Unchained - Bitcoin Miners Are Pivoting to AI. How Does It Impact Crypto? Bits + Bips - Ep. 943

Episode Date: November 8, 2025

Subscribe to the Bits + Bips newsletter: https://unchainedcrypto.com/newsletters/ Check out our sponsor Mantle!  As profitability tightens and competition soars, many Bitcoin mining companies are turning to artificial intelligence and high-performance computing (HPC) to stay relevant. In this week’s Bits + Bips, host Steve Ehrlich sits down with John Todaro, Managing Director, Crypto & HPC/AI Equity Research at Needham & Company, and Kevin Dede, Senior Research Analyst at H.C. Wainwright, to unpack the pivot that’s reshaping an entire corner of the crypto industry. They discuss how miners are courting AI clients, why Wall Street is suddenly valuing them like data infrastructure plays, and what this means for Bitcoin’s long-term security model. The conversation dives deep into hashprice trends, investor signals, power constraints, and whether these companies can truly deliver on the AI promise — or risk stretching too thin. Guests: Kevin Dede, Senior Research AnalystManaging Director of Equity Research at H.C. Wainwright John Todaro, Managing Director, Crypto & HPC/AI Equity Research at Needham & CompanySenior Research Analyst at Needham & Company Timestamps: 💡 0:00 Introduction 🏗️ 3:23 Why investors suddenly care about miners’ HPC capacity 📈 9:08 Why the Bitcoin Mining Index is outperforming BTC itself 🤖 12:49 Can AI demand really live up to the hype? ⚠️ 16:31 The red flags investors should be watching 💰 20:50 Why debt levels could make or break mining firms 🔄 23:14 Can miners truly pivot and deliver on the AI promise? 📊 29:42 Why hashprice is falling even as hashrate rises 🚀 34:02 The long-term potential for Bitcoin mining operations 🏦 42:07 Bitcoin miners vs. holding BTC on balance sheets 🇺🇸 51:10 The future of Bitcoin mining in the United States 🔁 57:08 Could miners pivot to securing other assets? 🧠 59:22 Should the U.S. government buy a stake in Bitcoin miners? Learn more about your ad choices. Visit megaphone.fm/adchoices

I think there's probably always going to be a place for Bitcoin mining. I don't think we have a clue about what artificial intelligence can bring to us. If you look out the next 30 years, life's going to be not incrementally different, but diametrically different. Hey, everyone. I am Steve Ehrlich, the executive editor here and also the principal writer behind our Bits and Tips newsletter. And really excited to talk with all of you today about the intersection of Bitcoin mining and artificial. artificial intelligence, high performance computing. Quick pause. If you're deep in crypto, but TradFi and Macro is a different language,

or you're in TradFi and Crypto feels like chaos, we get it. That's why we created Bits and Bips, a podcast that bridges those two worlds. No jargon, no gaykeeping. Just smart, clear breakdowns of how these systems actually connect. So you stop feeling lost when one side starts speaking their language. If this sounds like something you need, check out bits and bips. You'll find the link right in the description. Just scroll down.

Mantle is pioneering blockchain for banking, a revolutionary new category at the intersection of Tradfi and Web3. Follow Mantle underscore official to learn more. I have two fantastic guests. Anyone that has been following the space for any period of time will be familiar with their names. John Todaro, a managing director, Anita McCos, with John, welcome. Thank you.

Thanks for having me. And then we also have Kevin Dede. He is the senior research analyst at H.C. Wainwright. And Kevin, actually, I think you are the first person, first equities analyst to ever cover Bitcoin miners. I think I remember you reached out to me once when I credited someone else with that distinction. And you very politely reminded me that you had been doing it for at least a year or two. Yeah, I was pretty much ahead of everybody by years.

Yeah. But that doesn't mean I'm the best. I would defer to John on that. Experience doesn't necessarily mean you're the paramount of research. But thank you again. Great to see you, Stephen. Great to participate. Yeah, absolutely.

So, yeah, I mean, a lot to discuss here. I mean, I'll just kind of set the table briefly. And then I'll let you guys sort of take it away from there. But for years, I mean, Bitcoin miners, they were always seen as this like high beta play. on on bitcoin and going back to like around the beginning of the pandemic and so on and so forth when bitcoin did well they did especially well when bitcoin did poorly um miners struggled even more so that wasn't a cardinal rule sometimes things change but generally those were the trends even when uh michael sauer and strategy started getting into the game they didn't have much of an

effect because strategy was really just a one-off that we didn't have that digital asset trend that we do now. But fast forward a couple years later, and we're in a world of ETFs, and now we're in a world of a cottage industry of DATs, and all of a sudden there's a lot of different places that are fighting for the same types of investor dollars. And the economics of mining, I've spoken with both of you about it separately in interviews for stories. They're challenging, and they're becoming ever more so as Hachshare goes up and the price continues to languish. But miners now, they've kind of found the golden goose. There's an insatiable demand for storage capacity, computing capacity,

and turnkey may not be the right word, but a lot of the infrastructure that these miners are using to mint out new bitcoins can be repurposed for the likes of OpenAI and Google, Microsoft, etc. So let's just kind of start with there. John, we spoke a couple weeks ago for an article I wrote on this very subject, And one of the things that really stood out to me was sort of your statement that the demand, HPC demand has completely re-rated how your industry looks at Bitcoin miners. So maybe just first start by kind of walking us through the conversations that you've had

with investors, some of your clients, and like what was sort of like the genesis of that shut. Yeah, sure. So I would say almost all the investors were speaking with, It's around HPC AI workload opportunity around the miners. There's some questions we feel around Bitcoin mining, and there's some understanding you still need to have around the Bitcoin mining business. But for the most part, all the conversations we're having is around HBC and AI.

I think what really elevated it was you had your CoreWeb IPO. That kind of got things going a little bit. You had a few more leases signed in this space with NeoCloud, with CoreWeave. And then it really took a second gear with Oracle's RPO numbers that came out. And that was just a massive amount of capacity they were going to need. We were backing into, they have some capacity with Crusoe and some of these other ones out there. But in total, 4.5 gigawatts, what they would need to still go out and procure would be like 2.6 gigawatts. So you could very quickly back into just a lot of demand that is out there just to satisfy these customer obligations.

And so that kind of really re-rated the names additionally. Great. I didn't mean to interrupt you, but just I want to make sure that our audience is following everything that you say. There's a few terms that you threw out that you probably used multiple times a day. NeoCloud, gigawatt, and I forget what the third one was. RPO. RPO. Thank you. RPO. Yeah. HPC, maybe. HBC, I'm hoping we have a pretty erudite audience. I'm hoping they have that one down.

But those three. Could you just please briefly explain what they are? And gigawatt, I get is a function of, I guess, a measurement of, form of measure of power energy, but yeah, but just like put that in context so that people understand what one gigawatt means when it comes to like capacity. So it's a massive amount of power. I actually, maybe some of the XAI stuff that's been going on, I don't know if anyone has a full gigawatt data center up yet, right?

But they usually, they're in multiple kind of buildings where you're looking at 150, 250, 150 megawatts. But it's a significant amount of power. I think the numbers most folks are working off is the U.S. Data Center market today. like 25 gigawatts or so. So when you're talking in the magnitude of one gigawatt, two gigawatt, three gigawatts, it's a significant amount of power.

And we're obviously in a power-constrained market right now, environment. So there isn't a lot of this available power. And that's been the advantage of some of these Bitcoin miners have had, is they have these pipelines of power. Neo-Clouds or GPU as a service companies, so CoreWeave, Lambda Labs, Nebias, those are the names to think of. We were on the CoreWeave IPO, and I think they were the first, yeah, they're kind of the first and really only,

I guess Iron you could maybe call a neocloud now, one of the first and only neoclouds. And so that's GPU as a service where they procure the GPUs, and then they typically lease that capacity, including their software stack to end customers. Cori's big customers right now, OpenAI, Microsoft is that to deal with Google. So you have some of those major hyperscalers, the Googles of the world that you're signing with. But then they also work a lot with AI startup lab. So these smaller AI companies that you've heard about that have their own language models, they typically are using the Neo Clouds as well, like a core week.

Got it. I think RPO. I think that was the. Oh, the remaining performance obligations, it's basically looking at a customer backlog. So if Oracle and some of these companies come out and say, hey, we got, you know, some of them are doing a two-year RPO numbers now, but you're saying, hey, there's a two-year customer backlog. And so you added up that data center capacity to support and actually to generate that revenue. right and actually take that in okay all right great thank you for that and uh i've always learned from a long

period a long time ago it's never don't be afraid to have stupid questions because someone else has they're not stupid so um every dynamic industry develops its own esoteric lexicon so there there are no stupid questions okay yeah that's great um so so kev i want to just kind of ask you the same question to kind of level set here um like what are the types of conversations that you're having with investors. Is there anything John said that you disagree with? Oh, no, no, not at all. John and I are completely on the same page.

I went to the first core scientific analyst meeting, and there were maybe five guys, five research analysts from the cell side. And then the one they held this past June 24, all of a sudden, the room was full of 120 people because they had just signed a deal with CoreWeave. And one conversation I had with a by-side analyst there was, you know, we're just hoping, I want to mention the company, just bulldozes all their facilities and starts from scratch in designing tier three, tier four data centers to address the need that John's so well highlighted.

So I think it probably makes sense to put some numbers, some visuals to what we're talking about here. Put together a few charts. I know the production team is going to help us, hopefully put up the first one. Sort of just comparing a thank you, real. So pretty much comparing just over the last three months, an index, I think it's from coin shares, looking at a blended index of various Bitcoin miners and how it's comparing to the price of Bitcoin. Very clear transfer. Bitcoin is down, I think, 12% in change over the last three months or so,

and this mining index is up 120%. Clearly, there's an interest. story to tell here. And it really, I guess, comes down to AI Helium, AI excitement because of that constrained capacity that John you were talking about. But maybe you could just explain to me in our audience what this chart means to you. And obviously, this is a good place to also highlight how not every minor has the same AI type flexibility or is going at AI and HBC the same way. Yeah, I would say, you know, this definitely to me looks like the HPC outperformance aspect to it. You've seen some of the, you know, Dats out there that haven't performed great either, right, down even more than the price of Bitcoin.

So the Bitcoin exposure, obviously, I don't think it explains that by any mean that it's all got to be HBC outperformance. In terms of kind of the minor, you have a host of them that are going down it. I would say, you know, Kevin, keep me honest, I think all of them, are now, you know, clean spark is even come around to it. And they used to be very, very negative on HPC. They come around to it. Marathons leaning in a bit into it. So I think almost all are definitely talking about it. Obviously, the ones that have signed leases right now, there's four, maybe five, I believe that have signed leases with customers. And then you have maybe another five or six that are very seriously going down this with some looking to break down, poor concrete, some of the

the near-term stuff you can do before actually having a lease signed. And then even your other names that typically were just focused on Bitcoin mining, they're even at least talking about it and looking to see what capacity could they put on, what they have to acquire more capacity to go for HBC. But they're also talking about it. Yeah. Cab, what do you think? Yeah, there's no argument.

It's pretty cut and dry. I mean, as far as I see it, I think the only Bitcoin miners that really haven't leaned in are the ones of minimal scale. And that's because they're capital constrained and it's a little bit more difficult for them to get to the financial markets. But there's a whole slew of those companies and they're being left behind. It's also interesting that some companies have decided to split their mining operations and their HPC operations. And I think you're probably going to see more of that. I mean, Hives are a great case, right?

They created their company buzz, Bit Digital spun off white fiber. Core Scientific said that they're pretty much transitioning every. available watt into HPC, where in prior calls they had actually talked about maybe once the, if the core weave deal acquisition actually transpired that they would act, they would sell their Bitcoin mining operation. I think at this point, that thinking has changed a little bit. But now, John's right on the money. I mean, there's, there's really no dispute about what's happening here.

It's the, I think the bigger question, honestly, is how long does it last? Is it a bubble? What are the implications if the market does become satiated? And how does the end market actually pay for the huge amount of investment? I think those are sort of the outstanding questions, not really so much what the transition is because I think that's pretty much cut in stone at this juncture. It's the end game, I think, that really raises the bigger picture questions.

I want to talk a little bit more. Let John chime in on that. I mean, I think that's like the most controversial aspect of this whole period. Yeah, I would definitely agree with that. It comes down to how long it can last, how those contracts are going to look. if there's execution risk on delivering it, because you could start getting into some penalties with the tenant that you've signed, the ultimate CAPX, what that's going to look like. And then, yeah, the end use case for AI workloads. If we, you know, don't live up to the vision of all the AI demand that we're expecting, then we just don't need all the supply.

You don't need gigawatts upon gigawatts of data center capacity specifically dedicated to it. But right now, it's looking like you do. and like those numbers I referenced out of Oracle, these guys do have customers. It seems like a very real backlog that they need to get data center capacities to support that. But the big question, yeah, is, you know,

how much, ultimately how many gigawatts do we need to support those end use cases? And I don't, I just, I don't really know. I don't, maybe Kevin knows exactly, you know, kind of when that shakes out, but it's, it's, I think that's the hardest part of determining kind of what these end use cases ultimately look like, what that end demand is

and what time frame that occurs over it? I'll be straight up candid. I am a lot older than John is on these things, and I've been beaten down through the backside of so many cycles. It hurts to think about it. We saw a huge buildup in the late 90s, right, with Greenspan having the tap open, a ton of money flowing into the market,

and you could walk down to Sand Hill Road with a piece of PowerPoint, and walk out with a hundred million bucks to build something that, you know, it was allegedly going to go on the web. This cycle is a little different. There aren't as many potential players, I think, but the end result is probably more dynamic. I mean, you know, it's been 30 years, right, since I think I saw the first WWW entry at Silicon Graphics in San Jose.

and 30 years later, who would have thought, or at that point, who would have thought 30 years later, you'd be able to hook up a ride from anywhere to anywhere, have any meal delivered to your doorstep. It's really kind of amazing how much of an effect our daily lives that web generation has had. I, you know, in answering that question that we sort of talked about a little bit, I don't think we have a clue about what artificial intelligence can

bring to us. If you look out the next 30 years, life's going to be not incrementally different, but diametrically different. Interesting. So there's a lot to break down there. I want to kind of explore everything you guys said and a few more questions. First, let's just call this the Long Island Ic question, or the Long Island blockchain question. I know when some of the companies, it must have been like a year or two ago, started putting out press releases saying they're pivoting to AI and saw a nice stock bump without anything to back it up. Like today, obviously, there is substance behind the hype. It's just a matter of perhaps how much of it is hype versus substance.

But I'm just curious from you guys. Are there any companies that you think are really over their skis in terms of what they're promising to try to get the PR bump that investors should be aware of? And I know that you might be a little bit restricted in what you can say, specifically when it comes to companies. So maybe a way to get out of that or get or a different way to answer the question is, what are one or two signals that people watching this should be careful about if they're thinking of getting into some of these, these companies to ride the AI wave?

What should they watch for? I would say not all megawatts are created equal and really due diligence to the pipeline because what could be considered total power or power under, you know, some, characterize it as under due diligence, that power, they might not ultimately procure, right? Or they might not have the supply chain built out to get the equipment to procure that in any kind of reasonable time frame. So I would be kind of, you know, curious to dig in, or if I were a retail investor, I would look into, to make sure to look into the actual power pipeline that's there and make

sure you go through that thoroughly. what I think also is actually the more risky business model is the GPU as a service business over the co-location business, which is what a lot of the miners do. That is where I think you can actually lay around a lot of debt and you have the GPU obsolescence. That starts to become a concern. And there's shorter duration contracts. If you're signing an AI lab, maybe it's two years. If you are able to get Microsoft, OpenAI, some of those more quality names, it's five years. When you're looking at the Bitcoin miners through a co-location agreement, it might be 15 years, right?

And so there's just a longer duration with really good margins, too, for a lot of these. And if you sign an Amazon, it's pretty quality. So I, you know, I think for the guys doing that, assuming they can hit their execution timelines and don't have some of those penalties we talked about, then, you know, I think that's a fairly attractive agreement, right? And there's maybe not a ton of risk associated with it. Obviously, there's still that execution risk we could talk about. But I think that GPU as a service business inherently carries a lot.

more risk. And also it's a lot of CAP-X to layer on too. And so you're talking about a lot more debt. Both of them have a lot of debt, I will say. All right, Kevin, what do you think? John Spot on. I think one of the other variables I'd encourage investors to include in their calculus is what's happened to the stock already? How much is baked in? We saw iron take off, right? It went stratospheric. And everybody's like, well, that looks like a big short. But they came back with, You know, they came back with a fluid stack Google arrangement. So if I have that right, or was it Microsoft? Anyway.

Microsoft for Iron. Yeah, Microsoft for Iron. So that kind of solidified their position, right? And, you know, it kind of makes sense to think that they've got that sweetweather site, which is, I think, another two gigawatts coming online. And there's a lot of potential for that. but John's absolutely right. It's going to cost a lot of money to develop.

It's, you know, not necessarily easy. I think that's another. Yeah, I'm sorry. Go ahead. Sorry. Did I hear this right? It costs about $100 billion to build out one gigawatt of power, or am I making that number up? No, I think that's about right.

10 million per megawatt. Yeah. Okay. Yeah. So that's because I was listening to something about Sam Altman looking for a trillion to build out like 10 gigawatts or whatever. No, that's another data point to John's thesis. Spot on, right?

Huge demand, not just Oracle, but Open AI. And he's been on CNBC saying, hey, listen, you know, we'd energize every watt we could find. I think another thing that you'd want to look at as you start picking apart this universe is debt levels. and John touched on this, right? I think the nice thing about that powered shell model that John talked to is that it's very transparent, right? The GPU as a service, you don't really know what they're going to be able to lock down and what they're not going to be able to lock down. And it's only going to become more competitive, which you would think have negative effect on pricing leverage. So I think that's an aspect, too, that is.

is a ding in the GPU business model versus the power shell. The opportunity in neoclouds, though, I think, is a lot different. Again, John talked to this. A company can bring real value at it. It's one thing, like John and I could pool our money, you know, our $200 or whatever we could scrape together between the two of us and maybe buy part of a GPU, right? We're not going to be differentiated. We're not going to be any different than Irons or Amazon's GPU service. Where the real differentiation comes is in the software stack and that the ability for that provider to architect a system and solution for any particular client.

And I think that's going to become a big differentiator. You can look at companies like CoreWeave and Nebius as really differentiating. from some of the others that are just trying to throw a lot of money at this space. It's not just about the GPUs. It's not just about the megawatts. It's about that company's core competency. And I think that's probably a longer term, a much bigger differentiator. Right now it's the Oklahoma Land Rush, and anybody who can scrap together 100 megawatts can make some money, sign up a deal.

But what happens longer term, I think it depends on the value add those companies can tribute. I have one more quick question for you, John. You did mention something I was going to bring up. Sort of some of the contractual requirements for uptime performance, et cetera. I mean, Kevin, I see reports that you and some of your colleagues point out. I mean, John, you do the same. And pretty much every equities analyst that covers Bitcoin miners, we look at things like uptime, like actualized hash rate capacity, number of Bitcoin mine, et cetera, et cetera. And I mean, there's no penalty for, I mean, like, I know Marathon, I think last year, matter they got wiped out because of storms and lost like a huge chunk of their capacity for a while.

I mean, there's no penalty because they're not contract. I guess maybe if they are contracting out to third parties, maybe there is some sort of penalty. But I mean, for here, I mean, this is very high-tech stuff. I'm sure it's very complicated to run these types of, this type of infrastructure, just like it's tough to keep Bitcoin miners up at all hours of the night because they run hot and et cetera. I know HBC is kind of a new industry in and of itself, but how well will these companies be able to operate all this capacity so that they make the $10 billion their problems to make in one of these 10 and 15-year contracts? Are they expert enough to do this or are there going to be problems?

So a few things. One, we don't have a ton of data points yet. So it's yet to be kind of seen. You've got a couple of these sites coming online now, core scientifics, one with CoreWeave, applied digital site as well in the Q4 here. So you're getting just a few data points. But what I would also point out is, you know, having the power. So it's a powered shell co-location model where a lot of that GPU and interior infrastructure

is coming from the tenant. So that's coming from a core weave or an Oracle in those data halls. So they're responsible for a lot of that maintenance, kind of networking, any of the cabling issues that could result in downtime and reliability. For these guys, it's a little bit easier to have the data center going. I think what's the bigger risk is executing on the timeline and actually getting the construction done, getting the data center off and running. That could be difficult because there's long lead time items you need to procure.

Some of them have that, right? The electrical infrastructure, transformers switch gears, we always hear about. But others, you still need to go to procure, right? And so that could take some additional time. And if you have too aggressive with timeline, the contract. And this is where I think it depends on the customer, or sorry, we'll both, the customer and the minor that signed them. And we don't have a ton of granularity into this yet. But our understanding is a couple of them. If you miss two quarters, they become cancelable. And there also might be

penalties baked in where if you're missing a month or so on getting the site online, it can start to chew into the rental economics that you can get on it. So we do think the bigger risk is executing and getting the site at the timeline versus maintaining operations. But that, you know, it's a risk that could arise. It's just one that is kind of also just less due diligence at this point of the cycle, but it could be something. These are complicated, large data centers. But I would stress the point at the interior data halls, any of that uptime and reliability is typically run by the tenant. And so that kind of falls on that. And one just quick follow up before I go to the ads. I wrote a story a few months ago about how some of the Bitcoin miners might be affected by Trump's

tariffs when it comes to getting miners from Asia and how some of the miners were trying to strategically place themselves throughout Southeast Asia. So at least they didn't have to come directly from China or even be trans-shipped from China. Are there any tariffs issues when it comes to building all of this stuff, getting the chips that are really, that's really what's paying attention to? Yeah. So like I mentioned, so on some of the chips like the GPUs, these guys aren't necessarily procuring it. So that doesn't really fall in them if they're in the co-location model. But there's other pieces to the data center infrastructure that is going to be subject to

terrorists, right? And so, you know, it's impacting a little bit. I mean, the terrorists have been all over the place. I remember at one point we were kind of factoring like 10, 15% additional expenses. When it comes to the CAPX range, it hasn't changed a whole lot over the last six months. It's kind of that $10 million per megawatt number. Kevin was referencing. So, you know, there is some risk that maybe that jump in.

so like 12 million per megawatt as the terrorists evolved. But we've been fairly seen that pretty steady. It seems over the last five months or so that we should be. Did you see the iron numbers that came out, John? I think he pointed to 14 to 16 or something. Yeah. So I see it. I kind of see it creeping up.

And I don't know if it's tariffs or just demand outstripping supply. But yeah, John's absolutely right. I mean, there are generators that might be sourced from China. They're transformers that might be sourced from China. There's switchgear that might be sourced from China. So tariffs are certainly an issue. Okay. All right.

Well, now let's take this long anticipated ad break. I'll give you guys a second to get a glass of water. And then we're going to come back and talk a little bit more about the economics of Bitcoin mining. Quick pause. If you're deep in crypto, but Tradfi and Macro is a different language, or you're in Tradfi and Cryptofer. feels like chaos, we get it. That's why we created Bits and Bips, a podcast that bridges those two worlds. No jargon, no gaykeeping. Just smart, clear breakdowns of how these systems

actually connect. So you stop feeling lost when one side starts speaking their language. If this sounds like something you need, check out Bits and Bips. You'll find the link right in the description. Just scroll down. Mantle leads the establishment of blockchain for banking as the next frontier. UR is the access layer that transforms Mantle Network into a purpose-built vertical platform, the blockchain for banking that enables financial services on-chain. UR unifies and vertically aligns Mantel's focus on payments, trading, and assets, MI4, METH Protocol, Functions of BTC, supported by developer grants, ecosystem incentives, and the industry-leading distribution platform through the UR app, Reward Station, and By-Bit launch pool.

All economic activity, which, within UR will be captured by Mantle Network to further drive value to token holders and establish its significance in blockchain for banking. Follow Mantle underscore official to learn more. All right. So let's put up a couple of charts looking more precisely at what is happening in the world of Bitcoin mining. They really are sort of like inverses of each other and can certainly paint a very bleak picture, which is what I was alluding to at the very beginning. First, this is the hash price index. This is source from the block, but the data actually comes from Luxor, which is, I guess, I mean, they do a bunch of things in the marketplace.

And what you can really pretty much see is a trend that's just going down and down and down and down and for the most part, hash price is near at or around all-time lows, not to get into all the complications for people listening or watching that haven't seen. it, but basically this is just a metric of mining profitability that takes into account, like the price of Bitcoin, in general, like the costs associated with mining, it overhead, electricity, et cetera. And then also, it's sorry to correct you on this, Stephen, but it's a combination of the network hash rate, the overall network hash rate, and the price of Bitcoin. That's it. Thank you. Yeah. And then the second chart,

this is the hash rate. And this, aside from some variations, it just keeps going up. and up and up and up and up, and up, and up, which is really fascinating because in some ways, it kind of creates, it shows like the cash 20 to the Bitcoin miners are in. If they're committed to the space, profitability is going down and down and down and down, hash rate keeps going up. I want to ask you, I mean, it seems like, I want to ask you the reasons for the hash rate going up. Maybe it's obvious, more powerful machines have higher hash rate,

and companies that are sticking with this have to maintain parity in order to get their same amount of Bitcoin. But, like, I think these two charts really show why Dickpoint miners, at least some, have been so excited to perhaps have another opportunity to monetize their infrastructure. So, Kevin, why don't I come to you first? And what does this mean to you? What did these two charts mean to you? Well, we've commented on this for years. If you look at each low going back to the 2018, 2018, 2019 time period, each low is lower. Each bare market brings you a lower hash price.

And that's because machines are more efficient. So if you use the same amount of power and get a lot more tarahash, miners are also becoming more flexible in the way that they operate. They're not going to turn their machines on when power is expensive. So there's still a ton of opportunity in that flexibility. But you're absolutely right. There's no arguing it. It's a, it's a dismal curve and it looks difficult. But if you believe in Moore's law and more powerful processing coming to the market

and you have favorable power rates, then there's still. You know, there's still an opportunity to make money, Bitcoin mining. How much, I mean, John, before I go to you, I mean, in general, Kevin, I know the answer is it depends. But how much does it cost typically all in for one of the larger scale miners like Clean Spark, Marathon, Riot, et cetera, to mine a Bitcoin? Oh, okay. So there's a big, a big gap between the players there. I think some have their numbers down as low as 50 and some are higher close to 70 or or 80. You know, I defer to John on his opinion on that as well. It's, it's, I think the best answer is the way that you framed it. It sort of depends. Yeah, I just say it definitely depends. Obviously,

cost of power is a big, big factor in that. And they can vary from, you know, five cents per kilowatt hour down to cipher's got one fixed contract still at. I think it's Odessa for 2.7 cents a kilowatt hour. So that's a big delta there. And then also if you're counting depreciation or just cash Bitcoin mining costs, I think cash costs. I'm kind of around where Kevin is. If you throw in the ASIC depreciation, I think you would get a little bit higher than an 80K even.

Right. Okay. So, I mean, with that, I mean, when you, because you speak, you guys collectively, I speak to all the major Bitcoin miners. I mean, what are they saying to you when they, I mean, they're well aware of these numbers. well aware of how expensive it is, and you can be nine, ten figure outlays to get the latest generation of machines. And during COVID, I remember they used to have to pay like nine or ten, nine months or a year ahead of delivery because there was so much demand for these, these machines.

I mean, what, what are they saying to you for the ones that are still sticking with, with the Bitcoin mining? Like, how are they handling it from a financial point of view? What's the upside do they see in the market right now? Because it is somewhat languishing. And I know that when it comes sometimes to the overall calculus, it's not just the profitability of mining Bitcoin right now, but hopefully putting it onto their balance sheet and then reaping the gains as it appreciates years into the future. Like how do they weigh all of those things? It's a great question. I don't know that there's one single answer to it. You'll get different, I guess, different perspectives from each different management team.

There are many that are wedded to it. I got the sense. I don't know what you think, John. But I got the sense that Fred Thiel has a moral obligation to continue to invest in Bitcoin mining because he's worried about centralization within the Bitcoin network. He's concerned about that. He wants to continue to invest. I don't think Clean Spark would walk away from it. And I think John would agree. There are multiple sites that just might not suit HPC, right?

There's stranded power. Like, all you need for a Bitcoin mine is electricity and, you know, a Starlink terminal from Elon. And you're good to go. And that can happen anywhere in the world. So I don't really see it going away. It's just a matter, you know, and John pointed this out, it's a matter of how you're able to manage your power costs. If you work really closely with a power generator, you can suck up their load when they don't have a big load and a big, alternative load, but when their, you know, when their made load comes up, you can wind down,

and they'll pay you extra for that because you're helping them run their generation facilities more consistently. And so there are lots different ways it works out. And John highlighted Cipher. Seifer has, they have mining facilities behind the meter at wind farms. They can get their power cost below two cents per kilowatt hour. There's lots of opportunities. Now it's my turn to Kevin. I ask you to define a few terms. One, behind the meter, I'm pretty sure I know what it means, but there may be people listening that don't.

And then two, I just wanted to point out anyone who doesn't know, Fred Thiel is the CEO of Mara, which is the largest publicly traded Bitcoin miner by hash rate, correct? Correct. Yeah. All right. So behind the middle. What does that mean? So, yeah.

So you and I, we, we, we plug our computers into the outlet, right? That power is fed. But I think you're in New York. That's all John definitely is. It's con ed. I'm in New York today, too. That's all con ed. Right?

We're paying that. Oh, you're in Philly. I don't know. You've got to be part of the PJM grid, right? Then anyway, it doesn't matter. I don't know. That is Pico. All I know is we lose power whenever the wind goes over 10 miles an hour.

You lose power? You should have, if you were in West Texas, you'd have excess power. I get it. So, yeah, so you're paying what the power delivery company is making you pay. Because you're paying that metered rate, whatever it is. Most cases for consumers, it's set up by some regulatory agency, depending on where you are. But that doesn't mean you can't go closer to the generation. Right.

So you can set up a Bitcoin mine or an HPC facility for that matter, right where the power is being generated. Power goes directly to your facility. It doesn't go on the grid. It's not regulated. It's not monitored. it goes from creation to use. That's behind the meter. Sort of like the power generation

or power consumption version of Flash Boys, the Michael Lewis book. Cluster you can co-locate to the trading venues, the microseconding. Yeah, yeah, yeah, yeah. So that's a timing thing. This is a generation thing. How did I do, John?

Did I do okay? I think you did a great job. Yeah, it summed it up well. And I'm in Miami. that's the only thing, you know, not in New York. Oh, I'm sorry, you're in Miami? Oh, great. Enjoy. You know, you're dressed up.

I mean, it looks like you have a vest on. I does not like. I got the AC going too cold, I guess. That's right. I actually asked him the same question a couple of weeks ago when I spoke to him. You look cold. And then he's like, I'm in Miami. And I got.

I'm trying to get that power bill up so I could be like a Bitcoin minor. My power of us. But yeah, anyway, John, same question to you. I mean, when you talk to Bitcoin miners, that ones that, I mean, I think when we spoke before you pretty much said there's almost no miner that's just exclusively focused on mining, at least except for like the really small guys, you probably don't cover. But when it comes to them figuring out the long-term potential for their mining operations, like how are they weighing all the different things that we spoke about a few minutes ago?

Yeah, I think Kevin hit it. Obviously, you know, power cause, doing voluntary curtailment, or if you have forced curtailment, it's not that set up for HPC, although you could kind of balance the load. so you could have some sites that do HPC and some Bitcoin mining. But either way, if there's forced curtailment, maybe you would continue to do Bitcoin mining, right? And so you can get some credits back from the power provider or power down and sell that power, right, and generate a profit.

So from that standpoint, it kind of comes out of their power contracts. What I would say, though, if we just take a step back and assume the AI demand, it does come to the extent we think. it's a little bit of a developed country aspect, right? The AI, the infrastructure to support these AI workloads, it's mostly it's parts of Asia, it's the United States, it's Canada, it's parts of Europe, right? If you're in these very remote, like if you're just in a desert in the Middle East,

for instance, there might not be the fiber infrastructure set up, there might not be the water access set up to do HPC and to do those workloads. But to Kevin's point, you could have a Bitcoin mining site out there. You get Starlink, you send that up, and you can get, you know, if you have access to cheap power, then you can, then you can mine Bitcoin, right? So I do think if you play that out to the extreme, the U.S. likely takes the available data center capacity, the megawatts, and pushes them more towards AI and cloud compute versus Bitcoin mining that can be done anywhere globally, unless you have either governments or the companies themselves, like, Teal is saying where you have almost like a moral obligation to my Bitcoin. Now, at a company level, if you're not profitable, it becomes very difficult. At a country level, you obviously could see that. I mean, Trump's talked about he wants a certain percentage of Bitcoin mined in the U.S.

annually. But I would say most of the infrastructure in the U.S. is just more suitable towards HBC and AI. And you can do Bitcoin mining anywhere. In terms of the public mining companies, once again, it varies. Right. So Clean Spark and Marathon are on one end where they're more. Bitcoin mining, although even Clean Sparks get into HPC some. And then on your other end, there's Tara Wolf, core scientific, applied digital, who want to get to a point where it's actually almost all HPC, and I don't think really any Bitcoin mining. Gotcha.

And I want to ask to, and we'll get the next chart put up here that we put together, in terms of like the Bitcoin miners, I mean, a big part of the upside is, again, them acquiring the Bitcoin and then being able to hold it. I know for the, for the most part, they do not want to part with it, except I guess under certain circumstances where they need to pay for operational expenses. But here's a chart that really shows sort of, again, this is the same Bitcoin miner index, and then down below are, I guess, the largest Bitcoin debts, again, showing performance over the last three months or so.

Really hard to ignore these trends. Again, all the, all the debts are underwater. their MNAVs are for the most part, either treading water above one, some have fallen below. I know certain companies like Nakamoto, Stride, I mean, their stocks are down like 99% from peaks. So just going back to, again, like how the miners think about acquiring Bitcoin now, even during a period where it's now, maybe 20%ish from its recent high, maybe a little more than 20%, so they can hold long term. But what does this chart say to you guys?

How does the long-term potential of holding Bitcoin impact the thinking when it comes down to strategizing how much to put into HPC and how much to put into Bitcoin mining? That's a loaded question, Stephen. There's a, yeah, there's kind of a lot to unpack on that. I think one is just looking at the dots themselves. there are so many ways to build an investor exposure, right? You've got the ETFs. You can buy Bitcoin outright or you can get involved with these stats.

But in my mind, the Dats offer a lever. And I think the ones that are performing well have made it clear to their investors that these levers are working. One is that they've got to have some sort of core operating company as strategy did, right, or does in software development, right, that can actually cover debt expense. They also have to have some financial wherewithal in order to raise capital, right? So if they have those two things, they actually provide an opportunity for investors that, you know, an ETF couldn't manage or, investor on their own couldn't manage, right? There's no way that they'd be able to offer these complicated debt instruments or preferred stock. The crowning jewel of a debt, in my view, is the ability to take some or extract some value out of the underlying chain itself. That, I think,

is critical. You look at at SharpLink Gaming versus BitMine immersion, right? Sharplink's got a stacked deck and senior management and a connection, you know, to the actual development of Ethereum. What Tom Lee brings a marquee name, and that's a nice aspect, right? You see that too with the Trumps and American Bitcoin, right? So that's kind of the dat picture. And I think that's independent of what's going on in the HPC side. So companies, you know, Mera might be a case, a perfect case of this.

Some companies, it's just investors, I think, John chime in any time, I think investors like to put companies in buckets to kind of try to figure out, you know, what the opportunity is, what are the risks that company faces because they're familiar. familiar with that bucket. There are some companies that have both this mining exposure, maybe some HPC exposure, and some debt exposure. How should investors look at that? It's not necessarily clear cut. I don't know if that really gets to the answer that you were looking at, but. Well, I was hoping for one very simple answer that everybody can clearly understand. You have to ask a simple question then if you want a simple answer. That was a very, very complicated question with lots of, you know, lots of opportunity for variables to play in different ways. So don't tell my boss that because I want her to think I ask good questions.

John, John, go ahead. Yeah, I think I'm mostly aligned there. I do just think HPC is viewed entirely differently. So, you know, I think the overlap would be more so in some of the names that still have Bitcoin on the balance sheet. So to an extent you could almost look at them as a debt, but they have such just like the power pipeline we talked about, the HPC opportunity is so big. They're that most of that valuation is kind of based on that.

Most of the investors are, to Kevin's point, looking at buckets where they're going, okay, is this guy an HPC bucket versus like kind of comping them to adapt? But, you know, like you can say, hey, you know, you know, toss Bitcoin to the wayside. because everyone wants to go down HPC but to an extent like these guys built up very good balance sheets

by Bitcoin mining and Bitcoin has gone up a lot over these years right so to some extent some of them are sitting on you know 1.5, $2 billion in Bitcoin on the balance sheet so you know it's easy to say okay HPC you should all go down HPC but they have very healthy balance sheets some of these from Bitcoin miming

and in the end having Bitcoin gone up so you know So I think, you know, there's still a lot of, you know, credit to give to the Bitcoin, and I still think there's opportunities in Bitcoin mining. But, yeah, I guess I just don't think there's too much of a direct overlap with that. In terms of investors, I almost get none of the guys who are looking at the Dats, who are looking at the miners at this point. Because one question I did have was, I mean, I know minors, I'm sorry, Dats.

I mean, their capital raising opportunities right now are limited. I mean, nobody wants to buy now. I mean, it's a possible fundraise at, or virtually impossible to fundraise at, and that's less than one. Converts are starting to come out of the money, or start to fall out of the money. I know that there's a lot of stuff happening behind the scenes in terms of ownership of that type of debt. I'm wondering if, one question I just had is if, like, there's anything interesting happening in terms of converts when it comes to Bitcoin miners, either to finance Bitcoin mining or to finance HBC stuff. I know, John, when we spoke, I mean, you pointed out that a lot of the companies that are renting the space are paying for the buildouts, financing the buildouts for the infrastructure themselves. So it's not like the companies have to go do that on their own.

And then sometimes they can then go to a bank and take out, take out loans using the contract as a form of collateral for lack of better term to get more favorable terms. But investors, because converts were such a popular thing for Dats, I'm wondering if there's any like spillover now onto the mind. because of this HPC opportunity. Oh, I mean, a lot of them are doing convert. So you're definitely seeing it. A couple of them recently closed convertible bond transactions. So you're definitely seeing converts from the miners. Typically, the structure was you do project level financing debt for like 80% of the

build. And then as the site got online, you could flip to ABS debt with the facility securing it. But there's a lot that are tapping the convertible bond market. I think a couple of this of high yield issues. We were on some of the recent converts, but some of the recent high yield ones we worked on. But John, before you go on, wouldn't you say most of that is in connection with their HBC builds

versus their Bitcoin builds? Oh, 100%. Yeah. I would say it's all HBCC. Right. You've got the transparency. You've got a real operable business model, and you're removing that debt.

from the volatility inherent in Bitcoin. So the debt market exists in support of these companies purely on their HBC opportunity. Sorry to interrupt. Oh, yes. No, no, that's a fair point. And maybe Steve I miss heard you. Yeah, all that is in support of these HPC buildouts. It's very little that would go quickly my name.

But it could, I think maybe to Kevin's point, you could do some of that fundraise and then, you know, add some A6 here and there. But we're just not seeing it. There's just such a big debt build. or just a big financing component you need for these data center builds that I think we have earmarked for our models. Like any debt they raise basically goes toward building up these data centers. Gotcha. All right.

So let's wrap up by just kind of taking a bigger picture or looking a little, taking a wider vantage point and trying to understand what this means for for Bitcoin mining in the U.S. overall. When we spoke a couple weeks ago, John, and one of the things that you wanted to get across is that, This could really lead to a sea change in kind of how Bitcoin mining happens in the U.S. I mean, you alluded to it a little bit earlier. And I think Kevin, you did as well talking about how it doesn't take the same type of setup to do Bitcoin mining. All you need is a Starlink terminal and a little bit of power. And you can have a Bitcoin miner on the moon, I guess. But, like, what do you think, I mean, John, what do you think this means for the Bitcoin mining industry in the U.S.?

And I ask you this question with, like the background of, I mean, Donald Trump, when he was during the campaign last July, I mean, kind of tongue and cheek, he said he wanted all remaining Bitcoin to be mined in the U.S. Obviously, that was never going to happen no matter what, but it seems even less likely to happen now. His sons are involved in running a Bitcoin mining company. And even though the majority of hash rate is not based in the U.S., it's more distributed worldwide, the U.S. still has a pretty sizable, makes up a pretty sizable portion of global hash rate. So what are you hearing from D.C.? What does this mean for the Bitcoin mining industry in the U.S.? Yeah, maybe I'll defer on the policy stuff to Kevin if he has a good answer there. We're not speaking too closely with kind of policy folks.

What I would say, though, like there's still a lot of hash in the U.S. It's not to say, look, all these sites are doing HPC and there's no hash in the U.S. right now. American Bitcoin is a good example with what they're doing with hot. They've added a lot of hash, right? So a lot of these are still a lot of hash capacity coming on from the U.S. And so I would say near to medium term, I think Bitcoin mining in the U.S. is definitely still going to have a presence. And it's going to be fairly impactful. It's going to be kind of a large minority share of the total network hash.

My angle is just, if you play this out to the extreme where every megawatt goes to the most economic purpose in the U.S. And it has lost something changes on hash price, it would find its way into HBC, assuming it's suitable for that. Some of the stuff is just so remote in the U.S. that the costs are going to be so high that it probably does make sense to continue doing Bitcoin mining. Or to Kevin's earlier point with some of the curtailment out there, these grids might need Bitcoin mining actually to, to support them, right? Because you can bring down in storms in, in the winter and sometimes in the peak summer months as well. I think there's probably always going to be a place for Bitcoin mining. I was just making the case. The infrastructure in the U.S. is very much set up more for AI than Bitcoin

mining. And you are still, you know, I think you are going to see some U.S. companies take hash down, like Cyp. For instance, they're going to do HBC at their Black Pearl site. They're going to bring or reallocate and then maybe bring that down completely 10x a half. or so that's at that's at black pearl for bitcoin mining core scientific's bringing a lot of hash offline uh down the road right um wolf was probably going to be the same way so uh riot also of course the canada instead of going to bitcoin mining maybe it goes to hpc so you're either going to be flat to bringing that down but i would still say there's other companies that are focused on mining like american bitcoin's probably never going to do hbc it's not really suitable for them they don't own

the sites right it would make sense to do it um so sorry what was that i was just saying they'd have to change the name too. They would have to change the name. American HPC. Yeah. So, so look, I think there's always going to be here, maybe not always, but for the foreseeable future, there's going to be

Bitcoin mining in the U.S. But if you just play it out to the extreme, I think you would see more and more megawatts find their way into HPC hands versus mining. Kevin, what are you thinking? Anything to add? No, I'm with John 100%. I think I'd add one thing and people have tossed this

term about mullet miners around. That's sort of, you've got your HPC in the front and your Bitcoin mining in the back. I don't think the technology is there yet, but I think it's important to recognize that HPC loads aren't always consistent. And with the proper switching and sensors, you could make an argument that a 200 megawatt load, even though it's all signed up for HPC,

might also support 100 megawatts of Bitcoin mining, not all the time, but incrementally. That HPC customer is paying for the power 24-7-365, it would be a shame not to use it. And like I said, I don't think the technology is there yet. I agree, John, 100% with that on the fact that there will be some Bitcoin mining here and there in the States. And, you know, I stand behind Fred Thiel on that 100%.

But I also think that, you know, with some more engineering finesse that you can see the two cohabitate the same data center. And I think HUD-Aid is working to that, right? They've built racks in U-form, which is the standards for regular compute. And then they've contracted with Bitmain to have miners built to fit standard compute racks. So with liquid cooling, which is all sort of next generation HPC. And that's why I, you know, I'd argue that they can cohabitate. Actually, you know, I think Iron is running a site in BC that has both HPC and Bitcoin mining.

Two very quick things, and then we'll wrap up. I know this is probably not part of your direct coverage area, but just in light of, especially those, the hash price chart I was showing earlier. Do you ever see or ever foresee any of these miners pivoting to mining other types of assets? I know Marathon, I remember they were doing which one. It was called. Aspa. Thank you. They were doing that one.

I mean, Zcash is on the tear right now. I mean, light coins getting an ETF. Do you ever get a sense that any of these are going to kick the tires on any other P. P.O.W assets or do you think if it's going to be mining, it's going to be Bitcoin? I agree with your latter statement. but that doesn't mean people won't dabble. They cover a company called Bitmine, Bitmining rather, and they were mining a light coin and some other coin on a script miner.

And that kind of work, you go rewind a little bit, right? It wasn't long ago that Ethereum was mine. I think there will always be some aspect of proof of work, but clearly Bitcoin not only dominates proof of work, but it dominates the entire cryptosphere. Yeah, I was just that, you know, a lot of these management teams, they're kind of, some of them are like Bitcoin maximalists, right? You know, you know that term. So maybe less likely unless it's super economic.

It's actually an interesting question. One we haven't thought about in so long. It just seems like industrial scale mining has really been Bitcoin. But yeah, there wasn't that long ago where Ethereum was mined, right? And you still have, you know, Cor Weave basically came from that, right? Yeah. So, yeah, it's an interesting world. Just with Zcash on a tear, maybe it does make sense.

I was actually interested in pulling up the mining economics on Zcash, actually, and see if it's super profitable. But, I mean, I don't know. Yeah, for the last two weeks. You know, what's great, John, is you could turn down your AC. You could turn your AC down and turn up your Zcash mining. Yeah. Actually, I wouldn't be able to wear the best, you know.

I'll say one funny thing, though. I did a big profile on Zcash. I would say about a year ago, a little over a year when I was still at Forbes, and one of the people I spoke to there said that all the Zcachian, miners, when they would mine the Zcash, they would automatically sell it for Bitcoin. I don't think they're doing that anymore right now. But one other quick one other quick one, I just, I'm curious you guys' impressions or thoughts. In this new Trump administration, the government is starting to take some ownership roles

in certain species of companies, Intel. I think they're getting some, I guess, revenue, 10% revenue on Nvidia chips sold to China. They have a golden share with U.S. steel. do you ever foresee they're trying to promote the Bitcoin industry in the U.S. Do you ever foresee an opportunity where the government might say, hey, look, we want a piece of marathon,

we want a piece of riot or CleanSpark or something like that as a way of either adding to the reserve or just a way of doing business, given how that seems to be a bit more than modus operandi right now. Go ahead, John.

Sure. Well, not to speak for the administration. or anything. You know, I think maybe there's a case for it. I would make the case. You know, I think Bitcoin's a critical industry and, you know, and you need to keep it onshore. So, so maybe it does make sense.

And, you know, maybe American Bitcoin is kind of like the first kind of like foray into that, you know, with his son involved with it. So, you know, maybe it does make sense that, you know, the Trump administration just gets the federal government in total a little bit more on board with Bitcoin mining, being a critical. infrastructure. I could see it, I guess. But you know, like you look at David Sacks though, right? You know, there's obviously the crypto, but there's also the AI piece and there's just so much going on there. Interesting that the people point persons in charge of both. So it's. Yeah. So it's like, you know,

maybe you have both critical infrastructures and they're kind of a little bit competing right now. So it's interesting. Maybe you support both for your, your, your, but the problem is you just have, it's a constrained power market for now. So you need to really open up some of the power markets and get that going where there's just more generation so you can support both because you are just constrained regardless of whether the U.S. government starts taking stakes in some these companies. Yeah. I think if you look at it from, I guess, a solidarity perspective in terms of making sure that Bitcoin network stays cohesive and a federal Bitcoin reserve, it only makes sense that the government continues to invest in maintaining that value, right? If all the mining

was to emigrate to China or Russia or Kazakhstan, then there's less decentralization, which is a key aspect, right? You can't ever let one miner dominate more than 50% of the network. so or you risk it falling apart so that's i was going to that make a joke and ask you that defined that that sovereignty i think is is important sorry steven i was speaking over you no i was i was going to make a stupid joke and just say can you define decentralization please but uh but then i tried to stop that's actually a hard question that's a question all the uh layer ones are trying to figure uh yeah i guess if you really want to start splitting here as it can be okay um all right we're a couple It's over, so I appreciate the time, guys. Before we wrap up, is there anything either one of you want to add last minute thoughts before we let everyone go?

Just want to make sure that, you know, it was a real pleasure for me to share the screen with Johnny so much better looking. I mean, it makes me feel like that I can kind of fit in okay. So thank you very much for having me, Stephen. Yeah, thanks for having us, too. And yeah, it's always a pleasure to be up here with Kevin. Okay, great. And we're working people if they want to connect with you guys or hear your research. I mean, what's the best way to potentially get in touch? Email for me. Email for me. KD. First and last initial at HCWCO.com. Yeah, same thing. Email. Jay Tadaro, spelling of the last names up there. J.Tadaro at Needhamco.com. Okay. All right. Great. Well, thanks so much for sharing your, your,

time, your expertise. We'll have to do this more often as more big, like, multi-billion-dollar announcements come out. And fewer ads next time, Stephen, fewer ads. We get, we get very confused. That's how we pay the bills. We need, we need, we need ads. And we are very good. I'm joking. It's all right. I actually, thank you so much, John. Really fun to see you. Actually, yeah, have it likewise. Thank you. Okay. Yeah. Take care. All right. Thank you. Thank you. Bye.