Welcome, welcome, welcome. Thank you very much for coming to the HCW conference this year. This is the most highly anticipated event of all of our crypto stuff, the Bitcoin mining panel. Of course, we're going to talk about HPC and how business models evolve. And it's my honor to introduce Russell Cann, who will be moderating the panel. I used to do it a long time ago, but it's way above my pay grade now. And you've got a better man here. I've known him for years covering Core Scientific. He's phenomenal. This is Core Scientific, one of the many companies he's founded. He's still a chief development officer there. And most importantly, he's probably the best MC at any pool party. Take it away, Russell.
Thanks, guys. We got a short time here because I'm between you guys and President Bush and all that, but I want to get started. I'm just going to kind of go down the list on this. We have a couple of questions that everyone's going to answer, then we'll dive into specifics around the business models each company has here, but first off, just please give a quick introduction of yourself, your company, your capacity in the company. Talk about your specific operating strategies, your company was founded on, and how that has evolved over the years, either because of new tech or changes in tech or changes in strategy, and then just if there's been any kind of things around HPC demand, how those kind of things might have changed as well, but just a quick introduction of it. Matt, start us off.
Thank you, Russell. Great to be here, everybody. My name is Matt Schultz. I'm the Co-founder, Chairman, and CEO of CleanSpark. We operate 50 exa hash across 33 data centers in four states. Our business is interesting because we joined, we got into the Bitcoin space through energy. We owned a series of portfolio patents on distributed energy resource management. So we built microgrid solutions and worked in conjunction with military bases and embassies around the world, and so that kind of fundamental working knowledge of energy really helped us in the Bitcoin mining space because, as everyone knows, that's our biggest input. As far as evolution, I think we fancy ourselves as experts at land and expand, and we identify opportunities in utility markets regionally and build relationships and grow from there.
What that's presented us with is a portfolio of assets that have tremendous value for the Bitcoin mining ecosystem, but also lend value to potentially other types of computing.
Thanks.
Good afternoon, everybody. Salman Khan, I'm the Chief Financial Officer for Marathon Digital Holdings. Marathon has evolved over a period of time, to use the term that you use, and we own and operate approximately 60 exahash today and control about 1.7 gigawatt worldwide. We operate in four continents and have 15 data centers worldwide. Our story has evolved over a period of time. We used to be an asset-light company just two years ago, and last year, we ended up buying 800 megawatts at half the cost of build multiple, and as a result of that, we converted our asset-light model to 70% owned and operated, thereby reducing our operating costs or our electricity costs per coin as one of the lowest in the sector. While we did that, we did not stop there. We also bought generation of electricity.
We are not a utility company, just to be clear, but we like when we can marry intermittent power that is low cost and combine that with Bitcoin mining. And that makes it so much, you know, after every four years of having, we don't have to worry about the costs and the global hash rate when we marry the Bitcoin mining with wind farms, for example. We own more than 100 megawatts generation capacity in Texas. We also recently announced investment in Exaion. Exaion is a subsidiary of EDF, one of the world's largest clean energy utility company. And they're focused on as a Tier 4 data center operator. They have sites in Canada and in France. And their model is primarily focused, very different from traditional AI colocation service here in the U.S. They're focused on sovereign or edge compute. We're very excited about that.
That's expected to close by the end of this year. All in all, a global footprint, great people that we've been able to attract to our company, and very excited to be here in this room with some awesome panelists here.
Okay. Tyler.
Hi, everyone. I'm Tyler Page. I'm the CEO and founder of Cipher Mining. We are a data center development company. We began life as a greenfield developer of Bitcoin mining sites about five years ago. We secured power contracts and built from dirt our own first five data centers where we are operating 477 megawatts for Bitcoin mining today. That produces about 23 exahash. We have typically the lowest power costs of any company in the space. We've been very focused on securing low power costs in that business with a very efficient fleet.
What we have found ourselves in the middle of over the last, I'd say, year or so, and we're not unique up here on this panel, is a position where we were securing and developing large power interconnects for years before that became to be in such high demand because of the rise of large language models, HPC, and the desire for hundreds of megawatts at a single data center. I'd say about a year ago, we really aggressively positioned ourselves on the thesis that we could have very large interconnects in more remote locations, and the traditional HPC data center world would migrate to us. We're in the middle of that right now. We have 450 megawatts that is ready to go with a substation currently electrified, waiting for tenants. We're having lots of discussions around that.
So I think when we do this meeting next year, we'll have a very robust HPC business to discuss. And it is complemented very well by our existing mature Bitcoin mining business.
All right. Brian?
Sure. Thanks, Russell, and thank you, H.C. Wainwright, for throwing this conference. I've had a lot of very valuable conversations over the last couple of days. I co-head the data center business at Galaxy. Galaxy is a little bit of a unique company in that we have two large operating businesses underneath the parent company. The first and the one that we traditionally were better known for was the digital asset business. This business has a large global markets business, which includes a derivatives business, a market-making business, a very large trading business generally, and a lending business, as well as sort of a boutique advisory firm.
We also have an asset management business, which has about $6 billion in AUM and a staking business that has another $3 billion, a little bit more than that, on platform under stake for a little over $9 billion in assets on our platform. The other business, which is a little bit newer, is the data center business, and that's what I'm the co-head of. This came out of our Bitcoin mining business that we've been operating for several years. In particular, we focus on our asset Helios, which is in the Panhandle region of West Texas near Lubbock. We made the decision early last year to transition that business from a Bitcoin mining business to a traditional data center business.
And we signed first one lease agreement with CoreWeave covering about 200 megawatts of gross capacity, followed by a second lease with CoreWeave, which was an incremental 400 megawatts of gross capacity. And CoreWeave has also committed for the final 200 megawatts of the 800 megawatts that we have approved for at that campus. So we're currently building out that campus right now. We plan to energize in Q1 of next year. That's the business that, of course, is a little bit more similar to the other types of businesses that my panelists up here are also running.
Thank you. Asher?
Cool. How's it going, everybody? My name's Asher. As of, I guess, last week, I help run two publicly traded companies. One is Hut 8, which I'm the CEO of, and the other one is American Bitcoin, which I'm the Executive Chairman of, so Hut 8 is a business that we had merged. We founded a company called US Bitcoin Corp. We merged into Hut 8, and then I took over in the leadership role beginning of last year. When we started kind of the business, the thesis behind Hut 8 is that as technologies continue to develop and advance and kind of push humanity forward, their reliance and their consumption of power has ever increasingly gone up, and so how do we build a company at the intersection of energy and of technology? And how do we harness an electron to push advanced technologies forward, to push humanity forward?
And so at Hut 8, we see ourselves as what we call kind of a next-generation energy infrastructure platform. And how do we build campuses for any net new technologies that have large demands of power? So that was where we started, which was kind of Bitcoin mining and Bitcoin compute and securitizing the Bitcoin blockchain. Today, obviously, like many of the folks up here, tons of demand around traditional Tier 3 AI data centers for AI computing use cases, primarily training inference in some locations as well. I think for us, a big inflection point last year was when Coatue invested into the company and our institutional shareholder base kind of grew exponentially at that time. And we started having really two groups of shareholders. One that believed in kind of the long energy data center thesis, and the other one that was kind of core to Hut 8.
Hut 8 was one of the first publicly traded companies that held Bitcoin on the balance sheet. I think the first, the old CEO likes to tell me, but like, and so you have this kind of core Bitcoin audience and shareholder base, and what we realized was that those businesses ultimately have very, very different types of capital, cost of capital. At Hut 8, our goal is to drive volatility down and to lower our cost of capital. Whereas on the Bitcoin side, you monetize that volatility via, for example, convertible notes. You see kind of MicroStrategy having successfully implemented that strategy, and the goal is to keep vol, to be able to sell volatility, and so ultimately, we didn't see that these were two kind of asset classes that should live within the same capital stack.
And so we decided to spin out American Bitcoin into its separately run company. And so the way to think about it is American Bitcoin is an anchor tenant of low redundant data center capacity for Hut 8. And so Hut 8 is long energy and American Bitcoin is long Bitcoin. And so Hut 8. Today we have about a gigawatt of capacity under management. 90% of that is contracted. About 30% of that is power generation facilities we own. 70% of that is data centers primarily supporting Bitcoin compute. We have five traditional retail colo data centers in Canada that are smaller as well. And then we have about 1.5 gigawatts of own land and power agreements and net new development sites that we're expanding into right now.
And then with American Bitcoin, we really kind of took a first principles approach of why do investors historically invest in Bitcoin mining businesses? Why are they investing into these treasury accumulator businesses? And I think ultimately it's to get a levered kind of exposure to the underlying asset class. And so we kind of branded the company as a Bitcoin accumulator. It's not looking to be a mining company. It's not looking to be a treasury company. It's not focusing on how many exahash we have. It's not focusing on how many Bitcoin we have. But the ultimate goal is to increase Bitcoin per share. And that's kind of the ultimate metric. And we have a multi-pronged strategy in doing so. And so I think it's been a fun journey. And we'll talk a lot about the data center side on the panel today.
Thank you. Sam?
So my name is Sam Tabar. I'm CEO of Bit Digital. I'm also CEO of White Fiber. A couple of years ago when ChatGPT came out, we had sort of an epiphany moment that this is going to be huge. And we were really turned off by the Bitcoin mining business, especially because of the halving, which is basically your profits are basically cut in half every four years, which we thought was a crappy business model. So we abandoned Bitcoin mining and we focused on HPC. We landed our first cloud customer in early 2024, January of 2024, for a $150 million contract over three years. And we now have 23 customers. Our revenues are about $100 million per year. Backing that is a portfolio of four data centers across Canada and the United States. We also have a partner in Iceland.
That HPC business was doing so continuously so well that we decided to IPO that business just last month, about 32 days ago, and so we IPO'd that. That's called White Fiber, ticker is WiFi, and with respect to RemainCo, that being Bit Digital, we sold all our Bitcoin. We stopped investing in the Bitcoin mining business and we became an Ethereum treasury strategy, and so that's what Bit Digital is now, a pure play ETH strategy, but we do own 71.5% of White Fiber, and for us, our vision is that we think the two largest story arcs of our time are Ethereum and artificial intelligence, and so Bit Digital is an ETH play and owns 71.5% of a very successful company called White Fiber, which is a pure play on AI, and that's the position we'd like to be in.
Thanks, Sam. I'm also, for the sake of time, we're going to, I'm going to combine the next two questions. I want everybody to have a chance to answer them, though, okay? What are your company's strongest assets? Why do you see them that way? Has your opinion of that changed over the last 18 to 24 months? And then how do you think investors perceive your company? Particularly, what's the thesis on why investors, what kind of exposure are they getting to their company? I'll start again with Matt. We'll go down. And then after that, we'll be diving into some individual questions for each of you, depending on your strategies.
You're not making this easy. I would say the strength of our company is the team of people that we've assembled. We're very fortunate in the fact that we have 50 exahash across 33 data centers, just over a gigawatt of power under management, and we're able to manage to the industry's one of the top two most efficient fleets with the top two most efficient uptime. So our personnel, the team that makes that work, is really the lifeblood of what we do, and we've created this opportunity for growth by making commitments to rural communities throughout the United States, where previously there may have been a textile manufacturer or some other industry that has subsequently been offshored, and so these towns that bonded their share of development and distribution of electricity are now sitting on unmonetized megawatts.
So we create relationships whereby it's mutually beneficial for both the city, who's our utility, and for our company. And that really has grown over the years as a result of the relationships that we've developed. So how have I seen that change? Well, we're unique, I think, amongst all of our peers in that not only are we fully vertically integrated as a self-miner, we own and operate our facilities and 250,000 Bitcoin mining machines. But we also have, I think, the top six or seven largest treasuries of Bitcoin of any publicly traded company. And it's important to note that we mine each and every one of those Bitcoin. We've never borrowed money to buy Bitcoin or sold equity to buy Bitcoin. So we've got a very strong balance sheet.
We put up a convertible bond a year ago, $650 million, 0%, converts up 100 with a share buyback, so we've been very disciplined with our capital strategy, and we like to say that we operate at the intersection of capital efficiency, capital stewardship, and operational excellence, and so our people are really key. Now, how do I see that changing? When we entered the Bitcoin mining space, our first two locations were in metropolitan Atlanta. We have College Park, which has a total of about 100 megawatts of capacity, and then a second facility in Norcross, which is about 20 megawatts of capacity. Both of those facilities, when we acquired them, were traditional data centers, so we actually entered the Bitcoin mining space through acquiring operating data centers and subsequently fired all of our customers to convert those to Bitcoin mining.
The opportunities have now changed because we've found ourselves with the ability to deploy very rapidly. We secured 100 megawatts in Cheyenne, Wyoming. And the other bidder on that project was Microsoft. Well, why on earth would a utility pick a company like CleanSpark over Microsoft? And the simple answer was we were able to monetize those megawatts within 60 days, where to build out a proper data center to Microsoft's level is a four-year process. So what we found is that we can enter these jurisdictions, we can put up Bitcoin mining, and now we have a unique opportunity with the geographic diversity that we have that we can now create meaningful relationships with other data center development companies to build data centers while we're monetizing the megawatts through Bitcoin mining.
And if and when the time comes to shut down the Bitcoin mining assets to flip those two traditional data centers, high-performance compute, artificial intelligence, whatever the case may be, we can then duplicate those efforts. Now, in Cheyenne, Wyoming, we went from scraping the ground to having six-megawatt immersion-cooled pods deployed and hashing inside of six months. So that efficiency and speed to market is a huge differentiator. And through some of the relationships that we've developed, it's been very clear that there's demand for that because there are megawatts that have been paid for that exist that they're doing nothing until they're monetized. So we find that it's a very unique relationship.
The last thing I'll say about that question, when we had the opportunity to attend the Proto Miner launch at Russell's facility in Dalton, as we left there, the chairman of the local utility grabbed Gary and Harry and myself and wanted to have a conversation because obviously there's a ton of demand for compute energy. The challenge that they're seeing in the state of Georgia particularly is that they forward sold power. With all this increased demand, I don't think they contemplated five years ago that all these megawatts that were previously allocated for textile manufacturing would be soaked up very rapidly by compute opportunities. What they're now lacking is the flexibility or an interruptible load.
In many of these jurisdictions throughout Georgia, we have a distinct advantage because we can provide up to 200 hours a year of power back to the utility based on demand, allowing them a lot more flexibility, acting as a shock absorber. Now that we're seeing a tremendous amount of inbound inquiries, specifically for the Metro Atlanta facilities as an example, we have a unique position in that we can blend traditional compute, high-performance compute with Bitcoin mining to provide that shock absorber, that interruptibility on the network and provide a mutual benefit to the community. I think I answered all those questions, but we're really excited about what the future looks like for the combined enterprise.
Thank you. Salman?
We are with the 1.7 gigawatt capacity that we own and control today and the multi-gigawatt pipeline. We believe the biggest asset for us is a combination of the electrons that we own and control that can be utilized for Bitcoin mining today or AI in the future. Our goal is to maximize our dollar profit per megawatt hour, whether it's Bitcoin mining or whether it's AI inference on the edge. That's one aspect of it. I want to reflect a little bit on what Matt mentioned about people. People is one of the most important things for us. We've been fortunate to have attracted one of the best talents in our sector when it comes to our management team, when it comes to people in the field, in operations who are running on a day-to-day basis, and also at the board of directors.
We're fortunate to have attracted some really serious talent. Our company is not run founder-run. We are all hired help. And all of us have done some really interesting things in our prior lives, whether it's technology industry. Our CEO spent 40 years in tech. I spent two decades in oil and gas and renewable energy and energy transition space. And the resume continues across the board. And the people is an important part of our story. Why do I say that? It's not just the electrons. You need smart minds to convert that electron into value. And that's where the real value is created for us, where you take smart people, for example, to give you an example, somebody mentioned this is a difficult industry. I think Sam, you mentioned a few moments ago, it is a difficult industry to operate in. Nothing's easy.
It's not easy to make money, especially when every four years there's halving happening, so how do we plan that from a long-term perspective? Creative minds coming together. We looked at wind farms and we found opportunities to acquire wind farms for cents on the dollar because of the market conditions where it exists today with wind farms, and the marginal cost to produce those electrons is close to zero because all you need is a person or two to monitor the wind farms. When the electricity is generated from those wind farms, we consume that electricity for generating Bitcoin. Yes, it's not 100% 24 hours a day, but it's marginal electricity at almost zero cost to produce, and those are the kind of innovative ideas that come out not just from electrons, but having smart people in the room.
Thirdly, an important aspect. We are the second largest holder of Bitcoin worldwide in corporations. And I say that with conviction. We are not a treasury company like Strategy and others, but we do hold a large stack of Bitcoin that we've held for a long duration of time. We're in a full HODL position since last year. We made that decision as a capital allocation decision. And we've continued to hold Bitcoin since then. We also have a treasury operation within our company, some very smart people who know how to create value out of Bitcoin, not just hold it, but also create a yield around that. So all in all, in summary, it's not just one asset. It's a combination of different things. And it's a difficult space to operate, but there's value to be created for people who can think out of the box.
Thanks, Salman. All right.
So I'll give an unoriginal answer, but I promise I'm going to put an original spin on it, okay? I'm sure everyone's going to say that their team is their greatest asset. And of course, that's what I was going to say. And I would have said it first if I were sitting down there, but I was late to the stage. However, I do have an original angle to think about it. So this is an equity research conference. Thank you to everyone. I see some familiar faces in the room that met with me yesterday. I had one-on-ones all day. As always, HCW puts on a great conference with a great crew of investors to speak with.
I think the one thing I found that I want to remind everyone about our stock, if you're looking at Cipher, is that, of course, I'm going to argue we do have the best team in the industry, but let me tell you why that matters to you. And I think investors would be well served to evaluate any company at this conference on this front. What I have found is that too many investors in this space are looking at each of these companies as sort of like a bag of assets that they want to come up with a present value on and then say, "Is this stock cheap or rich?" And that's a wonderful value framework to think about investing, okay? I think that's a wonderful place to start.
The thing I would encourage you to remember, and certainly I think applies to our company, is that if you look at our team, we are where we are because we have a team that is particularly strong at originating great power contracts and building great data centers. We have, give or take, a dozen ex-Google data center employees on our staff, including basically our entire construction and operations team. Okay, so we also have a person who I think is the greatest originator of the strongest power contracts in the space, and the thing I remind people of is that we've got a wonderful Bitcoin mining business. You can have a view on that and value it. Kevin or Mike can give you a framework to help think about it from the HCW perspective in terms of what things are worth.
We also hopefully will have a series of HPC contracts where we're doing colocation, where you can come up with your version of the execution discounts and the net present value. The thing I'd encourage everyone to remember is those awesome people that originated the 11 data center sites we have now still come to work every day. We're going to buy more sites, and we're going to do more contracts, and we're going to build more things, and part of what you're trying to do as an equity investor is think about what is that upside? What's the sort of cheapness of the upside option in front of this massive wave of change that's coming to this space?
What I would try to convince you of for Cipher that's differentiating is, yes, it is our team that is our greatest asset, but I would argue it makes us the most attractive to an equity investor in terms of that upside potential that is created by this current marketplace. Told you I could try to be original.
Thank you.
Brian.
Yeah. My answer is sort of straightforward, which is that we are building a massive data center campus, and that campus itself is our most valuable asset, so we acquired this asset, which we call Helios, back in 2022. We purchased it because there was a massive amount of power that was available there that was already approved that could come online very quickly. We thought about that in the context of Bitcoin mining. We could mine Bitcoin reliably and inexpensively, but it was then very validating to us when early last year we pivoted the site into an AI and HPC site, and we're looking for tenants. We marketed that site to CoreWeave as well as hyperscalers, and the response was the same regardless of who we shared that with, which is like you're sitting on a pile of gold in this site.
800 megawatts at a single campus with already immediately available power, where we've already procured long lead-time equipment, was just incredibly valuable to these guys, and it was easy to set up site visits after that. We then chose to work with CoreWeave, and we've been scaling up the site from there. I will touch on the people component as well. I think the unique thing about Galaxy is that we are very diverse companies. So my team has a lot of smart and experienced people on it, and they're thinking about data centers every single day.
But we also have a lot of different pockets of experience in the digital asset space that we're able to leverage, which has been very useful when looking at pieces like financing or looking at new opportunities or identifying new risks that maybe we wouldn't see if we just had a group of people thinking about data centers every day. The last piece on our campus as well that I do want to highlight is not only do we have those 800 megawatts available now, but we also have an incremental 2.7 that we have applied for and we are waiting for approval for. So upon approval of that, we really believe that this asset can be one of the largest data center campuses in the world, and we're well underway in getting there with a Q1 energization.
Thanks, Brian.
Cool. I think the biggest thing that I've learned since we became public is the more you do as a company that we trade anywhere between call it $2-$3 billion in any given day. The more you do as a company that's smaller on that kind of market cap size, the less time an investor is willing to take to truly understand the business. So the more simple it is, the easier that people can go to understand the business. I think if you look at our company today, we have about $1 billion of Bitcoin on the balance sheet. Our equity stake in American Bitcoin is worth more than the whole company today in totality. Plus, we have the gigawatt of assets, 1.5 of net new assets.
But I think long-term, the fortunate kind of place that we've been able to operate from is in financing a lot of these opportunities that are in front of us. We don't need to raise net new capital. And so we've been able to take a much longer perspective on building the business and honestly haven't really cared about short-term share price. And as a result, have focused kind of heads down on what do we have to do and what foundation do we have to put in to actually build a strong business with a defensible moat. And I think there's two areas when we think about defensible moat on the Hut 8 side of the business. The first, and Tyler kind of piggybacking on what you said, is monetizing the gigawatt we have or the 1.5 gigawatts we just announced is great.
But really, to build a really meaningful company in the long term is being able to develop a flywheel of originating new sites to developing those sites and commercializing those sites. And I think renewable energy companies are a great example of that. They did that 20 years ago and built multi-hundred million dollar companies in doing so. And so I think, do we have a differentiated moat and flywheel of generating net new opportunities, developing those opportunities, building those and commercializing them? So I think that's kind of the first part. And I think we have experiences in different buckets from power generation and facilities that we own there. We have about 500 megawatts of behind-the-meter facilities that we run today as well. So deep experience there. And then have pretty large front-of-the-meter sites. We just announced 1.5 gigawatts with a gigawatt campus just a couple of weeks ago.
And so we believe the ability to continue to develop kind of that flywheel of net new growth. And then the second is, if you have the ability to find power, which we believe will become a scarcer and scarcer asset, then are you just in the business of selling kind of access to power, or do you have a right to actually build upon that power and build digital infrastructure? And then the question is, what is your kind of defensible moat around building that infrastructure stack itself? I think in building a traditional tier three data center, there are a lot of very capable people in the world that have built these data centers.
I mean, there's no secret that there's a set few firms that would be general contractors, A&E firms, MEP firms that design the majority of data centers that help build the majority of data centers and get hired. And so one area we spend a lot of time thinking about is, is there innovation that can happen on the digital infrastructure stack on the data center stack? I think in mining, an area we prided ourselves in a lot was being able to develop high quality at a low cost. So most recently, we developed a site called Vega in Amarillo, Texas, which was a liquid cool data center. And the thesis there was, can we develop a liquid-to-chip cool data center for the Bitcoin business that had analogies to AI data centers, and can we sell a low-redundant infrastructure stack at a lower cost to customers?
I think what we learned from that thesis, we built basically a 200 kilowatt per rack liquid cool data center for under $450,000 a megawatt. What we learned from that was customers cared about speed more than they cared about cost. And so v2 of that design that we're working on is, can we design a form factor that starts with low redundancy, but we design it in a way and spend a little bit more money upfront where we can actually upgrade it into Tier 3 standard so we can deliver speed of power, yet still use Bitcoin as the original initial underwriting to be able to capture land and monetize land?
And so, I think there's short term, which is build Tier 3 data centers that people know what they look and feel like with counterparties that they trust and deliver those to markets, and we get some type of NPV on those cash flows. I think that's right. Great. We'll create $2 billion of value of doing that. But long term, where is truly the bigger so what on what gives us a right to compete in the space? And the two areas for Hut 8 are the ability to procure power at scale and a continued ability to be able to do so. And second, a differentiated moat in designing and building data centers that is able to drive innovation around the infrastructure stack, either to drive speed and/or cost relative to quality. And so spending a lot of time doing those efforts.
And I think, I don't know, our conviction is that price will always catch up to intrinsic value over time. And as long as you don't have to raise capital, it doesn't really matter, and the market will work itself out.
Thanks, Asher. Sam?
Yeah, look, I think that the marker of a successful business is really simple. Do you have clients or not? So we have 23 clients. And the reason why we have 23 clients and growing is we have a very special approach to data center build-out. So we have what we call a retrofit model where we're able to build two times faster and 40% cheaper. In fact, we just got this client called Cerebras. Some of you guys may have heard that. We identified a mattress factory back in February. We're handing over the keys to them next week. And that is a facility that we were able to turn into a Tier 3 data center in six months. We were very allergic to greenfield. We think there's a huge execution risk in greenfields. That takes about 18 months, and there's a lot of moving parts on a greenfield.
So we focus on retrofits, which takes about six months. We just acquired a data center in North Carolina, a 1 million sq ft facility in North Carolina. That was a textile factory that has about 99 megawatts associated with it. We already have a whole bunch of LOIs attached to it. And the reason why we have all these clients is because we could do it two times faster and 40% cheaper, and we have a track record of that. And the latest example of that is Cerebras, where we're handing over the keys to them next week. That thing used to be a mattress factory just six months ago.
All right. Thanks. Next question. And I'm going to let someone who's passionate about it give the first answer, and we'll see where we're going to go. Where do you see the data center industry in the next five years or in five years' time? Not in the next five years, but in five years' time, specifically as it relates to Bitcoin mining, which we'll be after happening, HPC consumption, AI consumption, and then how that's going to interact with the utility grid, the global utility grid, but also the local utility grid. So that's three very distinct things. But five years out, who would like to take that?
I think it's a lock, when you look at Bitcoin mining, literally every U.S. corporation that does Bitcoin mining is talking about AI, right? Certainly AI is going to be an important play in the next five years. When we talk about Bitcoin, after every halving, it becomes more difficult, and cost is important. Being innovative as to how you mine Bitcoin is going to be important, just like we talked about. Marathon has been focused on finding opportunities where we can mine Bitcoin irrespective of what hash price does, irrespective of what global hash rate does. I think those innovations are going to be important in this sector to continue to evolve over a period of time. Modular approach to AI is going to be interesting.
When you think about the current model of AI HPC with large-scale models and large-scale data centers being built with reliance on one big customer, that's an interesting model for some, not for everybody. Mara has chosen not to follow that path. We are more a believer of inference on the edge, and that's going to be important because when there's data, there's security. Coca-Cola doesn't want to share their information with PepsiCo. JPMorgan would not want to share their information with Bank of America, for example, and that's where edge is important in our story because we believe in the future that becomes a 170 billion market or TAM in the next four or five years or so that allows us the opportunity to tap that and have customers who are recurring paying over a period of time as a recurring revenue platform as a service model.
The other thing is that it's very important to note that Tier 4 data centers, I'm sure you can talk for hours about that, Tier 4 data centers are very different from picks and shovels like Bitcoin mining, and you need a different level of expertise for that, so over time, you would expect that people will learn and evolve over a period of time, so there's a lot of learning and education when it comes to Bitcoin mining going into AI or software as a service or e-commerce, I'm sorry, software or platform as a service. In terms of inference, as I had mentioned previously, we are subject to closing this transaction. Q4, Exaion is an NVIDIA partner and GDPR certified, which means that they have access to compliance of data and also access to sovereign customers.
That's the kind of stuff we believe is going to create value from a longevity standpoint.
You also like the.
Sorry, I'll jump in real quick. I mean, I do think the curtailability of the Bitcoin mining load is extremely valuable as the data center industry evolves. By analogy, this is my second company I've been a part of founding in the Bitcoin space, and I used to travel around the world speaking to the world's biggest institutional investors to convince them to buy Bitcoin, and none of them would do it because it's like, well, I don't want the reputation risk, and I don't know, and now, of course, everyone knows Larry Fink is one of the big cheerleaders for Bitcoin, right, and I could have told you in 2018, 2017, that would eventually happen. I just didn't know when it would happen because it was just so rational that it would happen.
And I look at pairing Bitcoin mining with HPC data centers, traditional data centers as the same thing. I mean, you measure a peak PUE. By definition, most of the time, the data center is not operating at that peak, and there's extra power there. And so the ability for Bitcoin mining to soak up that excess and also work with the grid operators, to your question, to be able to, if they are in a position where they have so many data centers, they have to be able to curtail part of a load, it's very natural that the two eventually grow together. It'll be interesting to see how we get there, but it seems inevitable at some point.
I think.
You know, I agree with what Tyler was saying about pairing those assets together. But I think there's something that we don't talk about enough, and that is Salman's company holds billions of dollars in Bitcoin, and Michael Saylor's company holds billions of dollars in Bitcoin. And during the prior administration, there was a concerted effort to drive Bitcoin mining out of the United States. And I believe that providing and securing block space for those massive hot treasury companies, for American investors, for American corporations that hold Bitcoin on their balance sheet is a national security issue. We had a letter that was issued today referring a couple of players in the space, Canaan and Bitmain, to the Department of Treasury as a potential CFIUS risk.
They're talking about the consolidation of Chinese ownership of the block space and Chinese sovereign ownership of data centers and energy assets within the United States as a security risk. While we're focused on maximizing the value of each and every megawatt that we control, 1.02 gigawatts currently under contract with a 1.2-gigawatt pipeline spread over five states, we believe that providing security and support to the blockchain is, in fact, a national security risk. There's a value in providing that function that I think has gone maybe unrecognized. The Bitcoin mining space is really a challenge because each and every one of us, at some point over the last two cycles, have been the pretty girl of the dance. I mean, today it's Asher. Sometimes it's somebody else.
But as Bitcoin mining comes back into vogue, I think the companies that remain committed to operational excellence and capital stewardship are going to differentiate themselves because there is absolute need to maintain that support of the blockchain.
Yeah. And look, I agree on the securitization. I think on the question of kind of five years out from today, the question we always ask ourselves is just, what is our right to deliver this capacity? And I think five years out from today, the question is going to be, what does the supply and demand kind of balance look like? Today, if you have power and you have access to power, if you can execute on it, people can give you the trust, then you can deliver a data center to them. The reality is every hyperscaler self-builds their own data centers, and that's their preference if they can. I think you'll see a lot more of these AI labs starting to self-build these data centers as well. And so the question is, do they need additional demand and capacity than they can build themselves?
And if so, they'll sign colocation space just like they're doing today. And so the question in five years is, if that kind of supply and demand mismatch isn't there, then what can you deliver to them that they don't have? Is there something unique on the infrastructure stack that you can deliver? Is there price differentiation? Is it a cost of capital gain because you can drive lower costs of capital and you can deliver that relative to them building themselves? And so I think five years from today, if the power story and just being able to have access to power, like when we started these conversations two years ago, you have access to 1,200 megawatts, great, people are interested. I think today, unless you have a minimum of 300 megawatts scaling up to a gigawatt plus, people have very low interest.
I think those bogeys will continue to increase and those bars will get higher if they're not self-building and they're building externally. I think how we think about kind of the market is, what is our competitive moat in finding power and building to be able to build power infrastructure at scale? Secondly, on the infrastructure stack itself, how do you actually compete? I think thinking about kind of, I think with a lot of training clusters, you'll see them force into curtailing. You're not going to have them up 24/7, 365. A lot of the AI labs are already willing to start thinking about curtailability. I think you'll see that happening on training clusters. In regards to Bitcoin mining, I really want us to continue to build and scale Bitcoin mining in the U.S.
I think Bitcoin mining to just develop on its own underwriting in the U.S., unless you have some behind-the-meter stranded power, curtail power, low-cost power, et cetera situation, it's hard to compete against other places that have much cheaper stranded power and are very cheap to build. And so as a result, I think on Hut 8, being able to underwrite a site for a longer-term value of megawatts and then being able to have Bitcoin mining there and your underwriting is not purely on that use case, that's how you can kind of justify it. But overall, I think in five years, all of these companies, I believe, up here will all find their niches and will find their competitive moats and will be different types of business that evolve and change and target a specific part of the market.
And so I think the demand cycles will change and companies will prove their value over time.
I got just 30 seconds to add to that. It's related. One piece, the data center market's going to be significantly bigger in five years, and then a point that Asher said that I totally agree with is that these campuses, though, are multi-hundred megawatt campuses and at densities that we've never seen before, and so putting that together, we're going to have a lot of new players in the market that are not thought of as traditional data centers today because the traditional data center companies, rather, have not built these types of campuses before, and so that's where I do think there's incredible opportunity for all of the companies on stage, just as Asher said, to find their niche.
Yeah. All right. We could go on forever, but we're between you and the President. So we have somewhere else to go. Thank you, everybody. Let's give these guys a big hand.