$wulf and Colo economics - Full Space
Hosted by @₿itcoin ₿utcher 🥩 🐑 🐷 · 2026-07-09 · 98 min · Tags: WULF
TLDR
Discussion covers co-location business models versus neoclouds, TerraWolf's Anthropic lease economics, power infrastructure advantages, and long-term growth strategy including potential REIT conversion.
- Co-location providers own long-life power and shell assets while tenants supply GPUs
- Anthropic deal valued at $19B over 20 years with 3% rent escalators
- Yield on cost targets mid to high teens with 80/20 debt equity structure
- Grid-connected sites near population centers provide superior terminal value
- On-site generation and batteries can cut power costs by up to 80%
- TerraWolf plans to self-fund future developments from operating cash flow
- Management views creditworthy counterparties and power access as key moats
- Potential REIT structure discussed after organic growth phase
Speakers
- ₿itcoin ₿utcher — Hosted the space and guided discussion on co-location fundamentals and TerraWolf deals
- Daniel — Presented detailed valuation model distinguishing co-location from neocloud economics
- Perry — Explained equity versus debt financing and flywheel dynamics for future sites
- Patrick — TerraWolf executive shared deal structuring, terminal value thinking, and on-site generation plans
Notable quotes
- “A colocation company owns the slow assets and rents them to whoever owns the fast ones.” — Daniel
- “the scarce input isn't land or capital or even chips it's energized interconnection on a known schedule” — Daniel
- “Patrick will say that their goal is to earn in the mid to high teens a yield on cost” — ₿itcoin ₿utcher
- “we are in their ecosystem for a reason because we recognized early on” — Patrick
- “My cost of equity is twenty-five percent” — Patrick
- “we welcome regulation we want it” — Patrick
- “By this time next year they can fund future deals from their own cash and operating income” — Perry
Transcript
₿itcoin ₿utcher: Good evening everyone. This is Bitcoin Butcher. And my co-host this evening is Daniel, who I've had the chance to connect with in the last few weeks after he made a pretty impressive YouTube video that caught the attention of Tyler Page and Paul Prager on co-location economics. Daniel, how you doing tonight?
Daniel: Yeah, I'm doing really good. appreciate you inviting me on here. I'm really excited to, talk with everyone and, you know, discuss more about co-location 'cause I think it's just quite frankly not talked about enough, and it deserves more attention.
₿itcoin ₿utcher: Yeah, so, briefly, and then Daniel, if you don't mind, in between speaking, okay, you got the mute down, that's cool. I got Anthony on stage, I saw Perry in the crowd, Crypto Miami, there are members of this community In Wolf that are way more versed, I think I've done enough homework to start a conversation, but certainly I'm interested in people who've been in this equity longer than I have, and, so I appreciate everyone being welcoming and, but I also- I've had success in the past, you know, eighteen months, just creating these conversations with multiple companies, and, this is no different, and I see an opportunity, and it even caught the attention of the TerraWolf team, and about thirty minutes into this, we're actually gonna speak with a member of management who I have an idea who it's supposed to be, but it might be someone else or multiple people, so really a unique opportunity, and I'm just- Just grateful for what X is giving me, just an opportunity to network with a lot of sharp people like Daniel or Anthony or Miami or Perry, and it's, it's really a blessing. And, so we're gonna hop into it, if Daniel, if you don't mind, let's just- Take this very high level for people who maybe are newer, and I envision this being a few minutes talking through co-location versus, the NeoCloud space, and then we can dive into the Anthropic and Wolf co-location agreement, and then maybe go over some other sites, and then like I said, if there are members of the community, I see Anthony requesting, and I'll get him up here that want to speak. Just, send me a request and, only thing I ask is people unmute when they're not speaking and not to interrupt anyone. So Daniel, the floor is yours. You're the man of the hour. You've caught the attention of two of the top CEOs in the co-location space. Why don't you tell us what you put together and let's talk about just conceptually. We don't necessarily need to dig into the model line by line as much as trying to understand at a very high level the- Difference in how a colocation provider operates as opposed to a neocloud, I think would give people some good contextual, a contextual framework to kind of evaluate their investments going forward. And just as a disclosure to everyone, I do own TerraWolf, and this isn't financial advice, this is just, a combination of entertainment and education and, what you guys decide to do with your own portfolios, I can't help you with. So Daniel, floor's yours.
Daniel: Awesome, and I'm gonna try to keep it, as high level as possible. I, I do have some notes if, you know, we wanna take it to, at, to some degree line by line, but yeah, I'm more than happy to talk about Colo and as a whole, and also, like you said, it's really great to have other speakers on here too because You know, I- people have probably been in Wolf longer than I have been investing, so it's gonna be really great to get their perspectives as well, because I view myself more as someone who's learning, taking in a lot of information, and I'm trying my best to contribute to Finex, and, yeah, so this is more of like a, a thinking model rather than, you know, anything else, and just a little bit about me very briefly, if this is- Everyone's first time meeting me. I'm 23 years old, I work in public accounting, I also have experience, in consulting as well, and before I was investing in the equity markets, I was investing in crypto, I was mining Ethereum in my college dorm. So like a lot of other people, I kind of went from crypto bro to AI investor. And yeah, so I'll talk more about The model as a whole. My goal in making this model was not to be a poster boy for any certain companies or, you know, the colo guy. This is really just a framework for analyzing a new business model and understanding how the economics of AI infrastructure actually work. So I kept looking at these colocation companies, and I just couldn't understand why people were valuing them on EPS, which is, backward-looking earnings, or putting tech multiples on them, or even trading them like neoclouds. You know, the colocation model looks and feels like real estate to me, and the difference is that the tenants are hyperscalers, and the scarce thing you're renting them isn't square footage, it's actually power. So I built what I think is a more practical way to actually value these businesses. So what the campus yields on what it cost against what the market pays for that cash flow and what that leaves for the equity. And that's what I'm gonna be talking about today. So as Butcher mentioned, let's talk about what it actually means to be a colocation company and how that differs from a neocloud. A colocation company owns the slow assets and rents them to whoever owns the fast ones. So they own the actual land, not the compute. There are no GPUs on the balance sheet, and GPUs typically have between three to five year depreciation clocks. Those, those clocks don't exist on the colocation balance sheet. They own the land, the power, the cooling, and the shell, and the actual tenants show up with their own compute and plugins. Right now It's a lot of plug and plays. and that matters more than it sounds like it does. Shells, substations, and cooling are between fifteen to forty-year assets, and as we see with Wolf, these leases can start at twenty years and go up to however many long, like, this recent one with Anthropic can go up to thirty years, and that's, you know, that's potentially three decades of, of contract right there. the GPUs inside Again, are three to five years, so, you know, these colocation companies don't have those liabilities on the balance sheet because, you know, these GPUs at the end of the day are liabilities, and, you know, they are expensive. And this premium that these colocation companies have comes from scarcity, and the scarce input isn't land or capital, capital, or even chips, it's energized interconnection on a known schedule. Everyone wants megawatts in twenty twenty-seven But not everyone can actually deliver them, and that's what these hyperscalers are paying up for because as we can see with, you know, even the recent events of a week or two ago, it's gonna get harder to actually- Have power online that you can deliver to them or to hyperscalers. so yeah, does that-- Would, do you think that's like a general good, like, I guess? Place for now, like if there's, I don't know if there's any questions or you wanna like add on to that.
₿itcoin ₿utcher: No, it's great. I, I think without taking, I figured a few minutes not to make this a half hour, hour on Colo, I do think we have a fairly educated audience, but on the off chance there was anyone new, I like to take a few minutes and I view my, I have a similar background, Daniel, I, I do cut meat for a living now, but I was in- Public accounting and restructuring consulting after going to graduate school, so more of a financial background by trade for about ten years, and then came back to my family business. but what I would say is now in my role, I interact with people in my business in downtown Detroit that they might be shopping with food stamps or they might be, a C-suite executive, so I, I feel like I need to be able to speak with everyone And level set with everyone, and that's what I take pride in, in hosting these spaces is trying to connect sometimes pretty, sophisticated financial schemes that, or strategies, I should say, that maybe Main Street doesn't understand, and how do you simplify it to make it a little more understandable and palatable as people make investment decisions for themselves? So I think that is helpful, I would say, On the Google earnings call actually this evening, and we'll have a chance to speak with management later, I'd like to hear more from them about Google mentioning on the call that they'll be seeking more third-party data center providers. I think that's really bullish for Wolf, given their prior experience at Lake Mariner with, FluidStack, with, Google being a, credit, sponsor on that specific deal. So, I- Again, a really bullish time in the space. I do think there is a place for these co-location providers. There's this term in finance of off-sheet, off-balance sheet financing where If you can find someone like Wolf or Cipher or any other colocation provider that's willing to front the CapEx for these powered shells and then provide colocation services, that's, that's less CapEx that a, you know, a company like Google or Meta or any of these hyperscalers, they can dedicate more of their capital to their existing business lines or to the GPUs that they eventually want to sell compute off of. So those were just some of my initial thoughts on that, Daniel. So maybe this is a good way to transition into the most recent Wolf deal and we can talk about their, the, some of the economics with, Anthropic signing for the Kentucky, Kentucky site.
Daniel: Yeah, and, one thing I am gonna do is this weekend, I, I'm gonna release the, the Wolf model I have. I'm still tinkering around with it and trying to get a hundred percent, but I think I'm gonna release it just to everybody so that everybody can, can have it and play around with it, because again, I want- I want to, I want to stay high level here 'cause, you know, like, you can get really into like all the numbers in this. So, so you just want me to go through like some of the, some of the tabs and some of like the main points, is that like more so the direction you wanna take it?
₿itcoin ₿utcher: You know, I posted your video in the nest, so anyone who wants to walk through the model line by line, I think it's there for their use. I think it'd be more just very broad, broad strokes, and maybe you can, we can kinda go back and forth, but at a very high level. And Patrick, there's another video that Patrick, I have, posted, Patrick Fleury, that being the CFO of Terra Wolf, in his interview with Mcnally Money. I was able to watch that interview again, and he does a really great job of fundamentally expressing what exactly he's trying to do when he structures a deal. So at a very high level, we said there's colocation, Wolf is the colocation provider, and there's a counterparty that needs compute. In this case, it was Anthropic, and the site in question was four hundred and eighty megawatts gross, which I believe ended up being four hundred one, megawatts of critical IT and those newer to the space, gross megawatts are the total site, whereas four hundred one, the critical- Critical IT, that's all that is available for Anthropic to plug in their respective GPUs. So, so Daniel, where I'm coming from is I think once you release that model, people will have context from this conversation to play around with it and by watching your video, but I think it's even more valuable to just say at a high level Like Patrick said and what you communicated is, they're generating rent, they're providing this powered shell, and there is no- with the exception of building the data center and providing the liquid and power connections to the GPUs on beh- on behalf of their customers, there's really not much management, Or, I should say, additional CapEx that they have to provide, and this was one of the pushbacks when, specifically Jim Chanos had And his model was putting in like a million or a million and a half of maintenance CapEx per year, but I think you communicated well in your video that Wolf is only responsible for building the shell and then what their tenant does with that blank box, assuming that there's cooling and that it's, the temperature's regulated and power is delivered at, ninety-nine point nine nine nine percent of the time, like that's up to the client. So I think that was kind of where I was coming from, and if you wanna speak to some of the talking points. in your video, like yield the cost at a high level or how they're financing it with the debt component and then the equity component, which As a proxy, the cost of equity that Patrick had communicated was twenty-five percent, so that was kind of some of my thoughts without actually digging in, if you're cool with that.
Daniel: Yeah, 100%. We can do that. And, I, one thing I want to talk about, that I think is super cool about these colocation companies is that These are really easy to underwrite, and you, once you understand how to model out a rent escalator, you start to understand each year how much they can be making off. Certain contracts. So when I release this in the pro forma tab, there's gonna be a, a full rent escalator, and I just wanna quickly go through the mechanics of that so when someone's looking at those, looking at this specifically, they can have some context.
₿itcoin ₿utcher: Hey, Daniel. I'm sorry to interrupt, do you mind just for people newer to real estate and data centers, can you briefly define what a right escalator is and why that would be in place in a contract?
Daniel: Yeah, a rent escalator is a clause in a lease that automatically increases the rent over time, and it's very, common in commercial real estate, especially for like data centers. rent escalators increase a property's cash flow without requiring new tenants or additional leasing activity over a fifteen to twenty-year lease, even a modest two to three percent annual escalator can significantly increase total rent rental income. And that's a, a big proponent of a lot of the business model for these colocation companies. so for example, we're-- let's say we've got nineteen billion over twenty years, that's the headline, and the initial instinct would be to divide nineteen billion over twenty, and you have nine hundred fifty million a year. Well, that's wrong, and it's pretty much wrong about thirty percent. So these leases escalate, and the rent in year twenty is much larger than the rent in year one. So nine hundred and fifty million is the average of a rising series, but it's not the beginning of it. If you average, if you use the average as your starting point, you're quietly handing yourself twenty years of growth that you haven't earned yet, and It's important to understand that because when we're valuing these companies, we want to understand what they're earning each year, not the average of a twenty-year lease. So you have to solve it backwards for year one, and that's what I did, and that's what I'm pretty proud about of, of this actual model, and I think it's something that, you know, the, the CEOs might have appreciated because, you know Taking, taking the details from the headline and then building this out, it would be-- is really good for every single investor 'cause once you have the table, it's, very easy to, To model. So if rent grows a little under 3% a year, then year two is bigger than year one, year three is bigger than year two, and so on. You compound that across twenty years and add up all the growth factors and divide nineteen billion by that sum, and I, I promise that's gonna make more sense when you're actually looking at the escalator. So you have about seven hundred and twenty-five million in year one, and that grows to one point two billion by year twenty. So take that seven hundred and twenty-five million, grow it forward twenty years Add up every single year and you land exactly on that nineteen billion you see on the headline, and the whole schedule ties out. So it's important to see that, seven hundred twenty-five million grow into that one point two billion, because at seven hundred twenty-five million, the rent works out to about a hundred fifty-one, dollars per kilowatt per month. Which is at the low end of, where hyperscaler capacity is clearing right now. If you'd use nine hundred and fifty million, you'd be implying almost two hundred dollars per kilowatt per month. So that's just the rent escalator portion. There's also, There's yield on cost that I go into in the pro forma, and then I, I even have a sensitivity tab that if you disagree with any inputs or assumptions, you can test everything out yourself
₿itcoin ₿utcher: Well, so Daniel, you skipped over-- you didn't skip over it, but I think you take-- you're taking for granted your understanding of the subject matter really quickly, if you could explain to everyone how that NOI relates to the yield on cost, that would be appreciated.
Daniel: Yeah. So I guess, do- Should I go over what yield on cost, what you-- people will know what yield on cost is? essentially it's Stabilized annualized net operating income over total project costs. And, I'm trying to think about the best way to, to frame it. So NOI is what the asset earns in a year, and yield on cost is, is what that earning represents as a percentage of what you paid to create it. So one's a dollar figure, and the other is the same dollar figure turned into a rate. I guess that would be the best way to, to describe that, unless you don't-
₿itcoin ₿utcher: Yeah, if you don't mind me, where I'm coming from is Patrick will say that their goal is to earn in the mid to high teens a yield on cost, meaning for every hundred dollars in project costs, they're hoping to earn between fifteen and nineteen dollars of stabilized NOI. And why that's important is that's essentially A cash return on the investment, the best way of thinking about it is that cash generated, the NOI is the net operating income from the property, which is the rent minus any cash related costs. In this case, roughly eighty-five cents on each dollar has been guided that for every, dollar of rental income that comes in, eighty-five percent of it will turn into NOI. So then that final NOI number that You're referring to around seven hundred million, if I heard you correctly, divided by the cost of the project will land you somewhere in the, you know, thirteen to seventeen percent range. But the, the point of it is where the value is created is there's the concept that you have in your model of a terminal valuation where If you have this kick-ass lease with Anthropic, if you were selling it on the open market to a institutional investor or, insurance company, whoever it is, they would be willing to take What's called a cap rate, like a return on their cash of approximately, anywhere from five to seven percent, which I think Patrick had said six percent in his example. So why that's important is you created all this value, yet they only want six percent. So the difference, that spread, is where the value creation is for taking on the development risk. So hopefully, do you have anything you wanna add to that? But that's how I think of it is, and this is actually gonna be a question if Patrick is the team member that joins us later, I would be interested in how they think about their portfolio, whether they want to just imagine This huge property portfolio, or if they're open to selling, properties off to institutions to generate more, cash to, you know, the, you have the flywheel and later Years with escalating rents, but you, you have, you have a lump sum of cash, they've shown success lately in acquiring sites, so I'd be interested in how he thinks about that. But that's kind of where I was coming from, Daniel, if that makes sense.
Daniel: Yeah, that makes a lot of sense, and I, I appreciate you kind of, you explained that really well. I guess it's, it's, that was harder for me to, To like voice, I think you did a great job. And again, like, essentially what, when I look at the model, it, it, in plain terms, a dollar spent building this campus is worth about a dollar ninety-six a day it stabilizes, and that's the merchant development, you know, you build at, one yield and the market values the finished, Asset. So, yeah, you explained that really well. yeah. I think
₿itcoin ₿utcher: let's, if you don't mind. Perry, asked to speak, maybe Perry has something he wanted to add to it, and then we'll come back to you.
Perry: Yeah.
₿itcoin ₿utcher: Perry, good
Perry: evening. Hi, good evening, guys. I, I would just say that think of this as real estate, and I think this is a good segue 'cause I get a lot of questions about this. the, the equity versus the debt a-and the flywheel. So, TerraWolf has committed to twenty percent equity in their deals. So think of that as the down payment on a house. W-when we're buying real estate, whether it's for our, our own personal use or for investments You know, we're not worried about what the whole house costs, we're worried about saving enough for the down payment. So, tariff is at twenty percent, some companies want more leverage. I think Cleanspark is talking about five to ten percent Leverage. So how they have funded these deals in the past, CB One, Wolf Den, you know, the first Fluid Stack deals, that, that was with convertible debt And then with the Hawesville, Kentucky deal that was just signed, they raised that through equity, so they sold some stock. So what is the ongoing flywheel? I, I think it's a, it's a pivot point at this moment In their development, because if you look at their balance sheet, they're gonna, they have, they have five hundred and thirty million coming from the sale of Abernathy and then by next year, we're talking about four hundred million in free cash flow, just from the revenue of their existing, of net ap-- net operating income from their existing sites that should be completed by the end of this year. And then, they have a revolver. And so what are they gonna need for their next deal, which is Muskie? It's a gigawatt, you know, at ten million, you know, let's just use ten million a megawatt of development as a baseline, you're talking ten billion dollars they're gonna need. That means for twenty percent of their equity, that's two billion dollars. Now, from my understanding, they're not gonna need all two billion of that upfront. it's, it, the, I think it's a, the, the first phase is five hundred megawatts, so they're gonna need a billion, well, they, you know, have pretty much close to that. So, they might need to do, if they're buying other sites, they may, they may need to use some of the, the cash on the books, but, they may not. They can probably phase into that. And so we're at the point where the, the flywheel will no longer require dilution and, and the, the company has matured. So that's where we're at. By this time next year, they can fund future deals from Their own cash and operating income. and, and most people don't realize that I, I still hear comments about, "Well, they're gonna need to dilute, you know, how are they gonna get the debt?" The project debt is easy. It's just a matter of what rate you're gonna be able to get. And, From everything I've heard, it's gonna be six to seven percent, and they're pretty confident they can get that on the anthropic deal. So kind of use that as your model number and, You know, they're not gonna have any funding issues or dilution issues. That's kind of the easiest way to think about the colo business is, what is their down payment? How much real cash does-- do these companies need to raise? the, the- That is the, that's always been the difficult part, and as some of these companies mature, they're gonna be able to fund it from existing operations. You know, Cipher's probably gonna be next You know, some of these earlier stage companies are gonna have, you know, you know, they're gonna have to find a way to fund that equity, 'cause they don't have it right now.
₿itcoin ₿utcher: So, Gary, let me ask you this, you'd mentioned the Muskie site, which is over a gigawatt, and then there's also Morgantown, which is another gigawatt, and then last that I saw was the Lake Hawkeye site. Do you have any, just from your vantage point and your intimacy with the details, any prediction as to which sites are going in which direction? And, I mean, 'cause right there, those opportunities- Opportunities. I think Patrick had said in the video just, having Lake Mariner and, the, Hogsville, site signed, you know, now you have two consecutive sites that are additional two gigawatts Just that shows the growth opportunity, and you mentioned how they're gonna have to raise equity and that the internal flywheel will be able to self-fund it, and that's promising. but just as far as Lake Hawkeye or those other two sites, if you have any additional details you wanted to provide to those that would probably help people newer to the company.
Perry: Sure. number one, I don't think-- they're not in the headline, but- All of these sites are bigger than what they have outlined currently. I don't think most people know that. so if you look at the future growth plans of the company, you know, they've kind of guided they're probably gonna add one or two sites in the future, one in Northern Europe, one in the Nordic countries, and probably one site in the US. And then they're probably gonna be done for a while, but, like Hawsville, there's another five hundred megawatts there that nobody talks about. at, at Muskie, it's, it's a gig that can probably go to two gigs. same with, same with Maryland. Hawkeye, so the Cayuga site kinda got delayed, n-n-not for the reasons people are citing, it's not for the moratorium or regulatory issues, but mainly, they were gonna phase into it eighty-four megawatts initially, and they realized that the tenants want bigger sites all at once, and they didn't want eighty-four megawatts and, and have chunks. So there's opportunity at that site, I think they're looking at Bringing on generation. And so that three hundred and sixty megawatts that was initially guided, I don't know what it's gonna go to, but it's gonna be more than that, could be four hundred and fifty, five hundred. and the Short to medium term, Hawkeye, I'm sorry, not Hawkeye, Muskie is the next site that'll probably get leased. We'll see them build out five hundred There, and then probably the second five hundred at Hawesville will come after that. Maryland is kind of up in the air. The, the, number one, they've gotta close that deal, and we'll hear in early August the FIR announcement. I don't think there's gonna be an issue. it'll either be approved or it'll be conditionally approved, where they have to make some corrections, and they'll get it approved in three months. the main thing with that site is they need to build a pipeline. A-along the rail, the railroad lines, the railroad tracks, and depending on how accommodating Governor Moore is in expediting those permits and easements, will dictate the timeline of that project, which is why they haven't given any guidance. but if, if It goes the way they want it to go, figure six months to get that pipeline done, that's probably, twenty, you know, give it a couple years, you know, twenty twenty-nine type of, type of project. So, I, I would say Muskie and Hawesville, and then the additional two fifty Lake Mariner, they are in the queue right now with New York ISO For more power at that site, and they can build and lease that out as soon as that's approved. Now, it's an independent agency, but it's really not, right? Will they, you know, will they approve the power with this moratorium going on? They can, but it doesn't mean they will or want to. So could that be delayed a little bit? Possibly, you know, six months maybe or so? But again, that they can issue that power immediately too, we just have no idea. So, y-y-y-- I, I would count, I, I would say a lease from Muskii end of this year, early next year And, and then the second 500 at Hawesville, you know, will probably be next, that's kinda how I see it, and then, Lake Mead are tucked in somewhere in between.
₿itcoin ₿utcher: I appreciate that, Perry. while you were speaking, it looks like member of the Terawolf team joined us. My understanding is it's, Patrick. So, Patrick, if it's you, welcome to the program. How are you this evening?
Patrick: Hey guys, doing great. Thanks, thanks for having me. I think Terry's gonna be taking my job soon. Was
Perry: I good?
Patrick: Was I right? Pretty close. Yeah, no, you was well done. Yeah. And I see, I think we got Daniel S on. I watched His video and he also as a contender for my job as well, I thought he did an amazing job, so thank you guys.
Daniel: I really appreciate that compliment, thank you. It means a lot that, you know, you and your whole team left that, positive comment, and I never imagined that you guys would even see that, so I really appreciate that.
Patrick: Yeah, of course. I mean, look, you guys do a great job, and, you know, I've been talking to Perry for a lot of years, and, you know, there's an art in, you know, this is, these are complex deals, but they're very simple. They can be distilled to very simple math, and I thought Daniel, you did, an exceptional job in your, video of doing just that for people, and that, that's not easy to do. That's why I have a job, and I, I try to Exceptional job. So, thanks for having me, and let me know where you guys wanna take the conversation.
₿itcoin ₿utcher: You know what, we were talking about the Colo model and you just, acknowledged Daniel's Credentials and explaining it, certainly we're willing to talk more of the finance, but I think a interesting direction. I'm not sure if you- Caught the headline from the Google earnings call, but certainly explosive growth in their AI business, and most notably for, Wolfholder's purposes, the, proliferation of third-party data center usage. So any initial reaction to that, as it relates to your business, Patrick, that you wanna communicate to the crowd?
Patrick: Yeah, look, I, I recognize how privileged, I am and we are to have them as a partner, right? I mean, they're one of the largest companies in the world, just incredible team, incredible execution in every single segment of AI, right? Competing heads up with Nvidia on hardware, but, but You know, as I, as I tell folks, you know, we are in their ecosystem for a reason because we recognized early on, if you think about the evolution of the internet, right? Internet one point zero was AOL, Yahoo, Netscape, Google, right? Google one. Internet two point zero was cloud, Amazon probably, you know, as, is, and Microsoft are the victors in that, and then, you know, Internet three point zero is Is, is AI, and I think if you're competing on the retail side, i-it's really hard to know who's gonna win, but I think You know, just and, and I say this, out of utmost respect, so I'm not, I'm not trying to, you know, downplay anyone here, but, you know, ChatGPT and, and OpenAI has a huge mountain to climb, right? 'Cause they have to compete heads up with Google. Google's a four trillion dollar company that generates a couple hundred billion dollars of free cash flow a year, and, you know, when we did our deal with them last summer, had, I think, thirty-three billion of balance sheet debt on a two plus trillion dollar company So competing with somebody that, again, is the eight hundred pound gorilla in their business has one, you know, and I, I say to people like, "My kids know I have access to Copilot, ChatGPT, and Claude, and they still say, "Hey, Dad, Google it."" So think about that, like, to, to usurp them You have to come up with a product that is mind-blowingly good, so it's definitely possible, but in, in our view, we were like, "Yeah, that's not probable," and we don't know, we don't, we don't know who's gonna win. So I think where we are partnered in our ecosystem took that into account, right? We have Google credits leaving our project in Upstate, New York. We've got Anthropic, who is targeting the enterprise segment, right? When you think about where does all the data fit that will be mined for- Decades, it's my data, your data, everyone on these, this spaces, event at JP Morgan or United Healthcare or Ford Automotive Company, like all the stuff we all use every day, that's where like most, almost all of the data exists and that is private data, and that will be mined on an enterprise basis to provide services to all of us For a very, very, very long time. So to us, that, that was an easier kind of model to wrap our head around from, from when we think about terminal value, not only for our site, but terminal value for our customer. and so back to your, your original question, yeah, I mean, I pay attention, obviously, you know, I'm in touch with Google all the time, they're our largest equity holder, they couldn't be a, a more supportive partner. And so to me, yeah, seeing that their CapEx is, you know, up five to ten billion, and they continue to fire on all cylinders, super important, but I, I would expect nothing less. That team is absolutely incredible. They're among the best in the world, and I'm really lucky and privileged to be working with them.
₿itcoin ₿utcher: As a follow-up question, I'm just curious, you've, recently assembled quite the portfolio of disclosed power, Patrick, and I'm just, given the speed that you were able to execute Hawsville, from, I believe the closure was in February, into execution into a lease in July with Anthropic, I guess my question for you would be, you've shown the ability with your operations team In the organization as a whole to work quickly. Do you view yourself more as a, developer of sites or do-- is the end goal to hold all of these sites and accumulate those cash flows to internally fund? How do you guys think about that? would you be open to ever selling any of those sites after you've leased it out? That-- I was just curious how you guys think about that.
Patrick: Yeah, look, it's a great question. I, I, I would, I would say a couple things. obviously, you know, if somebody comes along and bids for the company at, at a price we can't refuse, then, then, we'll do our fiduciary duties and, and maximize value for shareholders. I think individual sites, like that's not something, you know, that, that, we've really entertained yet, and that's 'cause, as you know, you know, our, our sites are kind of coming online now, And so in the future, if, if maybe we weren't getting what we think as like the right terminal value, optimized value for those, would we consider it? Yeah, but I don't, I don't, I think we have years for that to play. And like the amount of thought and consideration on our team that goes into terminal value, and I'll say terminal value to me means there's two distinct terminal values. One is the site. Like generally speaking, we draw a three hundred mile radius around the site, and, you know, if you do that in upstate New York, you hit twenty million people, you hit New York City, Toronto, and Boston. You can get to all those people within three milliseconds. You do the same thing at the Western Kentucky Site and the Eastern Kentucky sites that we own, also five, six NFL cities within three hundred mile radius, roughly twenty million people. That matters for terminal value sites. the other thing that matters, in our opinion, is grid connected Versus, a-and cost of power versus behind the meter. Al-almost.
₿itcoin ₿utcher: Patrick, I think we lost you for a second. You still there?
Perry: I think you dropped down as a listener, if you can just request to speak again.
₿itcoin ₿utcher: I'm showing him still as a speaker of mine, so I don't know if it's glitched. let me try that. Send him a co-host invite just to- see, I don't know if he bounced off for a second there. Do you guys, Perry, do you still see him as visible? He's on the center now, but he should be coming back. I, I see him as a speaker now. Patrick, are you back? Now I see Patrick on mute. Are you, are you there, Patrick? Damn, he was just, he was starting to heat up. I- Do you guys, I, I don't have visibility to- Now he's joining back on? I show him as a speaker, does everyone else see him as a speaker right now?
Perry: We show him as a listener. I don't know if you-- there's a way for you to refresh.
Speaker 5: yeah, I see it as a listener.
₿itcoin ₿utcher: I don't know. Okay.
Perry: X is always glitchy.
₿itcoin ₿utcher: I tried re-inviting him. I removed him as a speaker and invited him. I don't know if that'll work. Otherwise, I might just have to start the space over. Why don't I start a new space, everyone, and, hopefully, I can shoot that over to Patrick. Apologies everyone for the glitchiness and technical difficulty. I'm gonna start something new. Alright, we'll give everyone a second to rejoin. I'm inviting some people onstage. Chloe, I see in the crowd, if there's any way we can, I can send this link to you for Patrick, I would appreciate it. Let's try and give Patrick another minute or two. I'm gonna. Okay, it looks like Chloe responded that he has the link. And let's get the man of the hour back up here, and I'll get him on stage. Again, I appreciate everyone's patience. I'm in a good WiFi connected area, so I- Wanna, I'll take the ownership of it, 'cause it's my space, but, I think this is an X issue right now. Give him one other minute. But yeah, I, he was starting to warm up and he had a pretty good, they were in the three hundred mile radius discussions on how they think about their power, Access and to different respective markets.
Speaker 5: I see him in the crowd now.
₿itcoin ₿utcher: Okay, let's hope this works this time. Patrick, I'm sorry for- Elon, maybe he can rent some of your guys' power for more compute.
Speaker 5: And if it stays like that, Patrick, just completely close the app, like, literally. Hard close the app, so restart the app, sometimes that helps too.
₿itcoin ₿utcher: Okay, while we are waiting on Patrick to come back up I'm gonna, Anthony, do you have any reactions to what he's said so far, or Perry, given your guys' experience in the, or maybe-
Speaker 5: Alright, he's here.
₿itcoin ₿utcher: Okay, even better. So, Patrick, if you didn't hear me, I think there's a great opportunity for you to speak with Elon about your compute capacity, and that might solve this for next time. I apologize, but I'm glad you're back up.
Patrick: Yeah, sorry. Thanks for your patience. I, I, all of a sudden I was talking to, I think I was talking to myself and I was really getting going, so sorry about that.
₿itcoin ₿utcher: I think where I last heard you, and Anthony, you can correct me if I'm wrong, but you were getting into the three hundred mile radius and the access to the population centers with your respective power sites. So if you don't mind picking up from there, I'm sure you wasted your first attempt On the mirror, but we appreciate you, repeating it, Patrick. Yeah.
Patrick: Yeah, no problem. I think I was talking about terminal value and, and how we think about terminal value at a site level, but also customer terminal value. So site level, as I mentioned, is location. We think location matters, y-you know, not necessarily in one or two years, but in ten or twenty years, 'cause you gotta think about what inning AI is gonna be in, not just now, but in ten to twenty years. so being near population centers, we think matters, right? Connectivity, quick connectivity into population centers, and then in addition to that, being grid connected matters, because the grid has thousands of points of fail-failure versus being behind the meter typically, right, has a few. it obviously matters what kind of behind the meter generation you're running. Are you running combined cycle? solar, batteries, SMRs, peakers, that matters for a couple reasons, one, reliability, but then also cost, right? So generally speaking, if you look at like the industrial wholesale cost of power throughout the United States, it's about eighty to a hundred bucks a megawatt hour, across a lot of different jurisdictions. And so if you're building behind the meter just to get online quickly and so- and, and- And meet hyperscaler demand, that's great, but in ten, fifteen, twenty years when your lease comes up, if your cost of power is greater than, you know, or in the range of a hundred to a hundred and fifty dollars a megawatt hour, you know, our view is that's probably not gonna be competitive because the US, you know, will get its act together, we will build more generation, and it will be all hands on deck, it will be solar, batteries, wind, SMRs, combined cycles, peakers, and that, yeah, we think will keep The cost of power in, over the long term, in that kind of eighty to one hundred dollar per megawatt hour range for industrial users. And so if you have a site, again, that's not close to a population center that has behind-the-meter power that is much more expensive, our view there is that may be great for now, but it doesn't have a lot of terminal value in the future. So we have looked at, I would say, close to a dozen very large behind-the-meter sites, you know, anywhere from Five hundred megawatts to five gigawatts, and I think as most of you know, we are a power company, that is our DNA. we have not done one of them yet because we just haven't been able to get comfortable that, again, all of those things check the box, plus, you know, when you're dealing with behind the meter power, the, it makes the service level agreements, or what we commonly refer to as SLAs Where, you know, ninety-nine point nine nine percent uptime is one of those, you know, it's a lot more complicated when you're off the grid. And so all of our sites, for the most part, or actually all of them, sorry, not for the most part, all of them are grid connected, and most of them have the ability to add generation at the site But that generation wouldn't be behind the meter, it effectively would be front of the meter, meaning we would work with the local utility, electric cooperative, to build power, sell that power onto the grid, and then have our hyperscale client connect to the grid to buy their power for the data center. That we think is preferable ten times out of ten, as opposed to having behind the meter. So that also goes into when we look at sites and think about terminal value and the ability to expand at those sites, that is, you know, part of being good partners and understanding how, utilities, regulators, independent system operators work, and making sure we have enough space, enough grid access, enough infrastructure in place that we can actually do that. So those, I would say are, you know, when we think about terminal value site-wise, those are the things, like, right, does, it doesn't have expansion? Is it close to a population center? Is it grid power? You know, how close is the closest gas line? Is the jurisdiction one in which, it, that, you know, they would welcome these jobs, welcome the industrial expansion? That's, by the way, becoming, you know, a much bigger focus across the entire United States. It, ironically, It really seems that data centers are the only thing that has united seventy percent of the country, in, in their, in their, hatred against us, although I think, you know, like I, I tell folks that are anti-oil, then you better start walking or, or same thing with data centers, right? If, if you don't, like to order food or, or really do anything on your phone, then, then I guess you can be anti-data center, but if you do, it, it's one of those things that we're gonna require. So Yeah, that's kind of how we think about it, and then I, I did talk to you a bit about terminal value for customers, you know, and then credit quality and, and, not only just credit quality, but the underlying business fundamentals and target business for our customers is, is important to us as well. So I hope that, addresses that question.
₿itcoin ₿utcher: Yeah, there's a lot to unpack there. I'd, follow up question or two, if that's alright. as it relates to Paul's post with on-site generation, you had touched on it. Is it-- I'm just more asking to clarify my understanding, but if you were-- I'm assuming this would be some sort of bi-directional, grid connection where you could-- You were saying you could, produce power on site that could be sent to the grid, but you could also- So, import power on behalf of the data center, is that correct? And if that is correct, what are, some of the proposed methods that you guys are considering to do on-site generation?
Patrick: Yeah, so think about it, it's, it's, it's really, you know, not necessarily bi-directional, but it's, it's two separate things happening, right? So, if you think about E-each of our sites, like Lake Mariner, is two interconnects into the 345 kV system, so there's multiple redundancies. Hodsville, Kentucky, I think has five. And then Morgantown, you know, which we're waiting for, for approval on here, has nine. So that's w-when you think about why Google, right, and, and some of our partners are partnered with us, i-it's not our data center building prowess. I mean, Google's built more data centers than we will ever build in our lifetime, but we showed them something that they hadn't seen before, right? Because most of their data centers are just like our homes. Or offices where they're connected to a local distribution network. So if a, you know, you have a windstorm or if a squirrel crawls into a transformer and, you know, the grid goes down, your data center's offline, which is why they have backup diesel gens for those, you know, so that they can effectively cool down the data center so that you don't melt the racks and the chips. What, what we showed them was, you can be at a former power plant site or industrial site, have those multiple interconnects that we just talked about, so you have redundancy, but more importantly, you can be plugged into the artery of electricity in a region. So in New York State, like, that is the highest voltage system, starts at Niagara Falls, runs west east, crosses top of the state, hits another three forty five kV line coming down from Canada, and all that power tries to make its way to New York City. So when you- You're plugged into that level of the system, it's really, really difficult for the system to go down, because if it does, it, which has happened one time in the last fifty years, the entire Northeast grid from Maine to Chicago is going down. And so That's, I think, what we brought to the table, again, which is, again, we are a power company, that also, like I mentioned to you, exists at Hosville, and it exists at Morgantown, like that is a common theme in our sites. And so the, the, your question on, is it bi-directional? It's not like, I would say simply, and I'll just do, hopefully Daniel West will appreciate this, but I'll just do some basic cocktail napkin math to explain like why we think this is so powerful. Okay, so if you take, like, Morgantown as an example. If you are, if you have a thousand megawatt data center there, let's just say, and let's, I'm gonna use round numbers to make this math easy, and you are buying power in that region, probably at an average of a hundred dollars a megawatt hour There are eight thousand seven hundred sixty hours in a year. That's just twenty four times three sixty five is eight thousand seven hundred sixty. If you're paying a hundred bucks a megawatt hour, you just multiply eight thousand seven hundred sixty times a thousand times a hundred. That's eight hundred seventy six million dollars a year in electric power costs, okay? What we think we can do is Finance the data center, a power plant, and a battery, basically for the same price that, you know, some of the folks in West Texas, like Stargate, are building their data centers. But what you're getting, and I'll break this down, if you can build the data center for about ten million per megawatt, you can build the combined cycle power plant for two and a half to three million per megawatt, and then the battery costs, you know, one to one and a half. You know, you're all in, call it around fourteen bucks, per megawatt. You can wrap all that into a lease with the hyperscaler, because, again, what the hyp- and I'm gonna explain what they're saving. If you have a combined cycle power plant, those generally have what's called a six, like roughly a six thousand heat rate, that's what you get from GE or Siemens. If you're buying gas, your, your cost of fuel for that power plant in that region's probably around an average of six dollars an MMBTU. So six thousand heat rate times six, six bucks an MMBTU is thirty-six dollars a megawatt hour in cost, okay? And then you add for operations and maintenance of those facilities, it's roughly another ten to fifteen bucks a megawatt hour for, you know, labor, maintenance, service agreement, all that kind of stuff. So just round it, let's say it's $50 a megawatt hour for you to produce the same amount of power that you're buying off the grid. But you're buying power at a hundred and you're producing power at fifty, so net net, you just saved $50 a megawatt hour. Because that's the spread you're making on the power that you're selling. In addition to that, if you have a battery, so that during times of peak demand, from say three PM to seven PM every day, you can effectively reload that battery overnight when power's really cheap off the grid, and then sell that power back to the grid when the grid pricing is at peak, so meaning it's more like two hundred, three hundred dollars a megawatt hour, we think you can now reduce That net cost of fifty probably down to like twenty to thirty. So just think about that for a minute. So instead of having a hundred dollars a megawatt hour of cost for power Now, what I'm telling you is I think I can shrink that to probably twenty or thirty, and that's eighty percent savings on a eight hundred and seventy-six million dollar a year power bill. I don't care who you are, like, you know, Google, Amazon, Nvidia, Microsoft, Anthropic, OpenAI, that is really material savings. And you're doing, you're getting what you want, which is a grid connection, and you're delivering what me, frankly, And all of us on this call as a consumer wants, which is you're not being a parasite on the grid, you're providing the same amount of power that you're pulling, and you're being actually a positive net source to the grid via the battery in times of high demand. That meets what, you know, whether you're a Democrat or Republican or you're Trump or you're a governor or you're in PJM or you're in ERCOT, that's what everyone wants. And that's, to us, like, that's where the space has to go. You can't just be a parasite on the grid anymore, that's not gonna cut it. And we're seeing that right with the groundswell, of negativity against data centers. But I also think our job is to educate the public that You know, we have, we, we, we have a great analogy, for example, on water, right? Because I think, I hear a lot of complaints about water and, sound, whatever. I mean, the sound standing literally right outside of a data center is about 55 decibels. That, that is below, I'm sure, what most of us listen to music in the car at, okay? The amount of water that we use, up, for example, up at Lake Mariner. The average swimming pool in the US, if you have a swimming pool, is twenty to thirty thousand gallons of water. Day one, we put ten thousand gallons into the facility. And then through evaporation, I think we've had to put single digit, like I'm talking less than ten gallons of, of water in due to evaporation, and it's a closed loop system. So that's it, right? So again, our job is to make sure we educate folks, and make smart decisions, like I'm talking about where you can provide power, you can provide a battery, other things, and we can kinda wrap that all into a lease and everybody gets what they want. So that, that's, again, I'll pause there, but I think that's hopefully Daniel S will give me the seal of approval on, on that cocktail napkin math, but I'm sure he, he can probably explain it better.
Daniel: Yeah, you got my approval, that was awesome.
₿itcoin ₿utcher: Yeah, I think, Patrick, I think that probably addresses my next question with respect to the deficit in PGM specifically, it came out seven gigawatt shortage, and then, I'd already mentioned Paul addressing it, but just- The New York moratorium, but I think what you're saying is, given your guys' expertise and experience in the power industry and the ability to creatively, create this new on-site generation solution that Wolf becomes part of the solution as opposed to the being vilified by whatever NGO who's using AI to type up their attack article. Is that a fair assessment? Yeah, 100%.
Patrick: And, and not only that, I mean, we welcome, as you've seen in, in Paul's post, we welcome regulation, we want it, 'cause I think there are a lot of bad actors out there, and there's also a lot of folks that are getting into these power queues that, you know, Shouldn't be in them, and are creating real problems for utilities and independent system operators. So I think, you know, weeding those folks out by having, you know, requiring things like collateral postings. I mean, again, as a ratepayer myself, I love that, and I, and I'm, I'm happy to post collateral as TerraWolf because, again, I think that weeds out a lot of folks that, you know, aren't real developers or aren't creditworthy. And these utilities have to go, and ISOs have to go make real economic impact decisions. They have to procure generation, they have to start making transmission upgrades, they have to build substations. That's real cost to all of us as ratepayers on the phone. And so, yeah, I, I think we absolutely welcome what the governor is doing. I think, you know, our sites, it makes, candidly during this pause, our sites are more valuable, they're similar with moratoriums that have taken place in Amsterdam and Singapore Poor and all the existing data centers that were there, released them at much higher levels, increasing the terminal value there. But I think, you know, I'm hopeful that this is a short term pause that results in more regulation that defines, you know, create, causes folks to post collateral, do things that like make sense, because I don't think the answer is just to say no, that, that, again, 'cause this isn't only- You know, a race, amongst hyperscalers, it is a sovereign race, and I think the US, you know, as, as a nation, like we need, you know, we- We gotta rise to the occasion here, so I, I, I, again, I hope, cooler heads prevail, and we think they will, and that, and that the right regulations will be put in place to govern that.
₿itcoin ₿utcher: Anthony, I had a question for you next, Patrick.
Speaker 5: Yeah, Patrick, I appreciate you taking the time. I just wanna clear a couple things up for people. The extra 250 at Lake Mariner, that you guys have put in for, that wouldn't be slowed down by the moratorium, right? in a sense because- that, that basically doesn't require an additional grid connect, correct? Am I right on that, or wrong on that, or what's, what's that aspect?
Patrick: Yeah, you are correct, but let me get into technically why. So the way that the governor has chosen to police the moratorium is through DEC permits. DC's, I think the Department of Environmental Conservation. So they govern things like wetland, like if you're gonna impact on your site, wetlands, wildlife habitat, you know, knocking down trees, stuff like that, you generally require a DC permit. And at Lake Mariner, we actually got a, what's called a negative declaration in twenty twenty-two from the DEC, so we don't need to, to, to, in, in other words, to expand And for another building, let's say, 'cause as you know, we have CB3, CB4, CB5, so to go to like CB6, we don't require a DEC permit. So technically speaking, You know, we don't, Lake Mariner isn't impacted by that regulation, practically speaking, right? And I think when I joined, I heard Terry talking about this, you know, the New York ISO is, is, you know, we enter the New York ISO queue, and they work with National Grid, NYSERG, and all the utilities to kind of go through that queue and determine who can get power and when. I, I, I, again, I don't- It, it is possible that, you know, New York ISO said, tells us, you know, in the interim that we can have power. I, I think that's highly unlikely in my, in my view, 'cause I don't think the New York ISO will, act independently of what the governor is doing. So I think, we were kind of expecting to get a line of sight as to when we would get that power, like around now, from the New York ISO. So I think that sort of line of sight as to when- We will get power, and the reason that's important is as soon as we can tell our customers when we definitively have access to power, like what that date is, then I can go out and sign a lease and, and contract. Until I have that definitive date, you know, I'm not going to. So I think we're in a, in a period of time now where it's in a little bit of limbo and gray area, where I just don't know. So that, that's the honest-to-God truth, but again, just Your question directly, Anthony, we d-we don't need a DEC permit, which is how the governor is, policing that.
Speaker 5: Yeah, no, I appreciate that. I was gonna say 'cause it's very nice if you have the teams already out there to keep them rolling, right? That would be a beautiful thing. that being said, I have one more for you. what you guys did at the Hallsville site is very, very impressive to acquire the site and then turn it into a lease within six months. Are there additional sites that you're looking at that have the Map right now, which site do you believe is the next to go? Just for people that are unfamiliar with the company and kind of a little background on some of the sites that you have on top of what you guys are potentially looking to do.
Patrick: Yeah. So next site up is Eastern Kentucky, so that's what we call the Muskie asset. That is an amazing asset. it is Definitive power in third quarter of twenty twenty-eight through our utility partner, American Electric Power, AEP, one of the largest utilities in the country, their subsidiary, Kentucky Power, is, providing power to us there, so that starts in third quarter of '28 and then ramps up, To about a thousand megawatts in twenty thirty, that site, candidly has, has the potential to be much, much larger, So I think stay tuned on that point, but if you kind of work back from third quarter of '28, you know, we wanna have shovels in the ground Twelve to eighteen months, ahead of the online date, so that is, you know, first half of next year. So I think you should expect that, that, you know, lease kinda comes, you know, sometime between now, and, and certainly this time next year. And then just remember, you know, we had a really competitive process in Western Kentucky and, and a bunch of folks that were disappointed that they didn't get that site. So we- Moved those folks into the data room for Muskie, and they've been working, you know, for quite some time already there. So, you know, I'm, I'm really excited about that site and, and optimistic, you know, that, that gets leased, you know, like I said, with certainly within, within the next twelve months, if not sooner. So that's, I think, the next site for us, and then in addition to that, you know, we look at sites constantly, and, you know, there are some other sites that, that have risen to the top of, of that queue, and, I would say generally because of the way utilities and independent system operators are trying to figure out who's real and who's not, the price of, of acquiring sites has actually- Come down because the capital requirements are going up. And so, you know, developers that once were asking for, you know, hundreds of millions of dollars are coming kind of back to us and saying, "Hey, you know, I, I don't have anything if I can't post an LC, so my asking price now isn't hundreds, it's more, you know, much smaller, 'cause they know that they can't post the credit, we can, and if, if they don't post the credit, then the utilities basically- And like, move their site to the back of the line. So I, I would say there's, you know, more stuff, in the quiver, you know, ready, ready to come.
Speaker 5: Awesome, yeah. No, I, I can I add just one little thing and then I'll throw it back to you, Butcher. globally, are you guys thinking really expanding the platform around the world? I know you've mentioned, Europe, et cetera, but like any thoughts on that and what customers are looking for, and I know, the sites right now are all US, of course, which is great, but just wondering your thoughts on that and different, assets potentially to acquire all over the globe, 'cause you guys have a
Patrick: Yep. Yeah, look, we've owned power generation in Europe as well, so yeah, no, we are, we are spending a considerable amount of time in Europe. Europe has, you know, a real need for data centers. I would say they're, you know, a little bit behind the US, but not much, but, yeah, there's a lot-- our customers have a lot of interest in being there. you know, we have a lot of interest in, you know, geographic diversification, and so, yeah, I think, that's definitely And we're working hard to, to, to bring that home.
₿itcoin ₿utcher: Let's go to Shepherd and then Crypto Miami after. Shepherd, floor yours.
Shepherd: thanks, Butcher. and whoever's speaking for Terawolf, thank you, I appreciate your time. and, and if you've already discussed this, I only hopped on around five minutes ago, so you can just brush me off and listen to the recording. It is for New York. do you see any headwinds in terms of your Cayuga site, and have you already gotten that DC, DC permitting? would you see that parallel from the other New York site that you're just discussing with Anthony?
Patrick: Yeah, great question. So Cayuga's, on a bit of its own path. So, so Cayuga requires town and zoning board of approvals, permits. so that process there is i-in and of itself, one that we have to get through. So I don't think Cayuga will be impacted by, you know, the moratorium. My, my sense is the moratorium gets worked out, you know, in the next- Six to twelve months, and Cayuga's on its own timeline, which likely takes longer than that on the local level anyway. So, the short answer is Working, working through it, working with the local community there, and, and that, that one's on its own timeframe, un-unlikely to be impacted by the moratorium.
Shepherd: Okay, thank you. I've just been, looking in terms of, the, the local, and understanding there be-because I've, I've passed through those parts, but then also under-understanding, all the challenges there with the zoning in Lansing, but then, in New York. So I appreciate the, feedback.
Patrick: Yeah, look, I, I am, that part of the world is very dear to me. My in-laws actually live in Ithaca, so you can imagine, I get an earful all Very important to me, it's important that we're engaging with the local community there in all the right ways, and so yeah, that one is, it's, it's, you know, all politics is local and that, that is family with a capital F for me.
Shepherd: Yeah, yeah, my, my, significant other, went to Ithaca, so, so, so, so I've been, been through there, and if you're in Lhasa, you, you know, as well as me in, in terms of, the type of headwind, that's currently facing basically you guys, trying to work through it the best you can.
Patrick: Yeah, we are. And look, again, I think a, a lot of that falls on, like, we just have to educate folks and disp Like that's not helping. And look, I, I, I'm sure people have seen, you know, there's a lot of foreign bad actors that are involved in that right now, like pushing stuff through all the social media platforms as well. So it's, yeah, it's a challenge, but we're working through it.
Crypto Miami: Miami, you're up. Yeah, thanks, Patrick, long-time, Wolf investor. and if you wanna hear how crazy some of us are, a group of us attended the Lansing, New York Town halls on YouTube, so we, we actually witnessed the hours and hours of discussion that were going on and, and of course, their town band and really, on the team wolf part, it was really a masterclass In community, community advocacy and turning that whole situation around locally, even before this New York moratorium came around, like just to watch you guys work with the town to kind of, advocate and educate was really, was really beautiful. my question really has to do, as a long-time investor, 'cause you've talked about this in the past, and, and, and just for some folks who wanna think about holding Wolf long-term And that is, and you just, you talked about this a, a bit ago, which is, you know, long-term goals in becoming a reit. And I wonder if you had any ideas about long-term goals in becoming a reit?
Patrick: Yeah, for sure. Long time, it was funny, you know, I'm sort of myopically focused on the next twenty-four to forty-eight hours these days, but yes, I, I think, as you know, the management team, board insiders here, you know, we own twenty-five percent of the equity of the company, so e-everything we do is to try to maximize value 'cause we are shoulder-to-shoulder, you know, with everybody on the scale that owns shares. So yes, I think, you know, as we build these data centers, you know, we generate a ton of depreciation amortization, right? Which really is tax shield for, for the company and for you as shareholders. So I think we can I think we can actually act like a REIT before we are a REIT, and what I mean by that is, you know, we can, n-not be subject to double taxation because we have A-NOLs existing today, and then B, we have a massive amount of depreciation amortization from building these, you know, four plus billion dollar projects. So I think, the long, long-winded way of answering your question, yes, I think REIT is still, end goal. The, the difficulty with REITs, is, as you can see with like Digital Realty and Equinix, is it's really difficult to grow when you're in a REIT. So I think what will likely happen in this space and for Wolf is, I think we will have really significant organic growth over the next like three to four years. Then once we get through sort of all the immediate near term megawatts, it's gonna be you gotta bring your own generation, and that's just a longer game. So that's gonna take, you know, depending on what kind of, technology you're building, anywhere from two to four years. So that I think will slow the growth for all of us, and so then I think at that point you'll probably see consolidation among all of our names, right? 'Cause if you can buy someone, e-e- You know, and take out, you know, a hundred million of SG&A at a twenty times multiple, that's a lot of, that's a lot of savings for the deal, for shareholders. so I think we'll kind of go and, you know, I saw that cycle in the early two thousands with IPPs, then I saw it in the twenty tens with, oil and gas companies, and it's sort of a natural evolution, I think, w- once you kind of, squeeze a lot of the organic growth out of a sector, then you gotta consolidate Then you start looking at, okay, maybe we basically break, break this company into like two, one is the REIT, one is like a development company, right? And, and that can house generation 'cause REITs can't. but yeah, I, I do think like once you get- Critical mass, and honestly, I, I think our business, you know, I think Digital Realty and Equinix do a great job, don't get me wrong, but I think our business is better, right? I mean, they have older cloud, smaller data centers, can't handle real, real big AI loads. they have a lot of customers, like hundreds, that generally are on, you know, two to five year contracts. There's churn, so therefore they have higher maintenance CapEx 'cause they gotta, you know, when a new customer- One comes in, one goes out, they gotta, change the configuration, you know, whereas I have one customer for ten to twenty years, it's a brand new Ferrari of a data center, it can handle super high power density per rack, and I've got pretty low maintenance capex. So, And I've got a super credit worthy counterparty, right, as opposed to hundred. So if that, my point to you is, if those businesses, just because they're called REITs and they're financially engineered to pay a dividend, you know, are trading at twenty-five times EBITDA, there's no reason why we shouldn't, you know, well before we convert to a REIT.
Crypto Miami: Great. And then one just short term question, because this has been brought up in other interviews, and that, that, you know, with the Anthropic deal, having a credit, you know, having the investment worthy, counterparty, of course, I think, I think everyone in this space is, is well aware of Anthropic's position, and they're probably gonna IPO and be credit worthy. But just sort of like, you know, as we're, as you're talking about the, the next builds The cost of capital and, you know, when you go to market in six months to eight months to a year, what other, what other alternatives to, to capital will be out there?
Patrick: Yeah, and you know, that's a great question. I mean, I'm, I'm Constantly working with Morgan Stanley, my bank group, with the rating agencies on innovating new products, and I think, you know, stay tuned, I think you're gonna see some new ones. But look, the, the, the two mo- scariest commodities in this whole space, in my opinion, are power and capital, right? And you're seeing capital now, I mean, that's impacting, you know, Amazon, like the hyperscalers now when they're going to raise capital, like their spreads are moving wider, wider, right? Oracle's, you know, was, was downgraded by one of the agencies, you know, by a notch. Their CDS, but so yeah, look, I think it's, It's a super capital-intensive business, right? Now, to be fair, you know, when you're talking about one megawatt, right? It's about ten million for the data center portion, right? So literally the building getting the water and electricity to the rack in a, you know, temperature-controlled, humidity-controlled environment. The chips And the rack and everything are, depending on what technology you're using, twenty-five to forty million, right? So you're talking about thirty-five to fifty million dollars per megawatt, for this industry. I mean, it, it-- so the capital intensity is undeniable. I think we are really comfortable in the position we're in because we're in, in our opinion, the longest live portion of that, right? If you look around at energy infrastructure in the US today, like look at Schenir Or every single pipeline that was built, or every single nuke that was built, every single one of those big pieces of energy infrastructure has significantly outlived its initial depreciation period, or if you're a nuke, your initial NRC license or what have you, I mean, it's tw- instead of twenty, thirty years, they're all fifty, sixty, seventy-year type assets. So I think that's what we have as well, i-i-i-it is long-lived energy infrastructure. You know, the shorter-term, more capital-intensive part of it is Silicon and the chip, which is the responsibility of our tenant. but yeah, from, from a capital perspective, credit worthy counterparties matter, and I, as, as, you know, Daniel S knows, like the, the simple math I even tell, like whether we're dealing with Google or Anthropic or AWS or Meta, whomever we're talking to, you know, I say to them, "Look, I need a mid-teens yield on my lease." It's, it's simple as that, and they say, "Well, why?" And I say, "Well, because if I go finance this thing eighty-twenty debt equity, you know, my cost of debt, assuming it's an investment grade counterparty, is probably around six percent." So eighty percent times six percent is four point eight, let's round it up and say five points of WACC. My cost of equity is twenty-five percent, it's really expensive, and so, you know, a twenty percent contribution, twenty percent times twenty-five, that's another five points. So my WACC is five plus five, it's ten percent. So by definition, if your yield on, on the data center, your yield on cost, as Daniel points out, which is literally just your expected NOI over your build cost If that's not in the teens, or if it's below your WACC, and I know some of my peers who have used like expensive preferreds or others, like they're signing deals where, where the, the yield is below the WACC, like by definition there's no equity value creation. That's effectively, that's called commercial real estate, right? Commercial real estate is a game where, if they go, they try to get a hundred percent leverage. The leverage costs eight percent, but and, and the cap rate is, you know, four. And so you're like, "Well, how does that work?" Well, how it works, right, is over ten, twenty years, the building appreciates, and that's where they make all their money, because they've got a negative spread effectively on the, the, the, the cap rate versus the cost. So that, that, like, to me, i-is, it's just like finance one-on-one. Like, I have my- Is about ten percent. That makes the biggest portion of that, which is the credit, that much more important. It's got to be a credit worthy counterparty. And so, you know, you look no far, like this week, right? The last couple weeks have been pretty tough for this space in general, and I've seen, like, I've got, there are some folks in our space that are trying to finance projects in the market right now with counterparties that aren't investment grade, and, you know, just, you can watch, I mean, that's hard. That, it, it, you know, 'cause again, I just-- there's no immediate equity value. And again, given how much equity we all own in the company, like, I'm not interested in making a bet on appreciation over a five or ten year period. I want equity value now.
₿itcoin ₿utcher: As a follow-up question to that, Patrick, what do you think from speaking with institutional investors? Daniel had kind of mentioned this in his video and, certainly you'd mentioned DLR and Equinix that are able to trade at higher multiples or essentially you have counterparties that are willing to accept a lower cap rate because they view it as a safer investment. What is the- The disconnect with institutions and the rest of the street right now that can't seem to understand your value proposition of signing these credit-worthy hyperscaler tenants or someone like Anthropic with a credit-worthy wrapper on it, like where, where can you educate people or where is-- certainly you're educating us right now, and I think I, you already know that I'm on your side of the argument with it, but I just, it's, it's strange to me that Someone like your company, TerraWolf, is being treated like a neo cloud when it's not even close to what the value proposition is supposed to be, and just wanted to hear any thoughts you had on that or when you speak with institutions, what their feedback is.
Patrick: I, yeah, I would say institutions get it. we, we have spent a lot of time and energy over the past two years, you know, cultivating all the big long-only accounts, and we talk to them all the time, so I would say they all understand the value proposition, you know, generally speaking, like, you know, whether I would say they get that, you know, when I talk to them about, hey, here's my average NOI, and if we take a fifteen to twenty times multiple, you know, make an assumption around how much debt I need at Hawsville, and, you know, round my share, fully diluted shares up to, to six hundred million, like, you know, you're getting twenty to Thirty bucks a share, like with no value, that's literally just what we have contracted with no value for the pipeline. So I think they get that, I think there's a couple things that go kind of beyond that. One is these stocks are incredibly volatile, and so, and having, you know, sat in that seat myself for close to eighteen years It's really hard, right? If, if, if, you know, being down seven to ten percent in a day and then up seven to ten percent, I mean, the volatility is kind of insane when you think about the amount of value that we actually have contracted. And again, that's not just us, it's the whole space. And I think that's also indicative of just the market we're living in, where, you know, headlines about, you know, DeepSeek or headlines about Meta or whatever move these stocks by an incredible amount, Yet what I continue to see and we continue to see is, it's just very steady Eddie, like demand is off the charts and there's not enough power to meet demand, but the market bounces around on all these headlines. So I think the volatility, and the beta make it really difficult for institutions to have a really big position. So how does that get solved? That gets solved by us bringing capacity online and getting cash flow through the income statement so that people can see us delivering and see the real cash flow. So I- I think over time, that volatility will come down as we bring CB three, CB four, CB five, other things online, all of a sudden now you have a billion dollar EBITDA business, right? And that's, you know, exit, that's probably, you know Over a billion dollars first half of '28 of EBITDA, so that's not that far away. So I think we need to do that and just continue to execute and, and continue to translate those, you know, construction projects into cash flow, and I think as we do that, the volatility will come down and then the institutional stakeholders will get bigger 'cause they'll, they can withstand it 'cause you'll have a lot less volatility.
₿itcoin ₿utcher: I appreciate that, Patrick. we're pushing one hour of your time, and I know you're thinking twenty-four, forty-eight hours ahead of now. And as much as I'd like to continue this, I, I value your time too much as a shareholder to keep you all night. So I think on behalf of everyone, we really appreciate it. I hope we can do it again sometime. It was nice to meet you over X here, and, everyone really appreciated It's you and the fact that, you know, the twenty to thirty percent ownership by management, including yourself, and standing side by side by us, that really, hits home and thank you, thanks for, and if you had any closing thoughts, but, we appreciate it.
Patrick: Yeah, look, likewise, I mean, I really appreciate hearing from this group all the time. It helps me get better and be smarter, and, you know, Perry, Daniel, S, you know, everyone else On this call, you guys know where to find me. I al-always, you know, as I hope everyone knows, always open to talking to folks, whether they're big institutional investors or retail, and yeah, we'll just continue fighting the good fight. Really appreciate all the support and, yeah, don't hesitate to reach out if you think I'm doing anything wrong. I would love, love, love to hear the feedback.
₿itcoin ₿utcher: No, we're, I don't have your contact info, but I'll, I'll get it from someone if necessary and, and your words call you, bro, if necessary. But I, I don't think, I think you guys are, I think you're doing what you gotta do right now, and the last thing I wanna do is bother you. I'd rather have you sign that East Kentucky site, so, by all means.
Patrick: Yeah, well, we're working on that. But again, I, I do, keep doing what you guys are doing, Daniel Good stuff to make it easier for folks to understand the business, but yeah, really appreciate everybody's support in this community. So thank, thank you for hosting.
Daniel: Yeah, and thank you for coming. as, as someone who I was around a lot of retail investors, I think what you guys are doing as industry leaders in colocation is phenomenal, and I really see colocation, becoming more popular and bifurcating from neoclouds as a whole, and that alone should be giving the likes of you guys and Cipher and whoever else is industry leading in colocation your own premiums, 'cause you guys have a really durable and awesome business model, and you're executing on Such a high level, so we really appreciate all your hard work.
Patrick: Thanks, guys. Have a great night. Thanks for having me.
₿itcoin ₿utcher: Thanks again, Patrick. Thanks everyone for joining. I will follow up with, post, linking the first and the second parts together. And thank you to my co-host, Daniel, and all the speakers. Have a great night, everyone.