interesting, it's really an infrastructure play. One of the early promises of crypto was it was going to make things more efficient, more equitable, and more transparent. A lot of that was held back by the lack of regulatory clarity that we had for the last four years, certainly, but really eight years.
As that's changing, I think you're going to see the acceleration there. We are close with our euro stablecoin called AllUnity, which we're doing with DWS, the Deutsche Bank subsidiary, and Flow Traders, which is going to be bought and approved and regulated. That'll be our first actual stablecoin play. In every other stablecoin, we are market-making in them. We are promoting. We're very much, from our trading business, in the heart of the stablecoin ecosystem.
Chris Ferraro (President and CIO)
Yeah. The thing I'll add to that for a little bit of color is I agree with Mike on the focus for 2025 in the U.S. from a regulatory perspective on stablecoins and other tokenized assets and market structure.
Rather than us just being an investor, market participant as a trader, maybe an issuer, rather than us just do that, I'll highlight the GK8 investment we made back in late 2022. It was actually early 2023 when it closed, was a way for us to not just be a participant, but actually be an infrastructure technology provider underneath what we thought was going to be this wave coming. Now, it's been a couple of years since it's actually come here to the U.S.. Luckily, that business has been able to make real inroads. I think we have 30 clients now in the book, mostly internationally.
The U.S. is going to be a sea shift change in terms of customer size and adoption for the technologies, both on the custody side as well as the tokenization side that we've been able to build and enhance since we bought GK8 here. I think while it's taken a while for it to really catch on, I wouldn't underplay how excited we are about the opportunity set for our actual infrastructure technology to be a real part of the financial infrastructure as it changes here in the U.S. going forward. We're pretty excited about that.
Mike Colonnese (Managing Director and Senior Crypto and Blockchain Analyst)
Really, really interesting, color. Thanks for that. I'm just curious to get your guys' views on a U.S. strategic Bitcoin reserve.
What's really the probability you see of the U.S. to start to actively acquire more Bitcoin this year and what the implications could be if that were to happen with other nation-states adopting similar strategies?
Mike Novogratz (Founder and CEO)
Yeah. Listen, I think it's higher than you might think solely based on the spectacular enthusiasm of many people around that Trump cabinet, right? Howard Lutnick, who is one of the major voices in that cabinet, has been a vocal proponent of both stablecoins and Bitcoin. He already has the ability, given the executive order within his perch, to, on a revenue-neutral basis, buy more Bitcoin. I think there's a shot higher than I would have thought six months ago that the government actually ends up buying some Bitcoin.
That is a really significant statement if it happens because it certainly will force other governments to think, "Hey, we should do that too." I mean, I think I said this on the last earnings call. I had direct conversations with one of the larger sovereign wealth funds who said, "Now that the U.S. is buying, we are buying." I think you are going to continue to see some of that. I do not think that happens until the third quarter if it happens because just getting the plumbing set up and the buy-in. I tell you, there are a lot of people in D.C. focused on it. It is quite frankly why even in a world where crypto assets are heavy, it is dangerous to be short.
This gap between crypto price action and all the exciting things we're seeing is somewhat explainable in that a lot of the exciting things are around infrastructure. Connecting that infrastructure back to building on blockchains and using those blockchains so they start getting more activity and more support from the market is just going to take time. We are in this gap between the promise of what could happen and people's enthusiasm and a risk-off environment for Ethereum or any of the other L1 or L2 blockchains. In time, they will be used a lot more than they're being used today. That is just the gap. Figuring out fair valuation versus future use is going to be the calculus and the science. Maybe it's art, maybe it's science to getting the markets right.
Operator (participant)
Thank you.
We'll go next to the line of Joe Flynn with Compass Point Research.
Joe Flynn (Senior Research Analyst)
Hi. Thanks for the question. I was wondering in relation to the construction CapEx and timelines. It was expected to roll on in stages in 2026. Can you maybe walk us through what that ultimately looks like? From the mention of the fixed cost on CapEx, is that more a function of the initial deal, or is that kind of the level called 15%-17% that we should expect on a go forward basis for the additional capacity? Thanks.
Chris Ferraro (President and CIO)
Hey, Joe. Good morning. Yeah. Let me hit timeline first. We're already underway a number of different subprojects in the development and spending CapEx and building. Right now, the anticipated delivery date for the individual data halls for phase I begins in early 2026 and should be completed into Q2 2026.
Think of 2025 as construction development and very quickly in 2026 delivering operating data halls and then flowing rent for the next 15 years-25 years, depending on options. From an actual use of proceeds and CapEx perspective, it is fairly straight lined over the year. If you think about equipment procurement, they tend to be 10%-20% down deposits and then delivery over a number of weeks and then final payment schedules once you take delivery and then installation. Same thing with the EPC contractor. That is really timing for phase I. Phase II and phase III, just to be specific about that, electrification of phase II is expected late this summer. That means phase II will be able to flow power late this summer.
When we think about development time for the actual data center that will be phase II, you should think about that sort of layering in towards the later stage of phase I project and then for 2026, basically having rolling delivery of data halls all the way up through basically 600 MW of gross capacity as currently planned. Phase III, which is the final 200 MW of gross capacity currently approved today, that's an early 2027 electronification of the substation. You should think about a similarly overlaid project timeline from a CapEx and delivery standpoint, knowing that you can't actually flow power until early 2027 anyway for the last phase. Your second question, can you just clarify what you meant on the CapEx side so I can answer appropriately? I might have lost you, Joe. I'm not sure what the second question was on CapEx.
I'm happy to hit it if you want to clarify it.
Joe Flynn (Senior Research Analyst)
Yeah. Sorry. It was more so related to the lease. So it was the percentage yield on CapEx structure. Is that something we should expect going forward? Is that more of a function of the first deal? With that, is there any escalator or anything over time, or is it expected to be largely fixed?
Chris Ferraro (President and CIO)
Yeah. I think these deals historically in data center world have gotten done in two different ways. Either sort of an open book percentage yield on CapEx, where you're working in concert with an end tenant to purpose-build a data center design spec with a CapEx that kind of falls out of it, and then you have an agreed-upon gross yield based on that. That's the construct that we've used largely here with CoreWeave.
I would expect, given that these end use cases are relatively new, that to be the construct going forward for our business because these have not been replicated multiple times before such that a developer like ourselves would or probably should on spec say, "We'll agree to a rent number with an unknown CapEx budget and unknown design spec." Where the market is today for developing these new data centers, they really need to be a collaborative design process with a resulting sort of agreed economic yield as a result. I would expect that to be very similar, at least for the foreseeable future as we develop the site.
Joe Flynn (Senior Research Analyst)
Thanks. That's good, color. Real quick, maybe digging in a little bit on that.
I guess going back to the initial term sheet and then ultimately the signing, were CoreWeave's end customers involved in the overall design of the project? Was it something that they had to sign off on and ultimately kind of how the deal came about, or is there a bit of separation between your guys' agreements with CoreWeave and ultimately whatever they sign?
Chris Ferraro (President and CIO)
Yeah. Our client is CoreWeave. CoreWeave is our main partner in all things economic, developing the site, specs, etc. Obviously, on the other side of that, one of CoreWeave's expertise is knowing who their clients are and what their clients want today and what their clients want tomorrow.
The design spec that we ultimately land on for phase I, phase II, phase III, it's a direct collaboration with us and CoreWeave, but it is definitely the result of demands both from CoreWeave to the extent that they're running their own proprietary cloud as well as what they know evolving spec needs are for other large hyperscalers in the world.
Joe Flynn (Senior Research Analyst)
Great. Thanks.
Operator (participant)
Thank you. We'll go next to the line of Bill Papanastasiou with KBW.
Bill Papanastasiou (Director of Equity Research)
Good morning, gentlemen, and congrats on a strong quarter to close off the year. For my first question, with the repeal of SAB 121 following the election, can you provide an update on how the team is progressing with the GK8 custody software licensing? What's the reception been? Seems like that's a bit of a low-hanging fruit given the tech. Curious how you guys are looking at this. Thanks.
Chris Ferraro (President and CIO)
Yeah.
Hey, Bill. Very excited about the opportunities that are in front of us on that front. I would characterize SAB 121 repeal rescission as a major step, but not the only step necessary to allow TradFi institutions to custody and touch crypto. SAB 121 was a surprising hurdle that was put in place that has now come down. There are other things being worked through on the market structure side on capital charges with a number of other government organizations that the biggest financial institutions in the world are actively working to today in Washington to sort out and figure out, but need to get sorted out first. There is more work that has to get done throughout 2025 on market structure and on stablecoins from a regulatory perspective than just SAB 121. That is a major hurdle that came down.
The timeline is still elongated in terms of figuring out when these institutions are going to come in and how and what they're going to offer. All of that is fair game for GK8. Outside the U.S., I'll keep pointing back to as we think about getting approval this year, which we're very confident in given our partner in Deutsche Bank and DWS, BaFin approval for the first regulated euro stablecoin, that company, even though we're a minority owner, ran a pretty robust process and ultimately selected GK8 from a technology perspective to be the technology backbone for issuing and maintaining smart contracts around stablecoin issuance, mint burn, etc. That is a thing I would point to as a tangible example of once the regulatory window gets clear and there's opportunity, GK8 provides a technological infrastructure solution that is a clear product-market fit.
There is still a little work to be done in the U.S.
Bill Papanastasiou (Director of Equity Research)
Appreciate that, color. I am curious, was there any revenue prepayment attached to the announced lease agreement? In addition to that, how quickly do you think it would take CoreWeave to execute on the remaining 600 MW option? Thanks.
Chris Ferraro (President and CIO)
For phase I, there was no prepayment for phase I. I will say that much. If I know the CoreWeave team, which I think we do really well at this point, they have a very large pipeline and are growing very fast. One of the key choke points for them is ensuring that they have the adequate forward data center capacity to fulfill the demand that they have on the other side.
My expectation is that we are going to continue to move pretty quickly at figuring out forward partnership and development of the campus.
Bill Papanastasiou (Director of Equity Research)
Thanks. Just one last question, if I may. How would you benchmark the latency specs between the Helios site in West Texas to the municipality of Dallas?
Chris Ferraro (President and CIO)
I would say we purposefully specced the fiber project to ensure that we had between 10 ms-15 ms travel path back to Dallas. We have two redundant paths back to maintain N+1 connectivity. That benchmarking was meant to be speeds that can service both not just training, but also inference loads.
The idea there was to ensure that even though Helios is in a relatively remote area, that it actually can service and is a good fit for what we think the forward on actually AI load use cases are going to be so that it's in play for many years to come.
Operator (participant)
Thank you. We'll go next to the line of Martin Toner. Please go ahead with ATB Capital Markets.
Martin Toner (Managing Director of Institutional Equity Research)
Good morning. Thanks for taking my call and congrats on this deal. How much CapEx? What type of redundant power will you use at site? How did that figure into the CapEx number if you can give us some detail there?
Chris Ferraro (President and CIO)
Yeah. I think the short answer is we're building phase I and future phases to a true Tier 3 N+1 redundancy across really all systems.
That is a big component of why the CapEx estimate is between $11 million and $13 million per critical IT MW. It is because every system from cooling to key critical IT, even down to the substation and fiber, has redundancy. This is the big difference from going from Bitcoin mining to two or three AI data center capacity. Yes, that is the major driver.
Martin Toner (Managing Director of Institutional Equity Research)
Great. Thank you. I think at the beginning of the call, you mentioned that the 1.5 PUE was like a peak PUE. Just wondering if you could kind of talk us through any dynamics there, and hopefully I did not mishear that.
Chris Ferraro (President and CIO)
Yeah. No, you did not mishear that. I think the idea is just that we need to ensure that there is enough gross power available at peak times.
Whether that's at extremely hot days in the peak of the summer in Texas or not, that we have and depending on expected workloads of the underlying critical IT, which is a function of an expectation of CoreWeave's usage and their ultimate end client's usage, the design takes all that work into consideration to ensure that we can effectively power 133 MW of critical IT with N+1 redundancy without having any downtime. The net output is we expect to have up to 200 MW of gross power effectively ring-fenced and dedicated to power the 133 at peak workload, peak time of day, etc.
Martin Toner (Managing Director of Institutional Equity Research)
Excellent. Thanks very much. That's it for me.
Chris Ferraro (President and CIO)
Yep.
Operator (participant)
We would like to turn the call back over to Mike Novogratz. Please go ahead and foreclose your remarks.
Mike Novogratz (Founder and CEO)
Guys, thanks for all the questions.
As you can see, we're very excited about the opportunities both in the data center world and in the crypto world. We're happy to have the AG thing put behind us. Listen, I think the next couple earnings calls, you're going to start seeing a company that feels more mature with two strong business lines, and we look forward to it. That's all I got. Have a great day.
Operator (participant)
We would like to thank everybody for their participation on today's conference. Please feel free to disconnect your line at any time and have a wonderful day.