Good afternoon, everyone, and welcome to Hut 8's Special Fireside Chat with Sean Glennon, CFO of Hut 8, and Asher Genoot, CEO of Hut 8. So today, Hut 8 announced two major announcements. We announced an AI infrastructure partnership with Anthropic and FluidStack, where Hut will develop at least 245 megawatts of infrastructure and up to 2.3 gigawatts of AI data center infrastructure for Anthropic, one of the leading AI labs who will be leveraging FluidStack's clusters.
We also announced the signing of a 15-year, 245-megawatt AI data center lease at our River Bend campus in Louisiana, with a total contract value of $7 billion over the base term and up to $17.7 billion if all renewal options are exercised. Google will be providing a financial backstop covering obligations for the 15-year base lease term, and we are really excited to get into this transaction with you guys today.
So, but first, some quick housekeeping items. Unless otherwise indicated, all figures discussed today are in U.S. dollars. Certain statements made during this call may constitute forward-looking statements within the meaning of applicable securities laws. These statements reflect current expectations and are subject to risks and uncertainties that could cause actual results to differ materially. Certain key risks are detailed in our Form 10-K for the year ended December 31, 2024, and in our other continuous disclosure filings. Except as required by law, we assume no obligation to update or revise any forward-looking statements. And with that, let's get into the AI infrastructure partnership that Hut 8 announced this morning.
Thanks, everybody, for coming today. This wasn't a conversion story. We had a new site that we developed, we originated, and we commercialized as AI demand came to reality, and so we did not convert an existing site, but we net developed it with the development pipeline that we've been growing. I think it's also one of the most thoughtful deals that have been structured, and we're excited to go into the nuances of what that means. We're able to retain the relationship, the upside of demand from Anthropic, FluidStack, and Google, but be able to have an investment-grade counterparty be able to provide a guarantee and financial backstop on all lease payments, which is about $7 billion for the term of the first 15 years, minimum power bills, which is about another $1 billion, property tax, insurance, and other pass-through expenses as well.
And so not only is this a backstop on the debt and the financing, this backs up the full obligations of the lease payments to us and therefore the equity investment that we're making and the return profile and the overall value of the site that we believe. That was extremely critical to us and why we were patient in finding the right deal that was structured thoughtfully with the right economics. We also have a holistic deal that we announced that was focused not just on the actual contract terms itself, but how do we actually execute on this project and how do we deliver it to the customer? One with partners that have scale, our industry leaders, and we believe a program that we can continue to repeat and scale as we now go into growth mode.
Awesome, Asher. Appreciate that overview, but now let's get into it. So maybe let's start off with how does this partnership, what does this partnership say about Hut 8's ability to support frontier model developers and effectively integrate us into the AI supply chain?
I think it validates our ability to work with high-class and high-counterparty partnerships. It validates our power-first development model that we've been discussing throughout the last year and a half. And it shows that we have confidence from some of the leading AI companies in the world to have our power development pipeline be able to help support them and continue their runway and give them the capacity to grow in advancing frontier AI technologies.
Very cool. So maybe let's start with how should we think about the long-term expansion potential of this partnership? Maybe walk us through in greater detail what that looks like.
So tranche one is the deal that we announced today. We have a single building that supports 245 megawatts of IT capacity. On the campus, they'll also have a ROFO to be able to expand an additional 1,000 megawatts of capacity. And then there's a third tranche, which is about a gigawatt of capacity as well and other sites within our development portfolio that they found interesting. And so to support them on a holistic 2.3 gigawatts of additional expansion opportunity that we'll be able to have them diligence, review, and scale into as they have their demand profiles come in. Most importantly here, it's proof of scalability and repeatability within Hut 8's platform and development pipeline. You're on mute, Sue.
Let's get into the actual meat of the deal. Let's get into the transaction overview. So can you walk us through the economics of the deal, including how we evaluated Google's credit support and what their guarantee actually covers? And maybe can you touch on, you know, we've seen a few deals signed with FluidStack in the market. What is differentiated about this deal in particular? And then finally, why didn't we take the high-yield bond road?
So I'll first start with kind of lease structure. The way that we've built our business and the way we've structured our reporting segments in the company is we have a power layer, a digital infrastructure layer, and a compute layer. The power layer is the actual vertical, and the power layer is the horizontal investments that we make and also the managed services and operating business. The digital infrastructure layer is when we go horizontal and actually build the building. And so if you go with a gross modified structure, that's essentially one business, right? You are building and operating a turnkey data center for a customer. And so that's where you see kind of 80%, 85% gross margins on the revenue that you receive from those inflows. But you really have two businesses there.
You have the business of developing and delivering a data center to the customer and the business of actually operating and maintaining SLAs and service level obligations of operating that data center. For us, it was really important in how our company is structured to be able to separate those within the business units that they operate. So we have a Triple Net Lease here, which is a fixed rate of return on the initial CapEx that we invest into the data center. Once we deliver that data center to the customer, we will have that cash flow coming in and all costs are passed through to the customer where it's really a rate of return on that capital deployed on that execution. Then we currently have an LOI to sign a managed services OSA agreement.
That will be incremental economics in a 15-year deal to operate the data center on behalf of the customer and have those relative SLAs and penalties on those obligations within that contract. So I think structuring for us was really, really important as we thought about our business units, how we think about lowering our cost of capital and the different risk profiles that the business units has in development and construction versus operations and maintenance. So that's first in terms of triple net lease structure and why you see that NOI contribution on the 15-year average at $454 million and such a high, near 100% flow through. So that's kind of part one. The second was it was important to us. We believe that power is scarce. We believe that near-term power is even more scarce.
And so because of that conviction in the asset, we have had many conversations with different customers and we said, we have something that we think is of value. We want to solve for two things. We want a good deal that is fair for us and fair to the customer, and we want a customer that we can scale with. Here, we have three customers that we can scale with within Anthropic, FluidStack, and Google and those relationships that we've built with them through the process and will continue to build. In terms of the actual structure, we did not want just a guarantee of the debt. That helps us finance the debt, which is great, but that doesn't help us in underwriting the project for our equity investment.
These are 15-year contracts, and we have to think about the creditworthiness of those counterparties for 15 years and how we think about that builds to the foundation of the business that we're growing. And so here, Google is providing a financial backstop to all of the lease payments in this initial term. That's about $7 billion. All of the pass-through payments, including power bills, minimum power bills about $1 billion during the 15-year term. We have insurance, we have other costs like maintenance, operations, property tax, and so forth. And so all of those together take us to a minimum of about a $9 billion financial backstop in this overall transaction. And so for us, the look-through is essentially Google is our credit counterpart and the customer in this lease during this duration, right? We didn't have to give equity upfront in order to achieve that.
If they do step in, we are not diluting our equity in that scenario either. They are the customer. FluidStack is not the customer at that point. We're so excited to grow and scale with FluidStack through this partnership with Google and Anthropic. The overall deal here that you see, our average cost to build would be around $9-$10 million a megawatt. The delivery schedule was important as well. We didn't want to overpromise just to get a deal done. We wanted to set strong conservative expectations of delivery. You see, early 2027 is when we'll start the first data hall delivery and deliver the full data center within 2027. We hope to exceed those expectations for the customer, deliver early, and give them their capacity.
But what we did not want to do was to deliver late and lock in the contract today, but hurt the relationship to be able to scale and grow with the customer, which obviously makes the upfront negotiation a little bit harder. The other element was the financing. Because of that, the structure of the look-through in terms of the IG creditworthiness of Google, we were able to get JPMorgan and Goldman Sachs to provide a financing partnership with us to finance up to 85% loan-to-cost of the data center. That's SOFR of 225, target of SOFR at 225. That's some of the best financing in the market today because of the creditworthiness of these cash flows and of this agreement.
Asher, just jumping in here quickly, it's kind of unusual to see a target delivery date on a page full of financial metrics. What's the rationale behind that?
Yeah, because for us, signing the contract is great. It's what is our potential expected returns. But here's where the real work begins. How do we execute this project and how do we deliver this infrastructure to the customer before they need it and be able to meet the timelines that we give to them and not be delayed and not have construction risks, supply chain risks, and so forth? And so we'll talk a lot about that in this presentation, but a part of us waiting and having shareholders be patient was because we want to put a holistic delivery solution and program together to announce to the market and to sign the contract with all of that in place rather than figuring it out after the fact once we've made the commitment to the customer already.
Awesome. So maybe you can walk us through a little bit on the economic durability and the revenue visibility of this deal. And I'd also love for you to touch on what is the difference and why does it matter that this is a ROFO versus a ROFR with respect to this deal?
So the initial base term is 15 years. That's about $7 billion of NOI contribution, roughly around $454 million per year on average. If the customer extends during the duration of the additional 15 years of extension options that they have, that second 15 years is about $700 million of average NOI contribution during that period of time. So on an average, you're looking at about $580 million across 30 years on NOI contribution per year. And so overall contract value obviously increases $17.5 billion. This campus is one that we believe a ton in and its ability to expand and to scale. And so the customer has a ROFO right. And what that means is they have the ability, when we have additional capacity to come online, to give us a right of first offer on what they would like to propose to us.
The reason why that's a little different than a ROFR is a ROFR really hurts marketability. If we go on and spend months marketing a site to a customer and that customer knows if I offer you and we get to a final deal, you have someone that can just jump in and get a right of first refusal and say, "We're going to take this instead," it really makes the ability to scale this campus and the marketability of that much more difficult. And so this is a ROFR we like because it shows our intent where we want to grow and scale with our tenants, but it also doesn't take away the marketability of future campuses in the case that the tenant just wants that optionality.
Yeah, that's a super salient point. So I think we should now sort of zoom in on some of the incredibly impressive partners that we've gotten to align with us on this deal. So from across financing, power, construction, and supply chain, how did these relationships come together? And given the industry's heightened focus on execution risk and recent delays, how confident are we in delivering by Q2 2027?
Very confident. A big reason that we've taken so long to announce a deal and to sign a deal was that we didn't want to just negotiate and get to the finish line on a contract with a customer and make commitments that we didn't have buttoned up on the other side. And so we worked really hard to get this program together and get these executed agreements where on the energy side, we have a strong utility power partner that can scale and grow with us. They were able to increase our initial 300 megawatts of utility capacity to 330 megawatts because we need to put a little bit more IT capacity into the building. If we can go back to the previous slide, thank you.
Then when we think about the actual financing, we talked about Google and their financial backstop and covering those commitments of the lease payments of $7 billion and covering the pass-through obligations under the lease as well. That's what gave confidence to JPMorgan and Goldman Sachs to come in as partners to have our target SOFR plus 225 cost of financing up to 85% of the overall build cost of the project and for them to be willing to put balance sheet capital as well to help drive that financing. And lastly, the ability to actually have the labor and talent, the engineers, the skilled labor in order to do so. So Jacobs is our design- build partner on this project and the program that we're developing to rinse and repeat into scale. They have over 40,000 people. They're a publicly traded EPCM company as well.
They are going to be managing our subcontractors, which we've all identified as well and already have staff committed to this project. Then Vertiv from a long lead time equipment perspective as well. Today when we announce this deal, we're not just announcing a customer. We're not just announcing a great credit counterparty that stabilizes the cash flows for this data center across the next 15 years. We are announcing an overall program to deliver this project on time and on budget and to be able to deliver that and repeat that program. This was key as well. We met with a lot of different counterparties. We wanted partners that could scale with us.
We wanted partners that had the infrastructure to be able to duplicate the model that we spent so much time building and grow from project and project to us and to repeat and repeat and repeat throughout all of the sites we have within our development pipeline.
Yeah.
One other thing I'd just jump in and add here is while JPMorgan and Goldman Sachs are providing capital, the underwriting process with banks like themselves, especially when they're doing something on their balance sheet, is extremely stringent, and so as we look to develop a construction project and a supply chain to get the project off the ground, they also had to do a lot of their own diligence, so you not only have us doing diligence, but you have JPMorgan and Goldman Sachs who have been in the market for a long time and are putting their balance sheets behind this project, effectively giving the stamp of approval that they think the program that we put together to build this is one that works.
Yeah, super, super important point there, Sean. Okay, so maybe let's pivot a little bit and talk about we've seen other deals signed in the market, other players that are having great success in this space, but what really differentiates Hut 8's development model from any other AI data center builders out there? What is it that makes us incredibly unique?
Yeah, I'll jump in here. And I think it really starts with the way we develop projects. For us, everything starts with the electron. We view electrons as a product and not a commodity. And so that starts with working with a company like Entergy, who provided us with an electric supply agreement that we could then sink our teeth into and understand that there's really reliable, uninterruptible power there for the long haul. Next, and I think people don't really understand the full gravity of this, a local ground game is incredibly important. You're starting to read articles about NIMBYism. And so we wanted to make sure that we were working with the local community to have a project that they would get behind.
So the state of Louisiana, West Feliciana Parish, both got really behind it, and we worked hand in glove with them to make sure that we were delivering a project that would deliver economic value, jobs, and a lot of high-tech improvement to the state. Next, once we have those in place, then you're able to commercialize a project. I think if you go to the market with an unbaked project and you start trying to market it to Google and the other hyperscalers, you're going to lose a lot of credibility really fast. And so once we have the first two on the left side of this page locked in, then we say, "Okay, let's go market the project to people." Once we've got a lease kind of in negotiation form, then we got to execute.
And as we walk through on the prior page, getting the financing locked in, getting the supply chain locked in, and getting the construction locked in really de-risks the project from an execution perspective. And we've preached patience, and we wanted to give you guys a deal that really spoke for itself. And as you put all these counterparties in this full development timeline together, you can see how we do that. And this is repeatable. We're very methodical about this, and we do this at all of our sites.
You guys might have noticed Sean's suit and tie. He is one of the former top bankers at Citi. Sean, maybe you can just talk a little bit about, as a former banker who did over 80 billion of transactions in the energy space, energy and infrastructure space, what does this deal look like from your seat?
Yeah, I mean, there's a bunch of boxes you got to check when you're doing a big infrastructure project. First and foremost, what's the creditworthiness of it? At the end of the day, all infrastructure projects are a look-through to the counterparty. Are you investment-grade or are you not? That's typically the big litmus test. So we've got that. Then you're going to dive into the economics. What does that look like?
Sean, frankly.
Sorry, go ahead, Asher.
Sean, let's do this actually. We have a slide where we kind of go through a recap of the overall transaction. Maybe let's jump into that. We have two more to get through, and then let's jump into that when we get there.
Absolutely.
Okay. So yeah, because we are a little bit tight for time here, so let's quickly go into Riverbend, which is now, I guess it's safe to say, our flagship greenfield project. Remember, guys, Hut 8 did not tear down sites. This is greenfield development. Maybe you guys can walk us through really quickly what drew us to this site, Asher, and how did it evolve into today's flagship product? Because Louisiana is not a traditional data center market where you see data centers of scale like this, right? So what drew us there? How did that evolve?
As Sean shared in the slide previously, we follow the power. We really looked at where is a strong power story within the U.S. that was underappreciated and overlooked. So Louisiana is a state that actually sits right in the middle of Dallas and Atlanta, which are two major data center hubs. The connectivity of this campus allows us to connect fiber north, south, east, and west with low latency. We've looked at having customers come to the site that offered cloud availability zone services as well, not just AI training. So when we look at the campus, there's about 3.7 gigawatts of generation just within a couple of miles away. Our switch yard actually connects to the River Bend nuclear substation. Then transmission. The transmission story on the site is also extremely strong. We have a couple of 230 kV lines, 69 kV lines as well.
We're only connecting to one of those lines right now for this first building. Because of that conviction in the overall power story and generation and transmission, the fiber story, we've actually increased our commitment to the campus size as well. We announced that we owned 592 acres. We've incrementally increased that to 35 acres, but more importantly, we've actually optioned the 2,300 acres contingent and connected to that property as well. We have about 3,000 acres of connected land to be able to grow and scale into this campus, into this gigawatt platform. We're very, very excited by this and think we have all of the right ingredients to continue to build the campus. I believe that's why we've had so much interest on this campus over the last year.
Yeah, no kidding. That's insanely exciting. So maybe we can quickly talk about how are we de-risking the long lead procurement items and labor for this full build, right? Because that's obviously, there's a lot of bottlenecks in the market right now. So maybe you can talk a little bit about some of the nuances that are unique to this site and actually Louisiana as a market in general.
So when we think about development and risk to development, we already started building the switch yard with Entergy, and we started procuring equipment for this substation as well. Some of this long lead time equipment, like 34.5 kV breakers from the substation on the low side into the actual data center building, that's malleable equipment that we would use at any data center campus we have, regardless of transmission level voltage. Right? And then there's some more unique products like the substation itself and if that voltage is more malleable. And so we started running the civil work on this site and having a lot of these long lead time items for the electrical infrastructure on the high voltage side.
When we announced this deal today, we also locked in the long lead time items as well within the data center from chillers to busways to CDUs and across the whole data center infrastructure stack. We have partners that have committed to the timelines that we have committed to as well, and that's what you see kind of on this screen today is we did initial work that we felt comfortable to from an investment perspective, but also made sure that all of the risk factors in actually deploying and executing this from a delivery perspective was already solved before signing the contract with the customer.
Maybe you can quickly touch on the significance of our skidded design and also the pilot that we actually did at this site.
Sure, so it's two different elements. One is when we think about scaling and growing, we believe that the more you can actually take out of the data center and able to skid into a repeatable assembly line in a manufacturing facility, the faster execution you can have on site and the lower risk you have of something not being executed well, and so when we think about the deployment of this data center, it is a traditional data center building, but the way that we have thought about our procurement of this data center, we're not just stick building the site on campus.
We're actually doing a lot of that assembly where we're connecting different components together, purchasing that full skid, having it arrive on site, and then you're connecting those interconnection points on that skidded design, which we'll go into more detail as the campus continues to mature and construct, and then separately from a pilot perspective, Louisiana has been an amazing state to work into. We've been welcomed by the governor, by LED, which is the Louisiana Economic Development Board, the parish president, Kenny, the local parish sheriff, the school district. I mean, we have a community locally and at a state level that have wanted us to come into the state. They recently passed a data center sales and use tax exemption as well.
And we work closely with the parish to put together a program that allowed us to be able to grow and attract the right customers from a tax perspective, be able to generate revenue for the local community as well. And so we feel extremely grateful, and that's why we have so much conviction and continue to scale within the state.
Awesome. Okay, so we're about to wrap it up here. We only have a few more slides left. So I want to quickly talk about maybe let's zoom out again and talk about why is this deal so important for Hut 8's long-term financial profile and our valuation. Maybe Sean, I'll let you kick it off there as the former banker and now CFO in the role.
Yeah, so before I go there, I'll actually start with why this deal is so great. I spent 13 years in power and utilities. It's an infrastructure business. I'm an infrastructure geek. I love infrastructure. And when you evaluate an infrastructure project, there's a kind of a checklist of things you need to go through and does the project check all the boxes. The first one is always creditworthiness, as I was saying before. Is the tenant or the off-taker going to be around to pay their bills for a long time? I think with Google's support here, that box is very clearly checked. Then you got to get into the economics. And we've already talked about the actual NOI generation, but the structure of the economics. We have an inflation sort of protection mechanism through the escalator that comes every year.
That's incredibly rare and important in infrastructure contracts and the triple net structure, right, which has a pass-through of all costs. Next, you're going to get into financing, both on the debt and the equity side. So from on the debt side, we've been able to secure 85% loan-to-cost at SOFR plus 225, which, given where SOFR is right now, is a sub 6% cost of capital on the debt. On the equity side, we also have to think about dilution. We didn't give up any warrants to do this project, which I think is very differentiated from some of the other things that have been in the market. And it protects our shareholders as we think about value creation. And then finally, having sat in a lot of credit committees in my career, there's like a kind of logo test that you do, right?
And that logo test is sort of like, does this pass the execution thresholds we need? And I think as you look at the logos here, we've got Jacobs, Vertiv, Google, Goldman Sachs, JP Morgan, Entergy. It's a screaming home run from that perspective, candidly. And I'm not just saying this because I work here. It's one of the best infrastructure deals I've ever seen in my career. And I'm really proud to be a part of it. I think that dovetails nicely into the next slide where we talk about what does this mean for valuation for the company. And if we could go to the next slide. Yeah, thanks. So we were joking around before we came here to this meeting about where we were a year ago. And I told Asher, I was like, "Look, this was inevitable. I saw the team here.
I saw what we had in front of us, and this wasn't an if, it was a when. Now it's going to be a how much and what's next, right? Because we know that's going to be your next question. As we look at where other companies are trading in the market from a dollar per megawatt perspective, and we apply that to what we have in front of us with an over nine-gigawatt development pipeline, that is extremely attractive to me. Moreover, if you think about the revenue generation we just created, that was about $1.85 million per megawatt. I'll let you guys do the math here, but clearly we're just at the ground floor of something that's pretty impressive from a value creation perspective.
Moreover, we just, and one final point on this, we just developed a site for mid-teens yield, and the market trades at around mid-single digits cap rate. So that's significant equity value creation right off the bat. And the fact that we've actually shown we can do this now, I think de-risks the rest of the pipeline tremendously. But with that, Asher, I'll give it to you to kind of wrap it up with some final thoughts.
We appreciate the patience that people gave us, and we hope that what we delivered today met true to the promises that we made in terms of giving a holistic solution to the first deal that was thoughtfully structured for the long term, not just wanting to get a deal done because there are deals that we walked away from and thought it was not the wrong decision, but today, we're confident in where we are, the deal that we put together, and more importantly, the program of partners that we have to repeat and to scale this structure in this program across this site and many other sites in our pipeline.
Awesome. Well, guys, that was great. Let's open up the floor now to a Q&A. We've got a couple of guys here in the queue. What I'm going to do is I'm going to read the question, and then Sean and Asher will answer. So the first one comes from our friend Amit at Millennium. So when is expected power delivery? What years for the ROFO, not the ROFO, ROFR, the ROFO portion of the FluidStack transaction? Thank you.
Yeah, so for the additional 1,000 megawatts, the conversation we had, we submitted that request with Entergy Louisiana. And we're actually meeting. I actually talked to Phillip May, the CEO, today after the announcement, and we said that we would put together a schedule and pipeline once we had our first deal that we announced for the initial capacity. And so we'll be releasing more of that information in terms of when that capacity becomes available and looking at doing similar chunks in terms of building sizes as we scale into that capacity.
Awesome. Okay. So the next question comes from Anonymous. Thank you for the call and congratulations on the deal. Could you please clarify if the $9-11 million per megawatt of CapEx refers to critical IT of 245 megawatts or gross of 330 megawatts? Thank you.
Easy. Critical IT. We think about cost relative to the revenues of megawatts that we're receiving.
Awesome. Another one here. So what do financial penalties look like in the event of delays? We've heard a lot about delays in the market. What does that look like for Hut if we are delayed?
The structure that we've created and the structure that we think is fair that we've had across many customer contracts that we've negotiated is that there's delay penalties that accrue in the six months following a delay, right? So that escalates in month one, two, three, four, five, six. At six months, there's no termination right on a delay. However, there is a step in right where the customer would come in and try to remedy and bring the site online. Once a site is delivered to the customer, then any delay credits that may have been accrued, those credits will be abated towards the first couple of months of rent, up to 60% of the rent cost until those rent abatements are abated and then the continued lease term continues.
And so we think in terms of the way that we've landed here as well, it's not overly punitive where you have a large termination rate if the project is delayed. However, it's fair and equitable in regards to having the right incentives to deliver on time and also having a remedy for the customer to step in and control that destiny as well if they don't. And I think this ties back to one of the earlier points that I made, which is giving realistic timelines to customers and then being extremely aggressive in trying to deliver them the best schedule in reality. And I think that's what we've done here today and hope to continue to build customer trust by doing so and by executing well for them.
Awesome. Okay, so the next question comes from our friend Tiago at WHG. So congratulations, team. Can you please provide additional details to tranches two and three? Do we have any insight yet on expected timing, the regulatory approval process, economics? Do you have any color you can share?
So tranche two was the 1,000 megawatts on the Riverbend campus that I had mentioned. In terms of tranche three, we have sites currently in our capacity and our development stage and sites in our capacity of exclusivity. So those sites were in active discussions on and diligence and negotiations on. And so we'll share more with due time. And for Anthropic, they have to look at their demand curves, how they want to phase into that demand, and ultimately how quickly they would like to look at the sites that we have and the attractiveness of those.
Awesome. Okay. Next question also comes from Anonymous. So how confident are we in the indicative SOFR plus 220 basis point financing costs? Appreciate if we can't provide much guidance on the expected margin of error here, but if you can, that'd be great. Would also like to confirm that the Google financial backstop covers the entirety of the 15 years contracted lease, AKA $7 billion. If we could reconfirm that.
Yeah. So on the first one, it's SOFR plus 225. We have a 25 basis point spread, so 225 to 250. And that's what we have with JP Morgan and Goldman and the terms that they have. And that aligns with kind of financing around a credit counterparty like Google. And that dovetails into the second, which is the Google financial backstop. It covers the entirety of the 15 years of contracted lease payment obligations, which is around $7 billion, but also because of the triple net structure, the pass-through obligations as well of the minimum energy bills or whatever energy consumption, the other pass-through things like insurance, other maintenance costs that are associated with the triple net nature. And so it covers the full costs of the lease from a tenant perspective when it comes to the lease and the pass-through obligations.
Awesome. Okay, so this is a big one. I've got quite a few people asking me this here. How are we thinking about funding the equity component of this project? So we've got the 85% LTC. What does the funding of that equity component look like?
Yeah, that's a great question. And so the equity is about $300-$400 million. Our balance sheet today kind of far outweighs that equity that's needed for the project. And so we feel comfortable that our balance sheet today covers the equity that's needed. We obviously will continue to look at opportunistic ways to bring on capital as we look at the overall development pipeline if needed. But today, I mean, we have 85% from bank balance sheet financing. And I actually got calls this morning on mezz financing on just the project level on the equity component as well. And so we feel today that there is no contingency or there's no need for financing in order to develop this data center.
At the same time, we're going to make sure that we continue to have a robust balance sheet as we look at our overall development pipeline and continued demand that we have across other campuses too.
Awesome. Okay, so from our friend Stephen Byrd at Morgan Stanley, hey, Stephen, is there a certain magnitude of SG&A that we should be factoring in?
So we'll share some of that in our next earnings discussion. When we think about SG&A related to actually the site construction of Riverbend itself, we have the team that's on it from an energy origination, pre-construction, design engineering, supply chain engineering, and construction perspective, and also operations. So we already have team leads on this project running that. The SG&A that would come is growth. We have demand for the gigawatts of capacity within our pipeline. We will grow our talent in order to support the needs of the demand signals that we're receiving from the customers. So SG&A will not be correlated to this data center and the execution of this data center. However, we look to continue to grow our teams to be able to actually execute on all of the demand that we see and the growth opportunity we see ahead of us.
Awesome. Okay, cool. And then the next one comes from George, our good friend at Craig-Hallum. So he asked, could tranche three come before tranche two, or could it happen simultaneously?
Yeah, tranche three could definitely come before tranche two. Tranche three is the existing assets that we have within our development pipeline under development under exclusivity. And some of those, as we shared in previous earnings calls, are within next year. And so those you can think of as kind of in parallel rather than one after the other.
Awesome. And then Joe, our good friend at Canaccord Genuity, Joe Vafi, does signing this deal potentially accelerate capacity from under development to under construction?
It depends on Anthropic's demand signals, right? They don't have an obligation to step into tranche two or tranche three. They have the option, and we have kind of announced a partnership in principle to go and scale and develop for them, right? That was important to us. We had a lot of demand signals for this campus, and we wanted partners that had an intent to want to scale and grow with us. And so here, Anthropic wanted to announce with us that they had that intent to grow and to scale in addition to just the first data center that we signed, and we're in discussions on other campuses already.
For the other customers that had demand signals at our Riverbend campus, we continue to have deep relationships with them, continue to have commitment to them as well, and are looking at other campuses with them as well.
Awesome. So also now from John, our good friend at Needham, appreciate the less aggressive timeline than peers. This is great. We talked a little bit about penalties. Can you discuss if there are ones related specifically to delays, and can you comment on supply chain to give us a little bit more confidence on that front?
Yeah, so the delay penalty regime is if we deliver on month X. If your delivery is late and you're X plus one, two, three, every one of those months, you'll step up into delay penalties. You'll usually see contracts that start anywhere from 0%-50% of the monthly rent as a delay penalty and go up to 150%, in some extreme cases, 200%. We went kind of middle down the fairway and feel comfortable with where we landed there to make sure the customer had confidence that we were confident in our delivery schedules. How do we make sure that we're not delayed? That caps out at about five months for six months for us until a separate right comes in. The delay penalties do have a cap.
Why that from our perspective was how do we solve for it to make sure we're not delayed? Because if you're delayed outside of delay penalties, you lose trust with the customer. They have billions of dollars, if not tens of billions of dollars of equipment that they've ordered based on your commitments. And so that's why for us, having a conservative delivery schedule, but being able to accelerate that was important. And also having the partners and contracts signed simultaneous with the customer obligation was important. So that's in terms of Jacobs, our design and build partner on this project, their staff to be able to actually support into GC this project, but also all the subcontractors that have the talent locally to be able to support the different skilled trades as well.
With Vertiv as a supply chain partner, making sure that all long lead time items we had solved for and we had committed dates and deliveries to as well, and having that unified approach where if we have delay penalties, our partners do as well, and we work together to execute on the schedules that we've committed to together.
Awesome. Okay, let's keep going here. So our friend Patrick at Piper Sandler, do you think having Anthropic as a partner is going to allow you to diligence sites any faster than previous? Will it speed up the power approval process?
Most definitely. Anthropic and other customers that were very late stage with us as well on this campus. I think I'm very grateful for the relationship that we built over the last two years. They've definitely compounded. We want to be good partners. We've always been transparent with them, what we believe we can do and what we can't do, why we bring in partners, what we're trying to solve for. And today, when we have new sites come in, we're able to get really quick feedback on if it fits within a customer's requirements. We save weeks, if not months, compared to early last year when we were first starting to build these relationships.
Awesome. So Brian from Clear Street, another good friend of ours, focusing on execution and replication is paramount. Were these the key gating factors in choosing your partners?
Definitely. Partners from a customer perspective of can we grow with them on a demand signal? We know there's demand in the market today. We know that customers need capacity. But are they customers that are willing to scale with us, or are we just giving up this campus for a one-time deal? That was important. Then who are the execution partners with us that can grow and scale? I'll give a shout out to JP Morgan. They have been by our side through various customer conversations that have worked, that haven't worked. They've given us guidance on some deals that we walked away from because we didn't believe it was a right long-term decision for us, and they supported us.
And so the partner, and if we think about JPMorgan, Goldman Sachs, their balance sheets and their abilities to be able to scale with us in terms of our capital needs from a project financing perspective. Then when we think about EPCM design build partners from an engineering and construction perspective, Jacobs is one of the only publicly traded companies within this space, and they have over 40,000 employees worldwide and have shown a commitment to want to grow and to scale. And you'll see that through some of the quotes in our press release with the vice chairman of JPMorgan, with the CEO of Jacobs. And then lastly, Vertiv as well in terms of the solutions to deliver.
They're one of the most well-known data center supply chain delivery solutions that's full stack across the data center industry, and their commitment to us to deliver these long lead time items and this equipment to make sure that we're able to construct and build for the customer.
Awesome. Okay, so we've got three more super important questions here that a few people are asking. Maybe I'll lob this one to Sean. So Sean, why exactly didn't we tap the high-yield market for this deal? Maybe walk us through that in great detail.
Yeah, absolutely. So when you're thinking about large construction projects and large infrastructure projects, the project financing market is always going to be the most efficient. So if we break it down into its constituent pieces, from a cost perspective, if we're at SOFR plus 225 right now, that pencils out to a little bit below where SOFR is. That pencils out to a little bit below 5.5% from a cost of debt perspective. High-yield bonds we've seen issue in the market were 7% plus, right? So we're more efficient there. Second, and this is really important, there's no negative carry associated with these deals. So what does that mean?
If we were to go issue $2.5 billion of high-yield bonds to finance this, and I'm just using a round number so that people can dig their teeth into it, and that was at a 7% interest rate, from day one, we'd have to start paying $175 million in annual interest. The way our loan is structured is a delayed draw, right? So we only use the debt when we need it. So if we have to make a big payment, we work with the bank operations folks, and we draw that capital down as and when as needed. 85 cents of every dollar we spend is from debt. And so there's a significant cost savings to us with that reduced negative carry expense. So those are the two key things.
And then I think sort of more subjectively or less numerically, working with the banks, this is our first project, right? And it's important that we nail this. And so having folks like JPMorgan and Goldman Sachs alongside us working through the lease, the engineering programs that we put together gave us a lot of confidence that they've seen a lot more of these than we have. So that was also a very validating point. And moreover, we take market risk off the table. This is done on their balance sheet. It's committed, pending some final execution. And so that's another really important piece because market risk is something the market is very volatile right now. So that was something we really wanted to take off the table.
Awesome. Super, super interesting. So a few questions on this one as well. Can you discuss one from our friend Chris Brendler at Rosenblatt? Can you discuss any concessions that may have been made in order to get the full 15 years backstopped without granting any equity or warrants? We talked about differentiators of this deal versus what's been done in the market. Can we double-click on that one?
Sure. Look, we just had open and honest conversations with customers. We said, "This is what we care about. This is the value that we believe these assets should trade at. And this is why we believe in our ability to execute, and we should not get a discount for being a first-time data center developer." And we brought in, obviously, partners to help build that conviction and that growth from a financing perspective, from an execution perspective. And I think we were okay saying no to deals.
We were okay saying, "If this deal works for you all and it doesn't work for us, we're not going to do it," and obviously, in the short term, that hurt a lot as deals were getting announced and people were saying, "Hut 8, where's your deal?," but for us, I think we felt like we knew what fair deals looked like and what other data center platforms were getting for their sites and their campuses, and so it was very simple.
It was, "What are the economic terms and structure that are equitable and fair between us and customers for the risk profiles and for the asset value of the assets that we have today and the timelines that we have to deliver them?" Then from a structural perspective, how does the structure of these agreements, the risks of each of those agreements, how the economics flow, how does that align with how our business is built, the responsibilities that our different business unit leaders have in managing their own P&Ls and being able to drive that into, "This is the structure that we are offering, and this is the kind of terms and contract we're offering. And if this is not the right site for you or the right opportunity, that's okay.
And we'd love to build a longer-term relationship together." And I think that approach has taken more time, but as trust has been built, I think that approach has done us well. And we have very open, transparent discussions with customers. And everyone wants to come and say, "Can you deliver me a data center in six months?" And you have to talk through, "This is what we believe we can commit to. This is the buffer we're adding for ourselves in case there's some execution or construction delays and risk. And we would love to deliver it to you early. And this is how we're approaching it." And I think that type of radical transparency has done us well with our shareholders as we were building last year. And I think has done us really well with our current customers and our prospective customers as well.
Okay, super interesting. Okay, so we've got time for about two or three more questions. So this is.
I have a couple as well, Sue. So we're going to let it rip a little since everyone's been waiting so long.
All right, for sure. So okay, next one is from Darren at Roth. "If you compare the market when you started the permitting process in Louisiana over a year ago to now, how would you characterize the overall demand environment? And are you seeing improvement in that market demand in real time?" I'll lob that to Sean or Asher.
Yeah, we're seeing a ton of demand in the market. We approached the Louisiana campus. We actually had two customers at first, and one of the customers couldn't get there for regulatory reasons. This was kind of end of 2024, and then earlier this year, we worked with a hyperscaler that was looking at a cloud availability zone, and then a demand signal just changed, and so we kind of changed our approach in how to develop campuses where we have customers that we're talking to where maybe multiple campuses might meet their needs, and we're able to kind of work in parallel where we're not losing the value of those megawatts by being locked into just one conversation. There's a commitment from the customer to drive towards the finish line as well.
And so for right now, I mean, since we've started this endeavor over the last 24 months in terms of these demand signals around AI data centers, we see demand signals increasing. We see focus on execution increasing and giving confidence on the ability to deliver on schedules increasing. And I think a willingness and appetite to historically work with developers that hadn't. And I think for us, when we were able to dive into the actual engineering discussions, the power discussions, I think the value and knowledge and depth of our team really were able to shine and come to light. And that's what's given us ability to have, I think, a very strong and healthy agreement today that allows us to give the customer what they need and allows us to build a great platform of digital infrastructure and power for our shareholders.
Awesome. Side note, George Sutton would like to give a shout-out to Sean and say thank you for wearing the tie today because this is obviously serious business. So you're welcome, George.
We have to please, George.
Okay, the next one is from Steven at Jones. He asks, "Initial data hall Q2 2027, how many gross critical IT megawatts is that? Any color on the additional data hall timing over the second half of 2027?
Four data halls. First data hall is 65 megawatts. The remaining three data halls are 60 megawatts, total of 245 megawatts for the campus. The extra 5 megawatts is the network stack in the first data hall. And so you'll see kind of the cadence of data hall one coming online, and then you'll have a fast follow within 45-60 days for the second data hall. And you'll see that cadence of data halls coming online.
Okay. And then from Mike. And then Asher, after this question, I'll let you choose a few if you'd like. So Mike from Northland, our good friend at Northland, wants to know, "Would Tranche 3 use the same partners outside of FluidStack and Google, or would it be different?
We really like these partners. We spend time talking to a lot of different folks. The answer to that question is a bit nuanced. The first answer is we've already been working with them on other campuses, and we've asked one from a Vertiv perspective, "Do you have the supply chain that can support us?" From a Jacobs perspective, "Do you have the actual labor, talent, and a team to be able to support the delivery of this project?" and we've had some open, real, honest discussions, which make that partner even better because they'll give you the truth, and we can work through that together. The other element that I've told them is I'm someone who likes to build deep relationships and partnerships with counterparties, and if they treat us well and we treat them well, then we scale and we grow.
And so that's the expectations going into these partnerships. We structure them to build a program to scale rather than just a single opportunity. And that's kind of the approach and the expectation set. If we don't receive that or we're not giving that, then I don't expect that to continue to grow and scale, and we'll have different folks. And to the extent that we can have the same partners that have the infrastructure to grow with us, we are looking to continue to scale. And we've chosen these counterparties because of the scale that they have today and their ability to scale with us across multiple campuses.
Very cool. Okay, do you want to pick the next one?
Yeah, happy to do so. There's some questions around how did you get the Google backstop without offering equity warrants? This goes back to the open and honest conversations of what a deal would look like that would work for us versus that wouldn't. I think it's finding the right places where we can concede because the customers have something they need and we need to deliver for them versus what we feel like is fair market and economical for our shareholders. How much debt financing is needed? I think we address that as well in terms of the build costs are around $9-$11 million a megawatt on the 245 IT, looking at 85% loan-to-cost financing with JPMorgan and Goldman Sachs.
Okay, I have one here that I think we should dig into, so from Alexander Whitelaw, "Compared to what we've seen in the market by peers who have recently signed AI deals, this deal on a risk-adjusted and commercial basis seems significantly superior. How do you go about educating shareholders and the market on the long-term commercial attractiveness of this deal?
We were talking about this in the office today, right? Because we were saying, "Look, if we announced the deal a month or so ago, we probably would have had a lot more appreciation on the day of announcement than today where the markets have been a little more bearish." I think at the end of the day, what we're solving for is what does this company look like in five to 10 years from today? And as we sign these long-term commitments, will we be able to deliver on the promises of those commitments to our shareholders and to the business that we're trying to build? And so I think these contracts, I mean, across all of these counterparties, were thousands of pages of agreements that we went through.
The nuances are in the terms and conditions of each one of those agreements, how they're structured, how risk is fairly allocated between counterparties. Those only come out via execution. So I think for us, for example, when we announced Google is providing the actual financial guarantee and backstop to these obligations, we didn't put out a number of what that is because they're securing the whole thing, right? So I think the market will have to ingest this. We gave a lot of information today. We gave information about a strategic partnership. But even on the data center side, we didn't just give the terms of a data center. We went much deeper into the execution of that data center and why from day one we're starting to deliver and to execute. We don't need to figure anything else out at this point.
We just need to execute. So I think for us, we understand that this was a lot of content for the market. As they digest it, I think they'll only feel better and better about the quality of this deal and the confidence that they have in our ability to deliver for the customer, make them proud, and be able to scale with them as well.
Yeah. Okay. So there's been a few questions around this, and this might be one of our last questions. "Will you sell Bitcoin to fund build-outs? Also, has there been any change in strategic thinking around Hut's working relationship with American Bitcoin? Do you still plan on leveraging American Bitcoin to monetize energy at future site purchases? What does that look like in a nutshell?
Yeah, I'll start at the last part of that. Answer is yes. We are a power infrastructure platform. We're powering compute, compute that's supporting the Bitcoin network, compute that's supporting AI technologies. A great example of that is our Long Hill campus in Corpus Christi. That's a gigawatt. We would have never underwritten that campus for an AI data center because of where it's located. We underwrote it because we knew American Bitcoin had the demand for that project, and there were other potential use cases. Today, there is demand for that campus from an AI data center perspective. So that shows that value and that opportunity. We actually have eight sites on the data center side with American Bitcoin that customers are looking at and looking at potentially taking over.
And so American Bitcoin continues to be an ability for us to scale and grow our development pipeline and have demand that's natural and captive within our platform. In terms of our ownership stake, that company continues to grow. It's been growing its Bitcoin yield. We see it as a free call option for our shareholders here that is non-dilutive. And it's a business that we have a lot of conviction and growth in. Because of that conviction and growth, our equity in Bitcoin optionality and upside is through our equity in American Bitcoin. That means that Bitcoin on the balance sheet is not something that we're religiously committed to holding, especially now as we've announced the first deal. And so we will look at the best ways to maximize the value and the return on invested capital for our shareholders.
And so we'll have more updates there in terms of how we think about the balance sheet and the composition of that balance sheet as we continue to grow into this development pipeline.
Awesome. So let's do one more question. "How do we think about backup power supply, micro grids, turbine solutions for some of these projects? How are we solving for that?
You're adding more variables and more complexity. So simple answer is if you don't need to, that's the best case scenario. The more nuanced and real answer is the grid is getting strained, transmission is getting strained, and people want sites that are larger and larger by the month. And so building on-site generation is becoming more and more of a topic and more and more of an opportunity as well, right? We have pipelines that run through our campus at Riverbend. We have pipelines that run through our campus in Corpus Christi. And so those used to be a liability historically because you had to build your building far enough away where you didn't want to be close to it. Today, they're an asset because if the pipelines have access to capacity, you can tap into it and build generation.
You have the ability to control generation growth, and you have to kind of work with the local utilities and their regulatory requirements around behind-the-meter generation and how you would actually structure those agreements. But you'll see that more and more a part of campuses as the demand for larger campuses scale. But ultimately, if people can get grid-connected power, they will prefer to do so. But available grid-connected power will continue to be scarce, and the speed at which it comes online will continue to take a long time.
Awesome. Okay. So Asher, should we wrap it up here? I think we've covered quite a bit, and we got to get back to work.
I think so. I really appreciate everybody for being here today. I appreciate all the partners that made this possible, all the local stakeholders within Louisiana that supported us from day one and our team for really busting their butts and working relentlessly to make this a possibility. I mean, the work really has just begun. Now we got to deliver the data center to our customer, and for our shareholders, we've got to show them that this is the first of a repeatable program that we can continue to grow and scale, so I appreciate you all for coming midday and listening to us, and hope you all continue to follow our journey.
Guys, the recording of this call will be available. If you have any more questions, sue@hut8.com. Happy to chat with you anytime. Talk to you soon, guys. Bye.
Thank you.