Skip to content
PodSubs
ShowsPeopleGuestsJobsNewsEvents
More
BlogPodCred
Menu
ShowsPeopleGuestsJobsNewsEventsBlogPodCredNetworksSign in
Join nowSign in
  • Shows
  • People
  • Jobs
  • News
  • More
    • Guests
    • Events
    • Blog
    • PodCred
    • Networks
    Join nowSign in

What Bitcoin Did

What Bitcoin Did artwork

AI Is Stealing Bitcoin’s Miners. Bitcoin Wins | Harry Sudock

Nov 4, 20261 hr 3 minAudio
ShareXLinkedInFacebookWhatsAppEmail

“We have approximately 2% market penetration of AI tools. That was all it took to sell out the entire planet’s compute.” Harry Sudock, Chief Business Officer at CleanSpark, returns to discuss why Bitcoin miners are moving into AI and why that could make Bitcoin stronger. We get into the economics driving the shift, what it means for mining decentralisation, and why Harry believes demand for AI compute is only getting started. We also discuss the backlash against AI data centres, including concerns about electricity prices, water consumption and noise. Harry explains why he believes building more power could lower energy bills, how data centers can benefit local communities, and why he would start an energy company today. Sponsors: Ledn - Explore Bitcoin-backed loans and get 0.25% off your first loan. https://learn.ledn.io/wbd BitGo - Secure your Bitcoin with institutional grade security. https://www.bitgo.com/danny Bitkey - Get 10% off the new Bitkey wallet with code WBD. https://bitkey.world/wbd AnchorWatch - Insured Bitcoin custody, security & inheritance. Book a consultation: https://www.anchorwatch.com/wbd Cape - Get 33% off your first six months with code WBD. https://www.cape.co/wbd Abundant Mines - Bitcoin mining made simple. https://abundantmines.com FOLLOW: Danny Knowles: https://x.com/_DannyKnowles Harry Sudock: https://x.com/harry_sudock

Audio and video play straight from the show's own hosting. PodSubs stores and serves no media.

Transcript

From the show. Tap a time to play from there.

So we have approximately 2% market penetration of AI tools. That was all it took to sell out the entire planet's compute. If we expect to have a better quality of life for ourselves and for our children, we need to generate and consume more electrons. The answer to higher prices is typically higher prices, which means we need to build more generation. We need to deregulate the ability

to invest in our grid. We need to make growth the priority because if you can grow supply, you can meet the needs of demand and you can serve the marginal consumer more effectively. Let's go, Harry. Good to see you, man. Good to see you, man. Busy. Oh my God. we got a lot to get into today um i've been speaking to a lot of people recently in the bitcoin world who are like a little bit worried about what's happening in mining or maybe more

precisely what's not happening in mining how everyone's moving over to ai worried about the hash rate dropping off and kind of security budget all these future scenarios and i want to run through all of this with you clear everything up find out what's happening in the mining space what's happening in ai you ready yeah okay so one of the things i want to start with is with all the big miners moving to ai at least to a degree who out there right now are the ones that are making money like who is the in the best position on a pure bitcoin side so i think there's really

there's really two competitive structural advantages that miners can accrue the first is the cost of acquisition for their land and infrastructure so cost basis right it's better to own bitcoin at a thousand than at a hundred thousand mining same rule applies it's better to own your site at a million and a half a megawatt than it is at eight million a megawatt yep and a lot of that vast majority of that is driven by the vintage of your asic purchases but a significant percentage still is a function of how cheaply uh or cost effectively you're able

to build the infrastructure required to run those machines um but with those power prices that you're talking about being a part one of the um inputs are they being driven up by the ai the demand for ai compute like with everything the question is where some places 1000 power rates are being driven up by inflation, I actually think that power prices writ large are going to go down because of the AI demand, because the demand signal is so strong.

And so what we're seeing, at least in the American grid, is the demand signal to build new generation is enormously strong. And what happens when you have that demand signal with the type of project structures that AI data centers represent? I think over time, you're going to see cost control or rate control imposed, at least at the household and small business level, even if some of these larger scale consumers of power may pay more.

But regardless, AI data centers pay more for power than Bitcoin miners do. Their incentives or their margin profile is not nearly as sensitive to power prices as a Bitcoin mining business would be. And so when we think about what is the structural advantage that a Bitcoin mining deployment can have, the first is the cost basis for the build and the ASICs. The second is the run cost, which is 80, 90, 95 plus percent indexed on the cost of the kilowatt hour on a consumption basis.

yeah so when you look at that i know there's a ton of variables like you say but roughly if you try to boil it down to like a price per kilowatt hour that you can mine bitcoin successfully at is it somewhere in the like five cent range yeah i think that you know depending on machine efficiency obviously you know this current having epoch in the 60 to 85k kind of bitcoin range you're you know break even somewhere in the sixes okay and where do you get that cost of power now

um we mean we're we're doing it right now you know across tennessee wyoming georgia mississippi but so if you're obviously all the miners were in texas this was like where the vast majority were um i think basically every pub co is probably doing an ai play now makes sense that's why that's their incentive um if you were in texas in one of those places can you still get that price power oh you can oh yeah you can certainly get you know depending on yes there's a lot of detail under the hood if you're in texas anything from four to eight cents for a miner is totally

realizable okay and so what do you think the state of this market is going to look like over the next five to ten years like the are there any public companies that are still solely mining bitcoin and not doing the AI thing? When 21 merges with Electron, they will. Okay. But as it stands right now, there's maybe no one. True pure play. Not top of mind. Okay. Do you see that as a problem in any way? No. Not at all.

I think that one, there are phenomenally successful private companies that are mining Bitcoin. Two, there are more companies that you've never heard of mining Bitcoin than you would imagine. Yeah. um when we look at you know the percentage of global hash rate that the public companies used to represent they represent today obviously they've taken a lot of hash rate offline over the last two years but hash rates only down you know a little less than 10 so it went somewhere

hmm it is down a little less than 10 but it's probably the only actual bear market we've had in hash rate ever like take out the tiny dip for China the China ban like really ever uh yeah and so do you think that that chart continues going up into the right over time or do you think we are entering a new paradigm with this um I think that the like with any commodity production business it should get harder and thinner margin over time because the lowest cost operators should

be the most successful and durable. And by being a low cost operator, you sustain a lower margin for longer. And so, you know, do I think that hash rate is going to go up into the right over time? Not at the rate that it did over the last 10 years. I think the next 10 years, if hash rate goes up, it's going to go up slower. I think the reasons for that are varied. The first is just that, you know, Bitcoin mining was kind of late to the bleeding edge for chips.

And so when the S9 came out, it was being produced on a 14 nanometer process node. And so as you go from the S9 to the S19, that transition is really where Bitcoin mining ASICs achieve the bleeding edge. And this is some of the more long range game theory there, which is that you don't need to produce a Bitcoin mining chip on the bleeding edge for it to work. It just needs to be efficient. It needs to be efficient. enough. It needs to be cost effective enough. The chip that's in your washing machine is

significantly lower down the efficiency curve than the chip that's in your cell phone. And so the question is, what is the cost differential between the process nodes that the machines are produced on? And what are the associated efficiency gains or losses based on that. And so if you told me I could have a half as efficient ASIC for 10% of the cost, the math starts to work again. And so I think that, you know, there's a hardware manufacturing

trade-off set that we haven't even started to attack as an industry or really as a distribution based because, you know, if you're able to get, you know, if you're able to get a space heater that has a Bitcoin miner or Bitcoin mining hardware integrated into it, as if that would ever exist, there's a great ROI opportunity if you're able to structure it the right way, or you're able to, you know, net it down against other costs that you're already bearing for your current quality

of life. And so I think that the opportunity for proof of work distribution doesn't look like public companies raising billions of dollars to build mega campuses because the market has sent the signal that there's a better use for those campuses. That doesn't mean there aren't huge amounts of power consumption opportunities to be pushed through a proof of work Bitcoin use case. And so some of those are going to look like smaller mid-sized businesses. Some of them are going to maybe look like huge businesses. We're still seeing, you know, the companies that are pure play, at least on the private side, do so at enormous scale and they're doing it successfully.

but I think that the idea that if the short list of public companies that did this at enormous scale over the last five years don't do this forever, Bitcoin is vulnerable. That's not a thesis I particularly subscribe to. No, me neither at all. And I actually think you can flip it the other way. I think this is really good for Bitcoin. This is decentralized in the hash rate in a way that we've not had for quite a long time. The question is, where do you think that goes? Do you think it stays within America or are the opportunities further out on the frontier? Is it Central America? Is it Africa? My sense is that there's going to be an all of the above

type of distribution. That means that America is probably a smaller percentage of global hash rate T plus 10. Good. Not bad. Distributed. Yeah. Like to me, I don't think in terms of good and bad. I think in terms of the market speaking to the economic realities of the world. And so the distribution of hash rate is very likely to happen. The reason for that is that there's a power cost arbitrage that's going to happen where you're going to take proof of work mining to

the lowest cost environment that it can achieve. There's also just a regulatory arbitrage that's present there as well, where if you asked me, am I willing to build an $8 billion AI data center in country X, Y, and Z? I'm probably going to say no, or at least our tenants will say no. But if you said, am I willing to put $45 million worth of Bitcoin mining hardware there and it pays off in 11 months, all of a sudden the answer might be

overwhelmingly yes. And so not only is there a cost of goods sold operating expense opportunity to more broadly distribute the Bitcoin network. There's also a pave the path regulatory aspect to it where the capital intensity of some of these other compute use cases is just too high to take on the geopolitical risk of certain jurisdictions. Even if those are great jurisdictions to build data centers on an engineering basis, Bitcoin mining can pave the way there. Oh, and by the way,

they're going to be able to put additional revenue back into those utility systems. They're going to be taxpayers. They're going to improve quality of life. They're going to be a force for property rights in the places where they go. And so over time, maybe the next five to 10 years are great Bitcoin mining opportunities there. And then there's an opportunity for a different type of data center to be the highest and best use of those megawatts. Our thesis has always been power first. That is the scarce and fundamental input to these tools of human

flourishing. Whether that's a large language model, whether that's a traditional data center, whether that's Bitcoin mining and proof of work, these are compute resources that are denominated in megawatt hours. And so the degree to which a megawatt hour denominated compute layer is getting built internationally, different priorities are going to be the most economic and cost-effective at different times in the life cycle of these infrastructure projects. And so the entrepreneur's job is to identify

the right opportunity, the right risk, the right cost of capital, and the right return profile, and then bring those ideas to bear inside of the market and get paid for it Bitcoin is a hard asset It took discipline to build your position and the platform you use for Bitcoin back loan should have the same discipline That Ledin the lead sponsor of what Bitcoin did When the next opportunity comes along a Ledin loan puts liquidity in your hands without selling what you spent years building. Loans are funded in hours, not days. There's no monthly payments. There's no penalty for repaying early and all the rates are published

upfront. The larger the loan, the lower the rate. Eight years of protecting client assets with a perfect track record and over 11 billion dollars in loans have taught leden what really matters your bitcoin stays custodied it's never lent out to generate interest and leden doesn't ask you to take its word for it every month leden publishes its open book report with all metrics independently reported it's proof over promises hard assets higher standards check your rate at leden.io forward slash wbd which is leden.io forward slash wbd bitcoin-backed loans involve risk do your own

research before taking out a loan. APR eligibility and terms apply and may vary by jurisdiction. Every Bitcoiner eventually has to answer one question. If something happened to me, could my family access my Bitcoin? Would my children inherit what I've spent years stacking? That's why Anchor Watch builds Bitcoin custody solutions designed to protect you and your family against real life accidents, errors, kidnappings, and even death. Every solution includes their inheritance protocol designed to help your Bitcoin reach the people you intend it for. and US-based customers can also add Bitcoin insurance backed by Lloyds of London.

Whether you're a self-custody expert or want multi-institutional support, your Bitcoin estate plan should not be an afterthought. Bitcoin is only generational wealth if it can be passed down. So make sure your family can access tomorrow what you've built today. AnchorWatch is your custody, your way. So visit anchorwatch.com to get started. That's anchorwatch.com. If you're a taxpayer in the US, you need to talk to your CPA about Bitcoin mining before the tax year ends. Mining equipment may qualify for 100% bonus depreciation. That could mean deducting the full eligible cost in year one,

and that's before any of the Bitcoin they mine. With Abundant Mines, you own the miners and keep every single sat they earn. Their team handles the setup, hosting, maintenance and repairs, all for a flat fee, and you keep 100% of the Bitcoin you mine. And here's what really sets them apart. If one of your machines goes down, their Hashrate Redirect covers you with Hashrate from their own fleet to keep your Bitcoin production going. To count this year, the equipment generally needs to be placed in service meaning plugged in and running by December 31st. So you need to move quickly. To find out more and get a free consultation, head over to abundantminds.com forward slash WBD.

And What Bitcoin Did listeners will get a free month of hosting with your miner purchase. Check them out at abundantminds.com forward slash WBD. So how much of CleanSpark now, as of sort of this moment, is in AI compute rather than Bitcoin mining? And what do you think that'll look like in say 12, 24 months? Yeah. So the current portfolio for the business is that we're running in the zip code of 50x a hash for Bitcoin. Those have been incredibly profitable deployments for the years that we've been building them. They're also the financing leg of how we evolve the business to take on the AI and more traditional data center opportunity in front of us.

Um, the current capacity inside of least AI data centers is 248 megawatts at our Sandersville campus. But what we are working on right now, and, and one of the reasons I don't sleep is, um, we've got an additional 885 megawatts in Texas. That's directly earmarked for AI data center capacity of that 585 megawatts are designated.

in the batch zero baseload, which means that all the regulatory approvals are in a strong position for those projects to move forward. The last 300, we'll find out early next year. And then as we look out across the remainder of our portfolio, there are sites that may be best suited for Bitcoin mining for the long term. There are many of those sites that are going to have a big AI data center capability because there aren't that many two, three, 400 megawatt

campuses available. And the demand profile that's out there today is so strong that a 20, a 40, an 80 megawatt site is going to get more attractive as the inventory gets a little shorter. So, you know, we're thinking, you know, step one is lease the big campuses that we have, get those built, get those cash flowing in parallel, come up with what is a smaller form factor footprint look like. But the beauty is, is that we can continue to cashflow those assets all

the way through that process. And we don't have to make a decision. We have the luxury of choice. But okay. So put that in plain English for me, like roughly what percentage of like all the power consumption that you guys are doing, how much of that is AI right now? um right now we are looking somewhere in the 15 18 range we expect that percentage to dramatically climb yeah so if you ask me this a year from now it may look like 70 80 okay um and then if you ask

me out past the having there's there's other forces at work that are going to dictate kind of how that ratio shakes loose. What I will say is that on a net new site basis, we are only looking for AI data center campuses on a growth basis. How do you sort of trade off the, I don't know if it's risk, it's probably not risk, but the additional requirement that you need a customer for those, whereas with Bitcoin, you don't need a customer. Is the demand just so extreme that you can build as much as you possibly can and you're going to be fine? Well, so I think that there's

some consideration for us around sequencing. So there are parts of the project that we won't do before we have a customer. Buy the land and contract the power, we will absolutely do that before a customer. Grade the land and gravel and stump the trees, we will absolutely do that before a customer. Build a substation, we will do that before a customer. Once you get to the concrete pad that the data center is built on, that's where we draw the line and say, we actually need the

customer and their requirements in hand and least before we're willing to take on that piece of the project scope, because then you're putting dollars at risk on a really speculative basis when if the technology were to change or the customer were to have a dramatically different set of preferences, we might have to go do something twice. But we know that land and power acquisition, substation development and campus grading all of those are well within scope for us to take on prior to a customer being locked in and like over the last 18 months two years data center companies

have been absolutely ripping but it's kind of slowed down a little bit it has the demand slowed down at all or is that just still through the roof no our view is that the demand is just as relentless as before maybe more so um let me ask it to you a different way are you using less ai today than you were a year ago absolutely not and that's what everybody on the demand side is also seeing but even even if that was true like i think the demand could have still been over egged a year ago and we it we so we have approximately two percent market penetration of ai tools that was

all it took to sell out the entire planet's compute. Wow. And those are basic tools, right? So the other thing is like the token density idea of use cases comes into play where, you know, think of a Google search as one token. Yep. Think of a open AI query, a chat GPT query or whatever your open source tool locally run looks like, think of that as like 100 tokens.

Think of video rendering as like 10,000 tokens. So the distance between a Google search and a unit of video rendering is like roughly 10,000 X. All of the use cases live closer to Google search than video rendering today. That doesn't include the advent of agentic behavior, you know, agent to agent interaction, all of those are much more token dense activities within, you know, the transformer process that, that sits behind these technologies. And so we're going to see

broader distribution and we're going to see deeper intensity of use cases both happen. And that's, what's guiding us towards this idea that the demand profile, not only is it dramatically underserved, it's also probably ramping up. That's, that's insane. two percent wait so two percent was the entire world's demand all of the compute that was built between 1990 and 2024 completely sold out with one to two percent market penetration of ai tools

that's insane um have you obviously have been following the pushback against ai data centers it's uh it's quite nice to see it not being bitcoin at the moment oh you mean i didn't spend the last five years telling people that bitcoin wasn't going to ruin the world only to find out that now apparently data centers are can we go through like the biggest concerns that they have i would love to okay so i think probably the largest concern from the anti-ai data center crowd is that they think it's going to cause their energy prices to go up yes that is yeah

let me i'll give you the four things i hear if you hear others i would love to tackle them first are utility prices the second is water consumption yep the third is noise yeah the fourth is light pollution pollution and i think like noise slash community disruption quality of life issues um energy prices you know i'm i'm a left curve guy typically so there's a great study there's a good twitter follow called crmue c-r-i-m-e u-x a-a-u-x it's a

he he's a data scientist he's great um he shared a thread about basically the market data doesn't support any of this okay where he did a state-by-state cross-sectional observational study where he just looked at all 50 states in the u.s which states have the fastest data center growth and the most data centers deployed and which states have the highest and lowest rate of change in electricity prices. Two findings. The first is energy prices have risen. Yeah. Go check inflation.

It's also risen. The second is the states with the most data centers, their energy prices rise the slowest. Wait, can I just go back to that first point you made? Have the prices of energy match the price of inflation the increase in inflation are they the same or is it increasing at a faster pace it's complicated okay like not everywhere is the same so we're sitting here in tennessee um 60 of our utilities are produced by our electric utilities are produced by hydroelectric and nuclear if you go somewhere else it's overwhelmingly natural gas so the the

commodity inputs to different segments of our grid are going to look dramatically different there's a lot of noise in those numbers. Okay. Okay. Sorry. I cut you off then. So, so the, yeah. So the, the, the first is that energy prices have risen, but inflation has as well. The second is that if you live in a state where data centers are growing rapidly, your energy prices have risen slower than if you live in a state where data centers have not grown rapidly. We can argue causation and correlation for sure. Yeah. But ultimately what we've seen

is that the demand signal to the utilities to build new generation and invest in transmission quality, those demand signals are coming from data centers. And so the utilities that are going to have the bandwidth to continue to have stability in their rates are going to come from the places where there's the most demand. I look at, or where there's the most demand with abundant supply coming to meet it. I think that the risk that we have is that data centers resemble the housing market, which is, you know, if you live in Austin, Texas, you've seen rents go down

the last four years because they basically deregulated the building of new housing. If you live in New York where they've chosen to opt for cost control rather than deregulating the building of new housing, you see regulatory capture if you have a rent control department and a blow off top in the rental prices for everything else And so the answer to higher prices is typically higher prices which means we need to build more generation

We need to deregulate the ability to invest in our grid. We need to make growth the priority because if you can grow supply, you can meet the needs of demand and you can serve the marginal consumer more effectively. but in that scenario can supply keep up with demand like demand at the moment is through the roof supply is slow to come online like will there be a period of time where maybe energy price increases is the reality um i'm hesitant to say that won't ever happen but i also think that the

utilities and the jurisdictions where these data centers are being built are taking aggressive steps to avoid that we're signatories of the president's ratepayer protection pledge as part of our data center efforts what does that mean so so the the president and the white house outlined a number of community values that data centers need to commit to meeting in order to be one good neighbors but two good economic participants in the local environments where they build and so one of them is we don't want to make people's prices go up how do we do that we pay for our

own infrastructure upgrades we build our own substations we invest in the utility when appropriate And so you're seeing this take a bunch of different shapes all over the country. But ultimately, the data center developers are responsible for bearing the cost not only of their electricity, but also the cost to the utility to serve that electricity. That sounds like it's in our house. Are there any of these data centers that are actually directly subsidizing energy costs for locals? We've seen a couple examples of that.

I don't think we've seen it totally broadly yet. what we are seeing is the data centers are committing to community benefit agreements or direct investment. We've done a lot of that transparently over the years. And what does that look like? Case study in Louisiana is that there's a large data center that went in there and they paid all the teachers a $50,000 bonus, all the public teachers. If you look at our facility, and this isn't even on the AIDC yet, but on just the Bitcoin mine, if you go

one of our sites in South Georgia. Before we got there, the city was down to their last police officer, like single police officer. And how big is this place? Many thousands of people. The county was handling all of the inbound calls out of the sheriff's department because the city couldn't pay for it. We came in. We are the largest taxpayer in the city. We're the largest taxpayer in the county. All of the police force at the city level has been restaffed,

and all of the teachers are making more money than they ever have in that location. Why? Because we pay a lot of taxes. And one of the interesting nuances there is that in Southern Georgia and across the MEAG network, Municipal Electric Authority of Georgia, those are utility-owned, or sorry, those are municipally-owned utilities, which means the city is the utility. And so when we pay for our power, the revenue is actually kept in the city. So they're getting paid twice. They get paid on the taxes. They also get paid on the power bill.

And the combination of those two things has let them do a tremendous amount of good in the community. We also cut checks to police department, fire department, local schools. You know, we put a big, a big CleanSpark logo on the high school football. Let's go. Jumbotron because nothing's better than football in Southern Georgia. but it's that kind of stuff where you know we don't we don't want to show up just to extract like that's not that's not the right way to do business when you look at these these data center tenant relationships those are decades long which means we're going to be a participant in that

community forever and we want to do that in the right way with the right priorities and we don't know what those priorities are you may go to one one city and roads are the most pressing issue Great. Let's get involved in helping improve the quality of roads. You go somewhere else, like where I mentioned, police department was the most important issue. Let's get involved and let's be a force for good in the areas of the most need. And we don't know what those areas are, but we build a dialogue and a relationship with local leadership as quickly as possible

because we want to come in and make things better. Can I play devil's advocate? And I want to make it very clear that I don't believe this, but it is some of the discourse you see around this is that people look at that essentially as like a bribe to have something in a community that people might not want otherwise um i think that that one that that's not a fair characterization of it because ultimately we are doing lawful business in a lawful way across this country

yeah i mean there's a lot of these arguments that aren't a fair character is it characterization I think that, you know, don't let, don't let the truth get in the way of a good story. No, I think, I think ultimately we, we view it as a win-win and that's something that we've always prioritized is we can't win and have the other side of the table lose. That doesn't work with the tenant relationship. It doesn't work with our contractors who are responsible for the construction and development of our assets. And it doesn't work with the, with the, you know,

regional or local community it doesn't work with the nation you know we we believe you know if we if we went out and and um and didn't have a national security viewpoint on energy and on data center capacity i wouldn't wake up as fired up every morning as i do yeah no i it's very clear that the incentives for everyone needs to be completely aligned you need to have those people on side and can we take a step back before we go on to the water thing and because i i've heard you talked about this before and I think it's great. Can you tell everyone, put it in context, why we

need to consume more energy? Yeah. So there's a great chart. I think it's from our world in data and it basically maps every country on the planet based on how many megawatt hours they have per capita. And then that's one access. And the other access is GDP per capita. That's not a perfect metric, but it's, it's representative, um, effectively the relationship between more energy

per capita and more GDP per capita. It's a straight line. Once one, one to one, there is no, you know, the Venn diagram is just a circle. And so if we expect to have a better quality of life for ourselves and for our children, we need to generate and consume more electrons because the things that it's given us. Quality hospitals, quality transportation, quality nutrition, quality education. Everything that we enjoy as a modern species is a function of our ability to

convert potential energy into real and consumed energy. I think one of the very clear, I saw a chart the other day which was the consumption of energy in the uk and it peaked on a per capita basis in 2003 and i don't think there's anyone in the uk right now that would say life has got better over the last 20 years like it's very clear we need to turn that chart around yeah i mean i i use i use silly examples because i think they're helpful anybody interested in going back to

horse and buggies anybody interested in walking i mean i like a walk you shout out steve to get to to get to get to New York? Absolutely not. So the reason that we're able to do that is one of the highest density storage systems of energy is gasoline for cars, is batteries for Teslas, is jet fuel for airplanes. That is energy. The ability to harness energy in a mechanical process is the reason that we have freedom of mobility. The reason that we also need energy is go look

at the population chart of Florida over the last 200 years, there's this funny invention called HVAC that made Florida a habitable place because in the summer it's hot. And so the ability to have a place where you're able to build a thriving economy, a state like Florida, wonderful state, it's only unlocked because we have access to the electricity needed to power millions of homes worth of HVAC. Otherwise it's uninhabitable. Dubai, Singapore, Hong Kong,

all these places. All these places. The test that I use other than I like the walking one, but walking is flawed because we like to walk. Go to your office manager and ask them to shut off the HVAC for the summer and see how much work gets done. We convert energy into productivity through modern life. In the world of Bitcoin, security isn't just a feature. It's the entire foundation. And that's where my new sponsor, BitGo, come in.

For over a decade, BitGo have been the backbone of the entire industry, securing billions of dollars of Bitcoin for major funds, exchanges, and institutions. Following their 2026 public listing on the New York Stock Exchange under the ticker BTGO, BitGo are now bringing that exact same enterprise infrastructure to individual Bitcoiners. Self-custody is still the gold standard. but as your stack grows, it's worth thinking about whether diversifying your custody can make sense. BitGo pioneered commercial multisig, and their cold storage combines distributed keys, offline key generation, and geographically separated vaults.

Head to bitgo.com to get started with institutional-grade custody. Your Bitcoin, secured like an institution's. That's BitGo. Recent hardware wallet vulnerabilities have definitely made me rethink my entire security setup, as I'm sure it has for a lot of Bitcoiners. And the uncomfortable truth is that no single security product can promise there'll never be a problem, which is exactly why the design matters. BitKey is a self-custody hardware wallet that uses three keys created in three separate environments. There's one on your phone, one on the BitKey hardware itself, and one on a secure BitKey server. Two of those three keys are required to sign any transaction, so a problem isolated to

just one key is never enough to move your Bitcoin. So if you are rethinking your setup right now, as I'm sure a lot of people are, you should really think about the BitKey. The hardware looks great, the app and the user experience is second to none, backing up and restoring your wallet is seamless, and they even have inheritance built in. So I really don't think you'd be disappointed. So download BitKey today, use the code WBD for 10% off the second generation BitKey, which is this device right here with the screen. Download the app, download BitKey, use code WBD and get 10% off. Bitcoiners hate single points of failure, but your phone has

three of them. One network, one identifier for life, and one carrier employee who can be taught into handing over your number. And that's where CAPE comes in. CAPE fixes all three. It's the only US carrier running two major networks on one SIM. So if one drops, you're on the other. Your identifier rotates every 24 hours. So to the network, you're a different person every single day. And the big one, your number only moves with a 24 word phrase that you control, just like a Bitcoin wallet. There's no call center worker, not even CAPE can move it. CAPE provides the same premium reliable service

you'd expect from any other carrier, but designed with privacy and security as its first principle. If you're a Bitcoiner in America, I don't know why you'd use any other carrier. Head over to cape.co forward slash WBD and use code WBD for 33% off your first six months. That's cape.co forward slash WBD and use code WBD. So how do we change the narrative then? We will get back to these points. I do want to go through them all. But how do we change the narrative around power consumption? Because there's a half the country in this like football team type politics is very against this.

um how do we convince them just be so great they can't ignore you that's the you know that's the steve martin school of economics which which i subscribe to is just the the answer is is that you may not be able to convince them you may not you may not be able to but if you build a product so great that it cannot be ignored there will be no choice but for the market to speak and to reward the maker of that

of that asset i think all the time about you know the billionaires should not exist rhetoric has anyone stopped shopping on amazon bezos is he's no longer the richest man in the world but he top three and the reason that he is so rich is because he built a product that delivered so so much more than his net worth to so many people Yeah We have the ability to order something and it comes the same day The same day Like not tomorrow, not in a week.

The same day his product is so effective and our need as consumers is so urgent and so rewarding of timeliness and quality and product selection that he earned 100 plus billion dollars by making a product that is so compelling. You look at Elon Musk, a billionaire soon to be trillionaire. They shouldn't exist. Well, they should when they deliver that much value. Go on the road. The Tesla is an incredible product. The ability to put satellites and

make everywhere on earth internet available, that is worth a trillion dollars of GDP and many, many, many times over. And so there's a, you know, this is the fundamental divide, which is, do you believe that there is a positive sum relationship between innovation and quality of life and the distribution of those goods and services to millions or billions of people? or do you believe that we are fighting over the same pie that we've been fighting over

since, you know, year zero? Yeah. We, we are not, we, you know, the, the, the other one I love is, you know, John Rockefeller on a inflation adjusted basis is still probably close to the richest person who's ever lived. And he grew up using the restroom in an outhouse. I don't want to go back to that. I don't want to go back to that. Would you trade places with him? not a chance you know the other one i'll keep going let's go the other one is you know warren buffett's worth x hundreds of billions of dollars um would you trade places with him

well i don't do i have to trade age exactly so now you're selling me that 60 some odd years of difference is actually more valuable than a hundred billion dollars fuck no it is i wouldn't trade places with oh sorry no i'm saying there's no way i'm giving up 60 years so great so now really what you're saying is the only thing you can't put a price tag on is time there's no price tag because i'm offering you a hundred billion dollars to give up 60 something years of your life

and you won't say yes no exactly do you think ai itself can change the narrative because like of these people screaming about data centers the vast vast majority are using ai and if they're not they're going to get left behind um i don't think that a lot of the people who are anti-data center understand the relationship between the product and the data center and that's not their fault that's because this industry is four years old on a good day and people are busy they don't

don't have time to look totally so you know so i i think i think the anti-data center perspective is is broken down you know roughly into like three or four higher level inputs the first is that wages have not kept up with inflation for the last 30 years yeah and that colors everything when people feel like their time is not being valued and when they can't get ahead in life

don't expect them to be happy they look for a bogeyman and and that's not irrational yep um the second is is highly related to that which is they hear about the data center they're very worried about utility bills and their ability to continue to to pay for their quality of life that being put under threat, very scary, very rational. Yep. The third is the biggest investment most people will make in their lifetime is in their home. And so if they see the build out of this infrastructure as a threat to their house value,

it's another place where economic insecurity can rear its head. Also, totally rational. I don't agree with any, at least with the last two, the first about wages, keeping up with inflation. I'm not sure I can help with, and I'm not sure AI data centers are going to have a perspective on one way or the other, at least yet. But in terms of utility payments and house value, I think we've got a great argument that we help both of those things. But the fear around those two items is reasonable and is something that I personally can empathize with.

yeah um and then you know the the last one is there's also a massive international propaganda campaign that's happening and being distributed broadly which unfortunately the people who are perpetrating that are really good at it because it's clearly working they're really good at it and they're also building out a shit ton of power they are um all right let's get back to this so water is the second one and this is like i think the most silly but i'll let you explain um we hear a lot about data centers being a huge threat to water security

i don't kind of know how to say this other than just that it's just not true so why don't you explain what their argument is in as much detail as you know yeah their argument is that the water table is shallow and water is scarce and and we're going to run out of it and that data centers are these huge water hogs that are going to that are going to accelerate the draining of scarce water availability yeah and i can't speak to drought conditions nationally there may be places that

have water insecurity that do have water insecurity this is that's not that's not the argument yeah the argument is that data centers introduce almost zero risk to any of that and certainly modern data centers introduce zero risk. So there's part of this issue that's a kind of 20 year legacy problem, where if you built a data center 20 years ago, the predominant technology for cooling was cooling towers and evaporative cooling. What does that mean? It means that you push water across the chip, the water gets very hot and vaporizes, and you evaporate it in a tower

on the side of the building. And then you add, then you pull new water, you push it across the repeat the process. Those data centers used a lot of water. Those are not the data centers that are being built. Why? Because the market economy is a powerful force and we had to innovate to not have to pay huge power bills. We are rational economic actors as a sector. And so the technology for closed loop water systems was introduced and the amount of water usage dropped by 99%. the wave to think about this there's a few ways to think about it one of them that i like is when

was the last time you filled your car with water never but your ac works there's water in your ac it's just closed loop system you use the same water every day the only time you'd need to add water to your car is if you pop a hose yeah that's never happened to you that's how these data centers work just like the radiator in the car so now you're still cooling it with water but you're just not evaporating it yeah so the way the way to quantify this is i'll use our our sandersville

project as a good case study less than one olympic swimming pool one time at phillip less than 10 houses on a go forward basis because there are bathrooms and kitchens at the data center so it's just not an argument it's just not an argument okay so that's water done other things that make make me insane golf courses use like 200 times more water than data centers if you told me and forgive your your listeners who are golf fanatics but if you told me i had to choose between data

centers and golf courses i'm picking data centers i'm picking neither i want to play golf well i don't want to play golf in a desert yeah and you won't and you won't have to certainly not because of data centers um okay so water is just not really a thing that's just complete it's a fact problem yeah okay noise is real though i so i've not been to an ai data center i've been to big bitcoin mining facilities noisy as hell yep um what is it like an ai data center much quieter so the first reason for that is bitcoin mining uses well air-cooled bitcoin mining in the way that

it's been largely done is used with a hot aisle cold aisle isolation and passive cooling environment which means one of the walls is really big and open to the side data centers are built in buildings so there are walls first piece of of technological difference the other is that the way that the engineering works and i'll go back to sandersville as another case study you're looking at mid 50s decibels 100 yards from the building what's mid 50s like what would this aircon in here be

oh let's find out okay i have a decibel here on my phone you know the aircon is going to get down to temperature before you manage to do this let's do it okay 53 so this is 100 yards from an ai data center and you're not building these next to houses is so so so another example in texas for us is we bought 270 acres we're going to build on about 10

to give you a sense of scale until there's demand to fill the rest of it no no we'll run out of power okay we'll run out of power long before we run out of land so why do you buy so much land because we want to have the setbacks on the sides we want room so i mean if you lived next to this amount of noise it might annoy you a little bit if you're 100 yards away if you're 100 yards away if you get to the property line you shouldn't notice it we'll go walk outside we'll do the decibel meter again and you'll but the important thing about and this is important about noise

noise is perceived logarithmically so if you go from 55 decibels to 45 decibels that's a 50 reduction in what you hear so every 10 decibels is a is a 50 reduction or a doubling okay the other direction okay so is there any like reality in the argument of noise is is this ever a problem anywhere yeah there are places where there's a problem there there are places where there's technological um missteps around engineering or they weren't as considering of that factor

on design there are places where it's built too close to things like there are places where where it's not ideal we have thought about all of this we've been operating in bitcoin mining for at CleanSpark five plus years, Harry for far longer than I care to admit. We've seen these problems. We've engineered around them. We are doing everything within our power to create a near zero impact to noise. So if you were to try and be as fair as possible to that side of the argument, as charitable as possible, is there anything that is frustrating

about having a data center somewhere near you in your community? I will quote the undersecretary for the department of energy and and i asked him to qualify what does a data center look like from from your perspective um boring and valuable like bitcoin so do you think how how do you think then this plays out over the next few years because it's

going to be a really hot topic going into the midterms probably even next election is there a point where all sides have to concede do you think yeah there is so two years from now hundreds of communities across our country are going to have a year or two of tax benefit in their pocket they're going to have a year or two of hearing it seeing it and living near it and the problem is is that a lot of these deals have been signed a lot of these projects have

been started but nobody has the seasoning of what is it like on a go forward basis once we have two three four years of this in play look look at loudon county virginia highest data center density of anywhere in the country also have some of the lowest property taxes also have some of the highest per capita gdp of any state or county and so you've got examples where and they've been data center alley for 30 years. The places where there are lots of data centers are having great

economic and quality of life outcomes. So can we talk about grid balancing? Yeah, because one of the sort of very powerful narratives around Bitcoin mining, especially when you look at ERCOT as an example, was that when demand got so high, Bitcoin is interruptible. You can turn Bitcoin mining off, give power back to the grid. And AI doesn't have that ability. Is that right? Um, kind of. So not in the traditional sense that Bitcoin mining would do it, where there's a lot of toggling and active management.

So the way that grid balancing interacts with the AI data center is different fundamentally than Bitcoin mining, where Bitcoin miners take a price signal from the power market and they react to that. they also engage in contractual curtailment. We're here in Tennessee, CleanSpark's the largest enrollee in the PowerFlex program of anybody in the system where there are contractual guardrails around how power gets managed. Two different versions of the world. If Texas is highly

economic driven, Tennessee is highly contractually driven, similar result. A couple of percent a year, you give back the right to the utility to move power around. If you look at the Senate Bill 6 out of Texas, they get to a similar concept, which is called their low voltage ride through provision, where if they have a catastrophic outage or system failure of any kind, data center loses power first. So it's a hierarchy of outage rather than a runtime

designation. That's a different solution to the same problem. But okay. So the classic example in Bitcoin was, I think it was 2021. There was the big cold snap in Texas. Um, people died like terrible event. And during that time, all the Bitcoin miners turned off again, because of the power pricing mechanism. If it gets too expensive, they want to turn off anyway. If that happened now in a similar scenario, it would be that instead of them turning off because of the economic incentive, they'd turn off because regulatory, they would just get turned off. They won't choose to turn off yes okay but you get the same outcome yes and does that not how much does that sort of

break the economic model of the ai data center in that moment it doesn't they sign up for it okay and they don't get paid though like bitcoin miners might get paid in that scenario they don't get paid correct okay but really there's if you're just a normal person living in texan in a scenario like that the same outcome yes the the big loads would get shut off first the households if they still ran out of power would also have outages, but they wouldn't be first in line. They'd be last in line. And do you think there's a scenario where these data centers will want a portion

of it to be Bitcoin mining to do the more interruptible load, the more pricing mechanic of the interruptible load? I think utilities may want Bitcoin mining I don think the data center When you say utilities do you mean like ERCOT might want it themselves And use it as a load bank Which allows them to ramp up power use that power But then when they need to. So we already have this concept in many places called spinning reserves. Where if you've got a peaker plant. You know, let's just use a peaker plant versus a nuclear reactor. Peaker plants want to run 10% of the year at most.

But they're considered what's called a spinning reserve. And so if you're in, you know, PJM, there's what's called a capacity payment. They get paid every day for being available, right? You're an on-call doctor. Basically, you don't know if you're going to get called, but you're paid like you're on-call. That same idea could be flipped from a peaker plant, which is highly intermittent on their generation, to a nuclear reactor, which runs in a flat line. and you have a load balancing resource associated with the gen that rather than spinning gen on,

you take load off. And that makes economic sense as well, because instead of running 10% of the year, you can run all the time. That's correct. Interesting. That actually was one of the questions I wanted to ask you is where do you think the power is going to come from in the future? What is the vast majority of the power right now for these AI data centers and how do you think now change. Yeah. So the way that these data centers work and the way that all electricity here works is that when you turn your light switch on, you don't actually know what made the electrons that make your lights work. The data center doesn't know what electrons make

its servers work unless you have direct serve behind the meter generation, which is a separate kind of topic here. But ultimately, as we look to the future of the American grid, it's going to be an all of the above strategy. There's a place for solar. There's a place for gas. There's a place for nuclear. And we need to invest in the build out of all of it. But critically, regardless of the generation type that we see grow, we need to continue to invest in our

transmission and delivery capabilities because that makes for a more robust network delivering power. And that ultimately becomes the constraint. The transmission is more of a constraint right now than generation. Okay. And so how do you move the pipes and plumbing around the system in order to serve larger and more dynamic compute use cases? That's where a lot of the conversation on the utility side is happening. But, you know, we had a panel with the same DOE official earlier this

And his viewpoint is he wants to add 300% more nuclear to the country over the next decade, or I sorry but over the next several decades Nuclear is the like when you look at all the different types of uh generation seems the best like it has the best trade in I forget about economics I don know the the economics of it but from you know base low power relatively you know, benign to the environment. Um, but that's been held up by regulatory for the last and they're hard, they're hard, they're, they're hard engineering assets. If you did a gigawatt

scale, you know, AP 1000, you're looking at a lot of transmission requirements to get the power off the plant and into the system. Okay. So not only is it the generation asset that you have to consider, it's also the infrastructure layers around it that you've got to implement in order to distribute that power to the users. Do you think we'll see an explosion, a growth of nuclear power plants over the next 10 years a proliferation a proliferation do you think but do you think like with the

regulatory tape being removed this is going to be you know where the vast majority of any generation comes from in the next 10 20 years i'm really hopeful me too really hopeful i think i think the next 10 years are going to be a lot of deployed natural gas generation resources and i think the 10 years that sets the stage for the 10 years after that where nuclear is going to become the fastest growing source of generation let's go all right i want to close out with a different question you started grid what eight years ago 2018 okay if you had have started a company today

would it have been bitcoin mining or would it have been ai i start an energy company and then i figure out where those electrons are supposed to go later but ultimately and that's that's how we view clean spark now is that we we are megawatt hour denominated megawatt and acres are the are the the sort of most common denominator for our business ai data centers are going to be the place where we realize that value but ultimately being a fundamental infrastructure business is one of the most attractive things to be able to do

unless i go back and learn to code you don't need to learn to code now um learn learn to orchestrate is there anything that i've not asked you on this i wanted to try and dispel a lot of the around data centers is there anything i've not asked that you think is important to talk about um i think that the data center conversation that we had around some of the economic insecurity and some of the more macro societal trends like that's where i want the conversation to go which is that the data center has nothing to do with the things that a lot of these people are worried about if

If anything they a huge force for good But the reasons that there is a sense of discontent those are incredibly relevant societal issues that we face that we need to tackle head on just as people And so I think that if we can sort of defang the data center conversation in order to have a more productive quality of life conversation, that's a really good outcome for this whole process.

And it makes kind of the heightened temperature from my viewpoint worth it because we want to build tools of human flourishing across our nation, across the world. I think data centers are a hugely powerful economic engine to introduce the digital reindustrialization of places that got left behind during the software boom, transparently. If you lived in a place that used to be a manufacturing hub and those jobs went overseas, you didn't benefit from the software boom of 04 to 2020.

You just didn't. And that's a huge shame because there are incredible people with work ethic and intelligence and quality that could have been working on those challenges, but we're just in the wrong place. The data center and AI revolution that we're living through right now, I don't think is going to leave those same people behind again. And it's about building the type of win-win relationships that we talked about in order to facilitate those economic advantages being reflected in a more distributed and more decentralized way.

but it's also this fundamental idea that AI is only successful when the digital and the physical deployments are able to coexist in a healthy and constructive way because the software doesn't work if the infrastructure doesn't get built, the physical infrastructure. That wasn't true for the build out of the web. It is going to be true for the build out of AI and that is the biggest opportunity that's in front of us so long as we are able to have some meetings of the mind to be able to realize this

and get those economic value propositions into the hands of the people who deserve it. We need to get to the root of the real issue. Broken money. Well, amen. I've got a cool idea I want to run by you, but I'm going to do it after the show. Thanks so much, Harry. Always love talking to you, man. Always great to see you. Thank you.

Open the transcript file

Hosts and Guests

Nobody is credited on this episode yet.

Credit the people on this episode

Be the first to rate this episode

More episodes

All episodes
  1. The AI Bubble Is About to Force the Biggest Money Print Yet | Lawrence Lepard

    Nov 20, 20261 hr 17 minAudio

    “They can never stop printing.” Lawrence Lepard returns to explain why he thinks the AI bubble could force the next big wave of money printing, and why Bitcoin could be one of the biggest beneficiaries. We get into the...

    Credit the people on this episode

  2. How Bitcoin Turns Trash Into Digital Gold | Daniel Batten

    Nov 18, 20261 hr 6 minAudio

    “We looked at all the other solutions and Bitcoin mining was the only one that made sense.” Daniel Batten joins me to explain how Bitcoin mining can make cleaning up pollution profitable. With his first landfill project...

    Credit the people on this episode

  3. Why This Bear Market Is Different | Michael Sullivan

    Nov 16, 20261 hr 6 minAudio

    “I just can’t overstate how much fear and anger cloud your judgment.” - Michael Sullivan Michael Sullivan is a writer and Bitcoin sentiment analyst behind Bitcoin Sentiment Weekly. He tracks how Bitcoiners talk online...

    Credit the people on this episode

  4. Can Bitcoin Survive If AI Breaks Encryption? | Michael Dunworth

    Nov 12, 20262 hr 32 minAudio

    “We’re playing for the machines. We just don’t know it yet.” Michael Dunworth joins the show to explain why he’s willing to bet a Bitcoin that RSA encryption will be broken within ten years, and why he believes...

    Credit the people on this episode

  5. Inside the White House’s Plan to Buy Bitcoin | Patrick Witt

    Nov 9, 202648 minAudio

    “I think we set in motion a number of things that will come to fruition over the years.” Trump put Bitcoin on the White House agenda. But will the US government actually buy it? Patrick Witt, former executive director...

    Credit the people on this episode