$WULF Q&A and business update with @TeraWulfInc Management

Hosted by @Perry Lin · 2026-03-09 · Tags: WULF

TLDR

TeraWulf management presented a strongly bullish transition from Bitcoin mining toward large-scale HPC and AI data-center infrastructure, backed by long-term contracts, scarce power access, and deep energy expertise. The team expects more than 500 megawatts online by the end of 2026, a Kentucky tenant agreement near term, and over $1 billion of run-rate NOI or EBITDA exiting 2027, while prioritizing execution and debt reduction.

Speakers

Notable quotes

Transcript

Perry Lin: Hey, good afternoon, everybody. This is Perry Lynn. I'm just waiting for Patrick to connect. I think there might be some difficulties there. Let me just see if he can get on. Sorry about that, everybody. Just waiting for Patrick to sign back on. Nazar, are you there?

Nazar: I am here, Perry. Good afternoon.

Perry Lin: Hi, thanks for joining. I'm going to introduce you to Nazar's co-founder, CTO of TerraWolf. I want to thank the entire team for joining. It's not every day that we get the management team of a publicly traded company on. So greatly appreciated. Thanks for sharing your insights. I see Kerry's on. I don't see the Terra Wolf account on. Let me just see if I can get him on here. OK, I sent I sent Patrick another invite.

Nazar: Maybe we can get started, Perry, and then Patrick will.

Perry Lin: Yeah, I'm sorry about that. I don't know. Sometimes you got to log back out and back in. Spaces can be a little glitchy. I want to get started. So 3 years ago, like, you know, a company like Terra Wolf, what you guys are doing didn't really exist. Now we have several HPC AI data center co-location type plays out there. For somebody that's not that familiar with the business model, they see a company that's still trying to gain profitability and may have some debt on the balance sheet. What are you telling the investor community? Why invest in TerraWolf?

Nazar: So maybe I think it may make sense to take even go three, four years before that to help highlight what we're doing today and how that ties to TerraWolf's story. Before TerraWolf, we were in the power generation business. We own over 5,000 megawatts of generation capacity, both thermal and renewable, domestically in the United States, as well as a number of overseas jurisdictions. And so we ran around developing, repowering, constructing, operating power generation facilities. And that's what we had done for quite some time. We noticed changes in the profitability of that business, right? As a wholesale generator, we were selling wholesale electricity on the grid. And the view that we had come to was just the business of selling wholesale electricity was challenged. And we started to think about how do we use our skill set and expertise in different ways. And we looked at data centers back in 2018, 2019 when we first had this thought. The data center business was a very different business than it was today. And so we ruled that out, and we got into Bitcoin mining. And so Bitcoin for us was really an expression of monetizing a kilowatt hour of electricity, understanding where that power is coming from, where we can procure a lot of it, how do we get so at an attractive cost? And the first couple of sites we had were zero carbon resources, both nuclear and hydro. So for us, again, this Bitcoin mining was really an extension of what we had been doing on the power side, which is rather than not putting power onto the grid, we were really pulling it off of the grid and running it through an ASIC and generating Bitcoin hash. And if you fast forward to the demand that has come from AI, it's a further extension of that. We're continuing to pull power off the grid. We're building the building to utilize the power. We're not owning the hardware. We don't own the hardware. We are tenants that are coming in on that hardware. But the same skill set and expertise that we've had for 25 plus years around power, power flows are ever present now in what the underlying customers are looking for. And so our role is really the procurement of power at a site and then putting all of that together in a data center building and being able to efficiently finance, construct, and deliver that to our customer. And so where we are today is we've signed over 500 megawatts of critical IT load over the past year. close to 10% or so of that is online today. The balance is being delivered over the next few quarters. As we go through the year here, all of that is scheduled to be online prior to the end of 2026. And the cash flow profile that we'll have at the end of 26 will be very different than what things look like today. And so a lot of the debt that you see today has been raised off the back of those long-term contracts. FluidStack and Google were the most recent transactions that we completed. So for investors learning about the Tarot story or trying to understand it, it's really we've got long-term contracts with high-quality customers going on 10, 15 years. where we have a fixed amount of revenue we can earn, 85% cash flow profile off of that. And that's all supported by equity that we've had from the company and that debt we've raised. And that debt, again, as I said, has been more than covered by the contracts we have in hand. We are guiding the market that we will continue to sign another 250 to 500 megawatts of contracts per year. Those signed contracts usually deliver 12 to 15 months after we sign the agreement. And so that cash flow that's coming off from those contracts, we are projecting to continue to increase, year over year in those 250 to 500 megawatt chunks.

Perry Lin: That's great, Nazar. I think Patrick is on. I just want to welcome him. Hi, Patrick. Hey, how are you doing?

Patrick: Hey, just I heard Nazar's answer, which I thought was great. Just one other thing I would add is, and you know this, but a lot of folks don't, but I think TerraWolf has very unique management shareholder alignment vis-a-vis a lot of our peers. You know, when you look at the board, the management team, and all the insiders, we collectively own, I think, you know, between 20 and 30% of the equity of the company. And I don't think you have that level of alignment really anywhere else in the space.

Perry Lin: No, that's great. That's great. Something else I've noticed is none of your peers has the diversity of site. and power markets as you guys do. Maybe you can walk us through the existing sites and the progress that's going on there. I think you have 5 sites currently across 4 power markets. Maybe we start with the OG, the Buffalo site, Lake Mariner. I don't know who wants to take that. Maybe tell us what you have left, when it may come on board, the extra power, and then talk about the future plans for the Bitcoin mining over there.

Nazar: Great. I can start. Patrick, please, chime in here as well. So in New York, we have two sites. The OG is Lake Mariner. That's where we've signed a bulk of the capacity that we've contracted so far. About 500 megawatts of gross capacity we've signed up under these long-term agreements. We're in the queue for another 250 megawatts of power at that site. We're expecting to hear back from the ISO in the first half of the year. And we're expecting that, we'll be able to add another 250 megawatts of gross capacity to that site in outside of Buffalo. That's the Lake Mariner site. And that's where the construction that I'd mentioned earlier is in full swing there. We've delivered Wolf Den, CB1 and CB2, CB1 and half of CB2 to our customer. The balance of CB2 will be delivered prior to the end of the month. And then CB3, 4, and 5 are all leased to FluidStack and their underlying customer. So that's the OG site in upstate New York. The second site we have in New York is Lake Hawkeye on Lake Cayuga in the Finger Lakes region. We have in-hand approvals for a little over 100 megawatts of capacity. We are waiting for feedback on our interconnection request to get the site up to 300 megawatts. Again, we're expecting that interconnection request feedback on that by midyear as well. And that will, you know, that will give us 300 megawatts of gross capacity to play at the Lake Hawkeye site in upstate New York. Moving on from there, we have a joint venture with Fluid Stack down in Texas, in Abernathy, that's outside of Lubbock. It's 84 megawatts of critical IT capacity net to us. And that is in Texas, but it sits within the SPP service territory. So it's Texas, but not ERCOT exposure in Texas. And then we announced a couple of transactions here earlier this year. One was in Kentucky, in Hawesville, Kentucky. That was a former aluminum smelter run by Century. We acquired the site from them. 480 megawatts of gross power that is available now. That power used to feed the smelter that Century is running. We are going to repurpose that power to run through the data centers. We are in discussions with customers as we speak, and we're targeting delivery of the data haul capacity there in the second-half of 2027. And the final site is in Maryland. It's the former Morgantown coal facility. It used to be the home of a 1 1/2 gigawatt co-pired facility that was retired. There's currently a little over 200 megawatts of operational capacity at site. That's liquid fueled. And we are envisioning to bring 1,000 megawatts of load, 1,000 megawatts of Gen. and 500 megawatts of battery storage to that site. And we are expecting the power to be available at that site in late 28 going into 2029. I don't know, Patrick, if there's other things to add there.

Patrick: Nope, you covered it.

Perry Lin: I'm going to dig a little deeper. At Lake Mariner, how much power do you have dedicated to Bitcoin mining currently?

Nazar: We're currently running around 150 to 170 megawatts of Bitcoin mining capacity. And so we, the infrastructure that we had there was for over 200 megawatts. We've repurposed some of that equipment as we've been ramping up the AI side. And as the AI capacity ramps up, we will likely see a corresponding ramp down in the Bitcoin capacity at the site.

Perry Lin: Got it. That was my question. So we don't have to wait until the halving in 2028. That's probably going to be transitioned a lot sooner.

Nazar: Yeah, I mean, as capacity comes online on AI, you will see that. We, the kind of the current fleet efficiency we have is in the 17, 18 joules per ter-ash range. And so it's a decent fleet. And yeah, we will kind of continue to run it as we have power available. But as that AI capacity ramps up, the Bitcoin mining megawatts will ramp down.

Perry Lin: Got it. And then at Hawkeye, I think is the first phase still 100 megawatts? Is that kind of the target?

Nazar: That's correct.

Perry Lin: OK, so potentially I'm guessing you're not going to go seek a lease until you have the power secured. So I'm guessing that's 2027 ish or 2028 where we can see the Lake Mariner and Cayuga leases?

Nazar: If you look at just the customer discussions that we've been having, we've been having discussions around, let's say in that 250 to 500 megawatt gross capacity zip code. And with the pending interconnection requests at Cayuga, you know, when we have 300 in hand, we think it'll be a much better and deeper discussion with the customers that we've been talking to around the capacity. So So I think once we kind of get a sense from the ISO where that the timeline for those incremental approvals that will then dictate timing on the back end with respect to when we can energize that site.

Perry Lin: Got it. The most questions I get is, you know, about Kentucky, because I think that's going to be your next leased site. And, you know, everybody's asking when is that going to be leased, when is that going to be leased? What kind of customer are you looking for, and are you seeing improving lease economics?

Nazar: Sure, and I think Perry Paul is on as well, so if you want to get him to chime in here, he's he's here to jump in. But Kentucky, you know, the power's available, and so those discussions with customers are underway as we speak. That power translates to about 380 megawatts of critical IT load. We are guiding the market that we expect to have that capacity online in the second-half of 2027. Typically, it takes us 12 to 15 months to turn a contract into fully deployed and online capacity. We're sitting here in March. And so, you know, we think over the next couple of quarters here, we will be able to announce a counterparty for the capacity at in Kentucky. And again, that 380 megawatts of critical IT load is squarely in the target range of all of the usual suspects in terms of how much capacity that they'd want. So those discussions are underway as we speak. And over the next couple of quarters here, we would expect to be able to announce a customer at the site.

Perry Lin: That's great. Is there a preference? Are you looking for like hyperscale or diversity, or would you prefer Google FluidStack just take that down because you're familiar with them? Or is it whoever's going to give you the best terms?

Nazar: Patrick, you want to take that one? Sure.

Paul: Actually, I'd like to take it. Could I take it?

Perry Lin: Welcome, Paul. Surprise guest CEO. Terrible. Paul Prager, welcome.

Paul: So a couple of things. I just want to, you know, provide a bit more background because I think Patrick and Nez are really humble folks. First of all, the Kentucky site was found by Kerry Langless, our CSO, who I think is on, and her team. And it's a bit of a unicorn because it's immediately available power. It's got four interconnects. It's a great location for the hyperscalers. And What we like to do here is focus on the execution story. You know, the whole team is really an energy infrastructure group. The C-suite here has worked together for over a decade managing power plants, some of us for as long as 25 years, like Nazar and I, together. So we like to focus on execution. The way that that is enabled is if we don't have to worry about surety of payment, And so Patrick's focus is exclusively on the highest financial credit to be our customer. So we solve for the customer on the basis of highest financial credit and culture, meaning can we work with them in Google and, you know, the end user of the sites that we have. And, you know, I think We have partners who are collaborative and who work hard with us to ensure we could deliver at the spec and the schedule they request. You know, when you have clients like Google or folks like, you know, if you will, a company like Anthropic, what you have is you have best in class. And these projects are highly complicated. These are really difficult things. So inevitably, you're going to have bumps in the road. You want to have a customer whose culture is oriented towards working collaboratively with you. So when we think about the customer for Kentucky, we're going to want to work with a customer that's got the right culture and the highest level of financial credit. That enables us to then just focus on building this thing to meet their schedule and performance requirements. So that's how we think about our future customer, and we will land that customer before the end of the second quarter. We have term sheets in the house, and we're negotiating them real time. Second, I just wanted to get back to some of the things that NASA responded to. One thing that we're really pushing hard is the notion of regional security, and therefore regional diversity. So we love the element here that we have facilities we're bringing online in Texas, Kentucky, New York, Maryland, because I think that's critical, not only to us as a company, but ultimately to our customers. So, you know, you take a look at what's happening in the unfortunate conflict in the Middle East. Data centers are a target. You know, we like the notion that we are diversifying risk for our customers. And then the third thing, which you haven't gotten to yet, but you will, and Nazar will be best to respond to that, is how the business is evolving from the standpoint of, you know, bringing your own generation. Nazar could speak to our qualities there as power plant owners and operators. You know, we've had six gigs of power. But I think as the business evolves to that, you either have to bring your own generation or you have to have nameplate capacity. to give you some leverage in getting into the queue. So I think that's what's exciting about all of our sites. At Cayuga, we have the ability to add generation. At Kentucky, we have the ability to add our own generation. In Maryland, we will be a surplus generator. So as you think about our siting, think about optionality and generation, regional diversity, and then again, focus on our execution orientation, which means we want to sign up the highest quality credit customer. In Kentucky, we've already announced we're working. with Fluor as our EPC provider, that just reinforces our ability to deliver to meet the needs of a high-quality customer. So that's how we're thinking about the business.

Perry Lin: No, that's great, Paul. That's a great segue into the Maryland site. I wanted, you know, housekeeping question on that. I know there was a report out from the PJM possibly, or maybe I may incorrect on that, but do you think that deal's gonna close? Are there any hiccups on the approval on that?

Paul: Listen, FERC is the ultimate approval. I'm sorry.

Perry Lin: Yes, I'm sorry, FERC.

Paul: FERC is the ultimate approval authority. I think what you have in the United States, which is what makes it great to be here and live here, is everybody's got an opinion and everybody wants to express it. FERC is focused on market power. We obviously don't have market power or the ability to manipulate pricing. And that's what FERC will make the determination on, not about really anything else. The fact that there are parties that want to say, oh, we're worried about what they want to use for end use there or how they will repower or what will be the ultimate, you know, fuel source or fuel mix. It's all good. Everybody's entitled to an opinion, but that's not really where FERC focuses. Genon has sold this to us. They believe we're the right buyer. We believe we're the right buyer. We believe FERC will make that determination. And at this point, while we will respond to any concerns that have been aired in front of FERC, we don't think that's really what FERC authority is all about, so we're highly confident we will close.

Perry Lin: Oh, great. Oh, great. That site, just from the PR, I don't think everybody understands. So you're going to be running a power plant, right? So those 200 megawatts are not going away. That is still going to serve the local market. You're just going to add capacity to 500 megawatts of nat gas and some battery storage. for your data center customer. Is that about, is that what you guys mean by bring your own generation?

Paul: Sure, Naz, you want to take that?

Nazar: Yeah, that's correct, Perry. The existing generation will remain online and we will be adding significant incremental capacity to the market in the form of that efficient gas power generation and the battery storage capacity.

Perry Lin: Yeah, that's, I think that puts Terawulf in a unique spot because everybody's talking about bringing your own generation. I don't think the hyperscalers really want to run the power side of things. They're partnering with a lot of, you know, utilities and companies like Constellation, but those guys don't really understand the data center side. So Terawulf can do both the energy and the data center side. Do you think there are other opportunities other than running your own sites, maybe joint ventures or a services business where you can provide that expertise without having to bring capital to the table?

Nazar: So we're not big fans of service businesses. We are principals, and so we want to use our skill set and expertise for projects that we're working on. So we will express that in doing more projects and bringing generation to other sites that we may be looking at. And so as you said, Perry, being able to bridge between what the grid is looking for, what these different regulatory jurisdictions, how they operate, the grid ISOs, how they work, the market structure, and then bridging that with the data center customer. Because at the end of the day, all they really want is the ability to get their hands on operational compute. And partners that can solve that for them and show up and turn over to them operational compute, Those are partners that they will continue to work with. And so whether it's at Morgantown or other things that we're looking at, we are increasingly having the discussion around what type of incremental generation can we bring and how can we support it. And as you noted and as I mentioned earlier, that's been part and parcel of who we are for a number of years before TerraWolf even started.

Perry Lin: No, that's great. I know you guys have a slide in your deck about your pipeline. but you just added these two sites. That takes you probably for the next three years, you're going to be pretty busy. Do you, are you still looking at sites? Is the pipeline still full in that phase three, phase four quadrant?

Paul: Yeah, so maybe I could start and maybe Carrie could get in here. This is an execution business. As I said, once Patrick locks down that world-class financial credit is customer, it's our job to build and deliver. And that's really where Nazar, Kerry, and I like to focus. I think Patrick leads the way in terms of transparency in our reporting and in control over everything that ops does. But we have wood to chop here. I mean, I'm in Texas on Monday, I'm in Somerset, New York on Wednesday. You know, we dealt with, you know, again, these projects are complicated, right? We were dealing with 30 knot winds in upstate New York, and we were trying to fly steel. I was there. We bought magnetic shoes for our troops up on the roof so that we could sort of continue to try and close in a building to get the mechanicals in. These are not easy projects, so the focus must be on execution. I think, again, that's why floor in Kentucky is a tremendous value add. So I think last year, Carrie's team, but, you know, she'll correct me. I think they probably looked at 200 projects and we ended up bringing home two. Nazar has let the market know that we are entirely comfortable with the notion of our ability to put in the ground 250 to 500 megawatts a year. Can we do more? Yeah, I think we can do more. But each project that we bring on, we need to make sure that we could deliver on because you are only as good as your last job. We have equity with guys like Anthropic and Google and G42 because we deliver. The day we don't, we got a problem. So my view is we need to make sure that we are executing, focused on the jobs that we have under contract and under construction, and that anything that we choose to bring on, we have got to have absolute charity we could do that. But Carrie is still driving hard with her team to manage the Ford pipeline. And it's not to foreclose the notion that we may want to bring on projects that would increase our annual, you know, stated projection of 250 to 500 megawatts. Carrie, do you want to add to that?

Carrie: Sure. Thanks, Paul. And hi, Perry and everyone. So since Paul talked a little bit about how we're viewing and pursuing the pipeline, maybe a helpful context is how we think our tenants are thinking about it and where the demand is right now. So I think it's clear to everybody that the key constraint in the market right now is access to power. Customers are still very much prioritizing sites that can deliver power in 2027. That's why Kentucky for us, and I think for the market, is really a unicorn and so valuable. There are very few locations that have that type of capacity at scale that can deliver in 2027. And where it does exist, you know, it's typically going to command a premium and will move quickly. So we're, you know, as Paul already mentioned, we're really confident on the discussions we're having already in Kentucky. And I think the site sort of speaks for itself and its value, and we're seeing that firsthand. Because 2027 availability is so limited, most of our new discussions now are sort of shifting towards 2028. And so we are actively looking at sites that may not be the unicorn that we have in Kentucky, but may start to deliver things that we can fold in in 2028 and beyond. I think hyperscalers in particular are increasingly willing to secure capacity further out to ensure that they have access to these large scale sites. Now, when you start thinking about that, when you think about deal structures, I think that also goes hand in hand with the timing of power delivery. So when 2027 power is available, customers are more likely to pursue a higher yielding sort of turnkey build to suit, excuse me, structure. Well, you know, when you're talking about an opportunity that might be 2028 and beyond, you know, our tenants might lean towards a powered shell or maybe a lower CapEx execution given the longer development timeline. So I think that sort of now gives you both what we're seeing on our side in terms of how we're pursuing the sites that we find most attractive and then the conversations that we're having with our potential tenants.

Perry Lin: So it doesn't sound like they're slowing down at all. They're still looking for sites and full bore ahead with securing power.

Carrie: Yeah, I mean, I would say And I think Paul, Patrick and Nazar would agree. I think the demand is insatiable. It's it's lining it up, sequencing and executing that are the important factors now.

Paul: Yeah. And remember, it's not just hyperscalers, which is where, you know, Anthropic's in the market, Amazon's in the market, Meta's in the market, Oracle's in the market. They trade probably a few hundred basis points. You know, the debt's a bit a bit more expensive. So not You know, it's something that we would have to consider. Google is aggressively in the market. You have the NeoClouds in the market. G42 wants more capacity. CoreWeave, Nabius, all these folks are there. So demand seems unabated. But again, Patrick, you know, he enforces a very strict discipline in terms of what we're willing to do in terms of the customer. because it has to fit our business model and he is driving towards the highest quality credit.

Perry Lin: That's great. I got a question for Patrick. Every one of these deals, the equity portion is financed with some convertibles. Is that structure still the plan going forward or do you think there are other options available?

Patrick: Great question. It's a tool in the toolbox, but I think when you look at where we're trying to drive things long term, yeah, I mean, obviously, if, again, it's one tool in the toolbox, Perry, but if you think about where we want to be in three, four, five years, that's about four times levered through the parent. So directionally, okay, You assume, as Nazar said, we exit 2026 with 520, I believe, critical megawatts contracted. Again, just like using very round numbers. I mean, Wolf Compute does 600 million of NOI in year one. Approximately, again, and I'm rounding numbers here, you know, 300 million of that is for Amortization, 300 million is kind of for interest. Again, round numbers, it's lower than that, but let's just make it simple. But then if you add, you know, another project like Kentucky, you know, that's about the same amount. So literally, like when I look at it, I think high level, we're exiting 2027 with over a billion of EBITDA. So today, right, I've got. 2.5 billion of principal amount of converts at the parent. So that's two and a half times levered. And if you look at Wolf Compute, right, we've got 3.2 billion of leverage there on call it 600 million of NOI. So out of the gates, I'm five times levered and I amortize down by about half a turn every year. So, you know, at the end of year three, you know, I'm in the kind of three times levered zone. to where those bonds are callable on purpose after year two, because ultimately, I want to move those bonds from the operating company up to the holding company so that we don't have any constraints on moving capital around at the subs up to the parent, and then B, about four times levered at the parent. And by the way, I've seen this movie firsthand in energy infrastructure, and it happened when I was at Blackstone with Cheniere. Now I think the same sort of progression where you started with 15% pick preferreds, then you went to bank debt mini perm, then you went to operating company debt and then holding company debt. That's where we are going and that's the roadmap we're following. And so is there room for more converts under that? Yeah. Are we going to have another, you know, 2 1/2 billion of converts at the parent? No. So I do think You know, there's, and as as Paul pointed out, you know, if you take our existing Wolf Compute deal, the customer matter because you're, you know, Wolf Compute is about 75% loan to cost, right? Total costs, like all fees and everything. And so you look at some of the other deals that have come out recently with very strong credits and they're 85 to 90% loan to cost. So Again, I think to Paul's point, you know, credit worthiness of the customer really matters because that influences the profitability leverage you can take on. But the end game, again, is to have a company that, you know, has moderate leverage at the parent and maintains an investment grade rating.

Paul: And hey, Patrick, just while you're talking, I mean, maybe it'd be helpful to explain how we think about the value in terms of our market value of the Kentucky deal. Do you mind going through that a little bit?

Patrick: Yeah, I mean, the short answer, right, is Perry, and you and I have done this math before, but you know, my view is very simply, you know, Kentucky's worth 10 bucks a share to the company. And again, let's keep it like super high level doing cocktail napkin math, okay? So It's 480 total megawatts. That's about, 380 critical. But again, just bear with me. Let's round and say it's 400. If you take the economics that we disclosed to you very early on in some of our first deals and what you've seen generally from some other folks out there, I think a very safe bet is given where leases have kind of are moving to now and where lease economics, if you take 400 times 1.5 million, per of revenue per megawatt, and that's critical megawatts, that's 600 million of revenue. 85% net operating income margin, okay, that's 510 million of NOI. And let's, again, round it down just to make numbers super easy. Really high quality tenants should probably command a 20 times multiple. 20 times 500 is 10 billion of enterprise value for that asset. And it probably cost me, again, round numbers, about three and a half billion of debt to finance it. So you're talking six and a half billion, six and a half billion of net equity value. And at today's stock price, we have around 550 million fully deleted shares. So again, just doing the math, that's 1180, but let's round down. Let's say it's 10 bucks. That's the basic math. So when you were asking, customer about customer quality sites, other things, Kentucky is and Morgantown both are incredibly valuable to this company because they move the needle in a massive way in the near term. And so focusing on those types of opportunities, I think as we move forward, that's more where I think our attention will be as opposed to like some of the JVs or other things that we've done in the past.

Perry Lin: And those sites are expandable. I think you've talked about bringing your own generation there, right? And growing those sites.

Patrick: Correct. And I think they are expandable. They do have the ability to bring on generation. And that matters because we guide to 85% net operating income margins. Candidly, I think we will beat that over time. I'm not going to change my guidance because I want to be able to beat and raise, but that is the value of big sites. You know, if you have very big chunky sites, you only have one security team. If you have four or five sites, you've got five security teams. Like, you think about what's in that operating cost that gets you from 100% to 85%, it's really chunky. It's really just property taxes, insurance, and labor. And the biggest piece of labor, candidly, is security. So, yeah, economies of scale matter.

Paul: Perry, you know, one of the things that's bothered me is that, you know, we've always said not all megawatts are created equal. And we have been really firm on that. And we are especially sensitive to it when people talk about pipeline. Because for a long time, people would talk about, oh, I have a two gig site in deep west Texas, you know, and therefore, you know, retail who haven't had the chance to see the kind of disclosure that Patrick does, you know, they're like, oh, they have a two gig thing. But the reality is, is you can't take a two gig load, right, and plop it at the end of this very long, skinny little transmission line and think that it works. That dog doesn't hunt. The grid operator can't handle it. The grid is not robust at this point. to manage that kind of load in Deep West Texas, not when they require 99.99% quality electricity. The beauty of Kentucky is it's got four interconnects. The beauty of the Maryland site is that not only is it in a robust grid, but I mean, there's a transmission superhighway right there on the site. So I think it's important when you value some of these pipeline opportunities that you really look at the node, you make a determination whether or not it's likely that you could actually put a big load on that and that the grid operator and the grid can handle it. And that's the beauty of Kentucky, that's the beauty of Maryland. Uniquely so, you know, you wanna talk about these projects in ERCOT, we don't have one there, But you're talking about a system that's 88,000 gigs of capacity with 280,000 gigs of demand. That means it's like musical chairs. They're going to be a lot of folks that are left out. And right now, some companies are talking about how they have that opportunity. But the reality is it's not a legitimate opportunity because either they're not in back zero, so they're not going to get the allocation for a while or worse. The transmission needs to be built out. Transmission's a regulated thing. It takes time. And without it, you're not going to be able to manage those kinds of loads. So you really have to look at each and every single one of these projected, you know, opportunities from the perspective of, is there customer interest? Is there real generation there to meet the demand? of a customer? Is there fiber available? And can the grid handle that kind of a load? It's the beauty of regional diversity, and in particular, these new sites that we've brought on.

Perry Lin: No, that's great, Paul. And as a segue, and Patrick knows this, I've been harping on revenue per gross megawatt for the longest time because just the economics in Texas aren't as good because the PUE isn't as good. And, that's the other downside of, hot weather and a site in the south. I think Anthony had a question, so I'm going to bring him up.

Anthony: Hey, can you hear me, Craig? Cool. I have a couple of quick ones. I heard you guys talking about this, but when do you, when will you know more if you're getting that additional 250 megawatts at Lake Mariner? And then also on Lake Cayuga or the Cayuga site, let's just say, on that one, when would you guys know more about the power situation there? Because I know we've talked a lot about Kentucky recently with that new site. And of course, the other news site in Maryland, but more on those two sites, if you guys can give a little more clarity on that. And I would assume if you get another 250 at Lake Mariner, if you take the Bitcoin mining off, you'll have like another 300 to 400 megawatts of capacity at Lake Mariner. Is that correct? Yeah.

Paul: So why don't I try and respond first? Maybe, Kerry, you could help. You know, with respect to Cayuga, it's a very exciting site. We got through the Zoning Board of Appeals. We redeemed a permitted use. An interesting element of that site is we have a lot of land and it's another deemed permitted use is power generation. So we are currently working with the planning board there, which is that's the protocol. That's the normal thing. As we review. you know, where we want to get to on that site. Currently, you know, there's, I think, around 130 some odd megawatts available. And our intention is to build that out larger. But we also want to think about whether or not we want to add generation to the mix, because that would enable advantageous permitting timelines. given the state is keen to encourage and incentivize generation. With respect to Lake Mariner, I think we'll get our next 250 here. You know, the way we've managed that is we've gone 250 at a time. Carrie, I think we're due to get our next 250 here near term.

Carrie: Yeah, I think we're expecting to get feedback mid-year, so call it summertime. We should have more visibility in line of sight on the next and final 250.

Paul: And with respect to the Bitcoin mining, listen, I like Bitcoin. I think it's great. But as a business, and all of us like it, by the way, NASDA was first on board. It was a very exciting asset for us. And it was really helpful because it was a flexible load, gave us an opportunity to express in a profitable way our megawatt capabilities. Terrell's future is high-power compute AI. We don't want to be in the Bitcoin mining business. We are currently doing that at site solely to provide support to the grid. You know, think about that. New York, a very robust grid. sophisticated grid operator. They want us to continue to mine there so they could cycle us as they need to to buffer the grid as they bring on new high-powered compute AI demand. I would imagine we're out of that business within the next 18 months.

Anthony: Okay, so when it comes to the Lake Mariner site, potentially additional, let's say, co-location deals there by 2027, is that possible?

Paul: If we continue to execute on site, which we are, I think Nazar can talk about our delivery schedule for our existing customers on site, we will be able to replace the load of Bitcoin mining with high power compute and AI.

Anthony: Yeah, that's that's what I was thinking. Okay, that was perfect. And when it comes to the Cayuga site, is that a 28 thing you're thinking now timeline wise or 27? I was just wondering kind of timeline because I understand the timeline for, of course, the Kentucky site and the Maryland site, just understanding the timeline more so on the Cayuga site.

Paul: Yeah, it's pushed out from 27 because of the near-term opportunity we have at scale in Kentucky in Hawesville.

Anthony: Makes sense. Thank you.

Perry Lin: Yeah, one more thing for Patrick. I don't know if everybody realizes, but these leases, NASA's going to be delivering these sites by the end of the year. That means like come January of 2027, you know, Terra Wolf's going to be spitting off probably, what, 50 million in free cash, give or take. What kind of uses are you looking for that? Are you, I know you have an authorized buyback, pay down debt.

Patrick: Yeah, debt, look, I would say primary focus, you know, now is debt reduction and then continued growth at other sites and funding that growth. So yeah, but again, I think, Perry, one of the things I think that gets lost on folks is, you know, the Wolf Compute financing is really a, like once the assets are up and running, Like I said before, I mean, you have 300 million of debt amortization every year, right? And so all of that accretion like flows to the equity, right? And where the really kind of big unlock for the equity happens, again, is in year, call it, you know, the bonds are callable after year two. So that literally we put those bonds in place in October of 25, So that's literally a year and a half from now, that is their first call. As you can imagine, once those are up and running and we've delivered that capacity and the Google credit backstop is in place, candidly, we should be investment grade. And so investment grade, right, is very different than the seven and three quarters that we got on the Wolf can be ponds when we brought them to market in October. As you've seen, that was the first of its kind. And imitation is the greatest form of flattery. A bunch of our peers have sort of taken that show on the road and copied it. And we couldn't be prouder of that. But the funding costs have continued to come down. And so I think when we look at it, the huge unlock that happens for the equity is when we go from amortizing seven and three quarter debt to literally bullet, probably at the parent or at an intermediate operating company, i.e. like five to seven-year non-amortizing maturity bonds, then to your point, now you've got, again, if I've amortized down to call it roughly 2 billion, now you're talking about a spread to the underlying tenant and or back. stop, which is Google A+ rated. And if I'm 25 to 100 bps behind them, again, I'm going to round up just to maybe, that's 5%. So 5% on 2 billion is $100 million, as opposed to the 600 million that I just walked you through. So that's where the cash flow and massive equity unlock for terrible folders happens. And that's really not that far away.

Perry Lin: That's A distinction that I think people need to understand because there's a lot of tailwinds, as these sites come online.

Patrick: Yeah, look, and Perry, to your point, that's one site. So we're telling you that we also think we're going to have the capacity at Kentucky, again, which is around, you know, 380 critical megawatts. Again, just to do that cocktail napkin math, I rounded up to 400. But again, you know, that's another 500 plus million dollars of run rate NOI exiting 27. So again, you put that together with Wolf Compute, even if we don't execute anything else, which again, as you've heard Paul, Nazar, Kerry all say, highly unlikely. I mean, you're exiting 27 into 28 with over a billion run rate NOI.

Perry Lin: No, that's great. We're just about out of time. Does if anybody else has a question, I think we have time for maybe one more. Someone raise their hands. Otherwise, I'm going to have Paul maybe close us out. Final thoughts.

Paul: You know, I was at the Cantor Fitz conference and, you know, it was a bunch of great meetings with a bunch of institutional masters and somebody said to us, Hey, I like you guys. If I was going to take a second position in the space, which of your peer group do you like a lot? And Kerry made fun of me because I said, well, if you want to take a second position in the space, why don't you just buy more Terra Wolf? I think I think Patrick, I think he's brilliant. I think he's led the league in terms of transparency. Nazar, Kerry. their teams. Again, we're together a really long time, Yasir and I, again, it's 25 years. So I think that we will continue to find, you know, the gems that are out there. I think we're the only team that has the ability to sort of manage power generation. I mean, it's what we've done. We're not a sort of data center company. We're an energy infrastructure company that builds the data centers to our energy infrastructure. And so I think it's a good, you know, it's a good chemistry that we have. We work with collaborative partners and we're really focused exclusively on execution. And I can assure you, you know, I'm in the office seven days a week. If I'm not there, it's because I'm on site. Same is true of NAS, same is true of Carrie, same is true of Sean Farrell, a chief operating officer. Our site managers are world class. Our project leaders, world-class, you know, we appreciate your confidence, Perry, and the folks that are on this call, because we wouldn't be here if it wasn't for the retail component. They were first in recognizing value in the HPC AI space, and without them, you know, we'd be nothing. So we're always happy to get on with you and answer any questions you have. But I'm very excited about where we are. I think if we could just focus on execution, given the quality of the customers we have, we're going to win. And again, we're aligned with you. As Patrick said, shareholders and insiders or management and insiders at 30% of this company, we couldn't be more aligned with the shareholders. So we want to keep going.

Perry Lin: Yeah, on that note, I think one of your board members is buying on the open market at around $15. So that's a ton of confidence. Appreciate all of you. I was not expecting the entire team, so a surprise for our listeners. But great job, guys. Continue the execution. Look forward to the next deal, and best of luck.

Paul: Thank you very much, Perk.

Perry Lin: Thanks, guys. Thank you.