Fred Thiel is the Chairman of the Board of Directors and Chief Executive Officer of Marathon Digital Holdings, bringing over 35 years of extensive experience in the technology sector. In this discussion we cover a wide range of issues including how Marathon is adapting to the wave of smaller miners gravitating towards artificial intelligence, what separates his firm from some of the other pure-play miners, and how U.S. politics and the upcoming presidential election will affect the outlook for the mining industry.
Forbes: Can you give an overview of how you’ve tried to position Marathon Digital Holdings within the world of bitcoin mining?
Fred Thiel: We look at ourselves as sitting at the point of convergence between energy transformation and digital infrastructure. From a bitcoin mining perspective, we’re the largest of the publicly traded miners today, operating 13 sites across four continents. We use a variety of different technologies and we’re the only one of the large bitcoin miners that has a fully vertically integrated tech stack. What that means is we operate our own mining pool, which is what orchestrates all our miners around the world and is the interface to the Bitcoin blockchain. All the other miners use third-party pools that are either Foundry, for example, or Chinese pools. We also operate our own firmware, which runs in miners and then we have developed and designed our own cooling infrastructure. We have single-phase and dual-phase liquid cooling, both of which don’t use water, so that is totally water-free, and this is a really exciting area as we move into providing digital infrastructure to the AI industry because it’s the exact same infrastructure that’s used by both industries. The other thing that is unique is we cofounded a company called Auradine, which is the only U.S. bitcoin mining ASIC manufacturer. Their miners can support multiple chip failures and still run so they can operate in wider temperature ranges. So what does that all add up to? To me, it means as a miner we have access to technologies that not necessarily other miners do and we continue to press the envelope on deploying that technology. Our data center in Abu Dhabi, which is all liquid immersion cooling, is a preeminent site. There are two data centers, a total of 250 megawatts, that have 99.8% uptime running in a very harsh environment, which is obviously the Middle East desert next to a body of water. It’s very humid so liquid cooling is the only way to go.
Our strategy initially was one of being asset light. We focused on just growing rapidly by hosting with third parties. Then earlier this year we shifted to vertically integrating and we acquired 54% of our capacity’s infrastructure, and as we continue to grow in coming years that number will get closer and closer to 100%. We tripled our capacity in 2023, we’ll double it again this year and we’ll continue to grow at a pretty great rate. Lastly, we’re the second largest holder of bitcoin of publicly traded companies in the U.S. and our strategic focus really is to amass as much bitcoin as we can. That’s our core treasury asset and we do that through mining and also going into the markets and buying opportunistically when it makes sense.
Forbes: What metrics do you pay the most attention to and how does the development and investment you detailed help give you a leg up on the pure bitcoin miners?
Thiel: A couple of things. I talked about how we operate 13 sites on four continents. We don’t have a concentration in a particular region or state. Riot Platforms is 100% concentrated in Texas today. They’re trying to solve that by acquiring Bitfarms and then they’ve acquired another small site, I think in Kentucky. CleanSpark was heavily focused on Georgia. Now they’re focusing on Kentucky as well. The challenge with being focused on just one area is you’re very susceptible to your energy utility and energy market pricing in that region. In the case of Riot, they have the double whammy of two big sites in Texas and you have curtailment issues, weather issues and community issues, which in Texas are not small things. In the case of CleanSpark, they operate lots of very small sites. Operationally that is a very different issue because now you have to have people spread all over managing these disparate small sites and it may mean you don’t have the opportunity to really aggregate the scale and get the benefits of scale in the same way. But Marathon is operating on different continents, so we also have the ability to avoid the risk of regulatory regime issues in the U.S. For example, in the Biden budget, there were attacks specifically on energy used by bitcoin mining. The Harris administration seems like they are befriending AI, but they are still going to be antagonistic to bitcoin mining. By the way, you can’t give AI energy at one price and not give bitcoin mining energy at the same price. Not only would the industry attack that issue, but it’s a type of commercial discrimination, which could be viewed as illegal. But more importantly, there is a huge symbiosis between AI and bitcoin mining because AI workloads vary just like demand on the grid varies and bitcoin miners have a unique ability to synchronize their load to AI miners loads such that you don’t get this huge variance of demand on the grid. So when you marry bitcoin miners and AI operators together, you get a much better grid profile. You have the ability to stabilize the load, and you have an ability to use more of the intermittent energy—solar, wind—than AI can use.
If you look at our technology piece, operating our own pool, it’s very transparent to our investors; they can see every block that we win. By operating your own pool, you do a couple of things. One is you pay a pool operator a percentage of your winnings for them to operate the pool. It’s 1% or 2% depending on the size of the miner; we don’t have to do that.
Then we have systems like Slipstream, a non-standard block submittal service that people in the Ordinals world really love. It allows us to get higher fees for submitting blocks to the network than the normal mempool would pay, so that gives us additional revenues on top of the traditional transaction fees that we get. We also have a Layer 2 called Anduro, which people are now starting to develop applications on.
Forbes: Can you speak a little more about how Anduro is constructed?
Thiel: It’s a Bitcoin L2 that is essentially a clone of the Bitcoin network. It can process transactions much faster. It can do things that you can’t really do as easily on the L1 of Bitcoin, but it anchors all its security, and it’s pegged to the L1. You can, for example, do transactions in a much easier way than you can with lightning. Lightning requires you to open channels and close channels. This is a true sidechain, so you could build a digital identity network on it. You could build a MLS for real estate on it. You could build a transaction processing system for credit card remittances, whatever you may want to have. One of the applications that’s been built on it, which will be released soon, is called Alice, and Alice is a full EMV system from the Ethereum world that lets you take an application built for Ethereum and port it directly to Bitcoin.
And then going back to our liquid cooling technology and the ASIC technology, we have technology advantages in that we can build miners specific to the use case. So in Finland for example, we have a small data center that heats 11,000 homes. How are we doing that? We’re doing it through district heating by capturing heat from the miners. We have the ability to build custom miners for use cases. We recently made an investment in a company in San Diego, and what they do is they go in and they over-dimension solar systems for homes with solar panels and batteries and they use the excess capacity to mine bitcoin. Realize that solar energy has zero marginal cost, and so if you put 10 kilowatts of panels on a house versus six kilowatts of panels, the only difference in cost is the capital expenditure for the panels. Then all that extra energy can be used for bitcoin mining. More importantly, you can also heat water with the heat from the bitcoin miners. And what they do is they provision this and then they sell electricity to the consumer at 25% below market price.
That fits into this model of the long tail of bitcoin mining, which is highly decentralized bitcoin mining, putting it out at the edge and leveraging zero-cost energy. So what you’ll see from us over the next couple of years is a move towards the cost to mine being covered by services we’re providing to the parties with which we’re partnered. That’s an obtuse way of saying we believe–and we’re in the process of signing agreements with large partners–that people will pay us for the benefits of marrying bitcoin mining to different applications. And in doing so, our cost of energy will go towards zero.
Forbes: Hash price continues to go down, and your stock price, along with other pure-play miners, has trailed that of Bitcoin and some of the other smaller miners that have made more direct overtures towards AI. What do you think of those things?
Thiel: A couple of things. At the beginning of this year you started to see the shift from bitcoin miners to kind of MicroStrategy and the ETFs because of the coming halving. Any miner was suddenly going to produce half as much bitcoin, and so their costs were going to go up. Essentially, they’d be less profitable, so it’s only logical to think that their market cap would be impacted. At the same time, MicroStrategy was out there buying $2 billion worth of bitcoin, and that became very attractive for people to do.
The investment community–I’ll go away from the retail side for a minute–the hedge funds, the volatility investors, love bitcoin mining because you get a lot more alpha than you did just by investing in bitcoin. It’s the same reason Warren Buffett invested in gold miners versus gold because he likes the fact that when gold price runs up, gold miners’ cost to extract doesn’t necessarily go up.
The same generally holds with bitcoin miners. There’s a six to nine-month lag as bitcoin price runs up before the cost to mine really starts jumping because the cost of mine is driven by either energy costs or global hash rate and the difficulty rate going up. And so if you have to go buy new capacity, buy new miners and deploy them, there’s a six to nine month lag in that process. So you have this period of high profit when bitcoin price moves quickly. When bitcoin price does not move quickly, it moves sideways, buying spot bitcoin is much smarter and that also gives you an insight into our HODL strategy and why we go and buy bitcoin in the market sometimes. But MicroStrategy and Marathon and the other big miners have options that there’s an options market. And so the ball players love it. They can do all sorts of very creative stuff and if you look at the stock prices, they tend to move pretty much in a pattern where they move up at the beginning of the week and down at the back half of the week. And that all has to do with selling calls and then buying back the calls after you short the stock. You can see very specific trading patterns for all the companies. The smaller guys were generally ignored previously, and now people have said the big guys are well-valued, and the smaller guys should be valued similarly on a per exhash basis. I think they’re giving the small guys too much credit. Look at Cipher, they’ve signed up 350 megawatts of new stuff, but it’s not going to come online for 12 to 18 months, yet the stock price goes up. So that’s retail chasing that. Our investor makeup interestingly has changed this year in that in the first quarter we had about 39% institutional and now we’re up closer to 60%. The institutional investors are seeing the value in our story because of how we position the transition to AI. I do not think that a miner like TerraWulf or Hut8 can go and build and operate a hosting center for AI very effectively. It is not the same thing as running a bitcoin mining site.
Forbes: What is the investor makeup? Are these buy and hold investors?
Thiel: Our single biggest shareholder today is BlackRock. They hold 13% to 14% of our shares, in both their managed and unmanaged funds. Same thing with Fidelity. You have basically BlackRock, Fidelity, Vanguard, and then some long list of others after that.
But I think what we’re certainly seeing when we’re out talking to investors is the people who are starting to talk to us with interest are the ones who are looking at what we’re doing around AI infrastructure and how that opportunity is very different from traditional bitcoin mining. Traditional bitcoin mining is a zero-sum game, and every four years, you get slapped in the face with a price increase effectively. As you start looking at digital infrastructure in the AI space, you now have the benefit of providing technology or services to these hyperscalers. The area we believe is most interesting is AI at the edge, which is all about inference. The inference marketplace is still two to three years from growing a lot because people are still busy building these LLMs [large language models].
If you read Sam Altman’s white paper that he published recently, basically, his argument is you have to invest heavily in LLMs so that you have essentially something for inference to query. The way I put it is LLMs are the brain and inference is how you make money off of AI. Inference is the digital twin that is sitting in a factory, running a factory, and leveraging a model that sits up in the cloud, and the model is updated daily, sent back down to the edge, and it operates at the edge. There will be tens of thousands, if not hundreds of thousands, of these modular data centers that do inference at the edges spread all over the place.
Forbes: Let’s talk more about the HODL strategy. You’ve hinted at some very strategic decisions that affect how you leverage your balance sheet when updating equipment and making investments versus procuring bitcoin on the spot market. Can you break that down a little more?
Thiel: When you are investing in bitcoin mining, you are betting on the fact that you will be able to produce bitcoin at a lower cost than you could acquire it in the open marketplace. So if I’m going to spend $100 million dollars and buy bitcoin, I get that bitcoin tomorrow, and then I get the benefit of its price appreciation over the hold period, and let’s just say I’m going to hold it for 36 months. If you assume it will go up 20%-plus a year, you have that price appreciation in the bitcoin, and your basis is today’s price. If you are going to go into the mining business, your bet is I’m going to invest $100 million dollars in infrastructure that will likely go live in about 12 months on average. It could be shorter if you’re buying an existing site that’s running and you’re just going to replace the miners. And so you need to over the following 24 months, mine not just as much bitcoin as you acquired where you bought the bitcoin today, but enough bitcoin so you’re also covering your cost of power, etc., such that at the end of the period of time you end up with more dollar value of bitcoin than you did just by buying it today.
What that requires is a belief in what the price of bitcoin is going to do in the long term. And so when we have access to capital, there are times when we will choose to go into the market and buy bitcoin versus invest it in adding more mining capacity because sometimes there’s a lag in how long it takes for us to deploy the capital when we’re adding mining capacity.
Forbes: You recently met with Donald Trump. Can you share your experiences or takeaways from that round table?
Thiel: Former President Trump has been very well briefed and understands the value proposition that bitcoin provides to the U.S. as a country and the opportunity that bitcoin mining provides for the U.S. to be the dominant player in processing bitcoin transactions and securing the bitcoin blockchain. It’s a little bit like the gold reserves the U.S. holds. The U.S. holds 40% of the gold reserves of the world because in the event there was ever an attack on the U.S. dollar, you have the gold reserves that you can use. The U.S. also happens to have custody of the majority of other countries’ gold. As you know, Russia got on the wrong side of that with the sanctions program.
Many nations have figured out that with the weaponization of the dollar, they need to be able to hold assets in something other than the dollar. So they’re now evaluating holding things in amongst other things, bitcoin as well as gold and other assets. And not that they will hold all their reserves in bitcoin, but they may decide to hold somewhere between 1% to 10%, depending on the country and their risk profile. I don’t think the Saudis, if you were to talk to them, would say that they consider the rupee, the ruble or some of these other currencies as good reserve assets, and so they’re still going to put their faith in the dollar. But they want to have a hedge, and they want to have something that provides a risk-adjusted return potentially better than just placing their money in dollar assets because a lot of the countries today outside of the U.S. are being very cautious about their buying of Treasury bills, for example. The sixth largest buyer of Treasury bills today is Tether [according to Investopedia]. It’s not a country. It’s a stablecoin.
So what you’re starting to see is people who want to have dollar-denominated stablecoins end up being the biggest holders of Treasury bills. Then you have to look at bonds and who wants to hold those? The Chinese certainly aren’t buying them anymore nor are a lot of other countries. So if a sovereign is going to hold bitcoin, then the sovereign has to be guaranteed that they can transact it. Because if all bitcoin mining happens in the U.S. or the U.S. has a significant amount of bitcoin mining going on, then the U.S., in theory, could sanction a country by forcing U.S. bitcoin miners through legal regulation to not mine certain transactions because they’re on OFAC sanction lists, for example. That could restrict somebody from transacting where, let’s just say you had 60% of mining in the U.S., well, six out of 10 blocks, meaning somebody would only be able to transact in four out of any 10 blocks. That happens, and if you’ve got enough countries in the world to go against one country, you could potentially almost block them out. So sovereigns who are going to hold bitcoin have a need to have mining in their country. Bhutan is a great example. Bhutan holds $1 billion of bitcoin. They also have a lot of mining in their country.
El Salvador now has stated that they no longer need third-party debt because they mine and hold bitcoin. So the U.S. should hold bitcoin just like they hold gold, and the U.S. should ensure that a certain percentage of the global hashrate comes from the U.S. so that all the bitcoin held in the U.S. by ETFs and these other funds can be transacted if need be.
Forbes: Trump has said that he would like to see all remaining bitcoin mined in the U.S. That might be just politicking more than a serious policy statement, but I’m interested in your thoughts, given what you just mentioned.
Thiel: I’m not going to put words in his mouth, but I think he views it as strategic importance for the U.S. to be a dominant player in bitcoin mining. And I think sometimes when he says dominant, he’ll say, we have to be the end all. He wants that in the way of bitcoin mining and in the U.S. holding bitcoin reserves just like the gold aspect. You’ve seen him recently talk about if the U.S. were to hold a lot of bitcoin it could potentially appreciate to a point that it could pay for part of the debt or even all the debt. Representative Paulina Luna, Robert F. Kennedy, Jr. and others are all believers in these programs. We think it’s certainly a very wise choice because there will only ever be 21 million bitcoin. So it’s just a question of who will own them over what period of time.
Being shareholders in and having cofounded the only U.S. ASIC manufacturer for bitcoin mining, we’re certainly big believers in bitcoin miners made in the USA. You could argue that Chinese companies make 75% of the bitcoin miners made in the world today. I’m not saying that they’re trying to do anything bad, but it does provide them an ability to essentially somewhat control the supply and the pricing of bitcoin miners.
Forbes: That concentration can be dangerous in anything.
Thiel: Exactly. So that’s why we’re big advocates. Regardless of who wins the presidential election, there may be opportunities to do something about that from a tariff perspective.
Forbes: Have you had any interactions with the Harris campaign?
Thiel: We have talked with the head of the DNC. We have talked with all sorts of senators and the Harris team does not want to engage with us. They have engaged with the AI industry. My belief is that, unfortunately, Senator Warren and her anti-crypto army have a fairly strong lock on some of the policy decisions that the Harris team is evaluating. Certainly, President Biden wanted to apply a tax to bitcoin mining as part of his budget. Part of this is they just don’t understand the environmental benefit that bitcoin mining brings and the fact that it does not add additional fossil fuel requirements, it actually helps subsidize the renewable energy side of the world. And if they understood that, I think they would be willing to listen. But there are many forces rallied against us for political reasons on that side of the aisle that I wish didn’t exist.
Forbes: When did you start imprinting “made in the USA” in the blocks that you mine and why did you do that?
Thiel: It was around the 4th of July [actual date was July 31]. It’s a way for people to know if that block is mined in the U.S. or if it was mined offshore. Offshore entities control the majority of bitcoin pools. We’re one of the few bitcoin pools that is not. We believe some people prefer transacting and putting their transactions into blocks that are mined in the U.S.
Forbes: Are you ever worried that there could be a certain premium on bitcoins mined in the U.S. versus those elsewhere? Could that potentially impact the fungibility of any of these assets?
Thiel: If you filter on OFAC, that doesn’t impact the fungibility because you’re not filtering bitcoins; you’re filtering wallets. How to look at it is somebody’s bank account is on an OFAC list. The dollars in the bank account are on the list. It’s the account, so the account can’t transact. It’s not that the bitcoin it holds is tainted. You may remember back in 2021, when we launched our pool, it originally had OFAC filtering turned on.
Forbes: You’ve mentioned in the past that mining of Kaspa is a way to diversify. How much Kaspa do you have and are you continuing to mine it?
Thiel: We continue to mine kaspa. We’ve grown our capacity a handful of times since we first announced that we were doing it. It’s dissimilar from bitcoin in that it’s a much smaller market cap and pool of liquidity. And so you can’t just add exahashes and grow. You have to be very careful because there isn’t a lot of liquidity compared to bitcoin. We mine enough so that we are not an overwhelming amount of the hashrate on the Kaspa network, but it is more profitable to mine than bitcoin on a per kilowatt hour basis. So if I invest a kilowatt of energy towards bitcoin, my profit using S21s could be anywhere from 5 to 8 cents depending on my cost of energy per kilowatt hour. In the case of kaspa, it’s anywhere from 35 cents to a dollar. I can take kaspa, I can sell it and I can buy bitcoin or hold it. Kaspa has had some very strong price appreciation lately. So we’ve had a HODL strategy around some of the kaspa.
Forbes: I published a list of what we called crypto billion dollar zombies in the magazine in March. Kaspa was mentioned as one of them. We saw the mining and we know it’s similar to bitcoin, but it’s hard to ascertain real discernible use for it beyond just people mining it and then sending it back and forth. Any more thoughts about Kaspa you’d like to share?
Thiel: Kaspa as a technology is very interesting. It is a webchain versus a blockchain. It’s web of blocks versus a linear line of blocks, and Kaspa has a very fast transaction speed. It also has a very high emissions decline. So it goes from emitting a lot of kaspa at the front end and then that subsidy drops very rapidly down. And so kaspa mining will likely not be a long-term business, but as long as we can mine it profitably and there’s a good liquid market for it, then we just view it as a great source of revenues.
Forbes: Is there anything I didn’t ask you that’s at the top of your mind you’d like to share?
Thiel: One thing that could be of interest to readers is that one of the primary drivers for the hashrate growth has been the upgrade cycle. If you take 30 watt per terrahash machines and replace them with 15 watt per terrahash machines, you just double your hashrate. Now that the upgrade cycle is finished, we’re entering a phase where people have to add actual capacity.
I think you’re still going to see global hashrate inching up, but you won’t see the step function change that you saw. In 2023 hashrate doubled, then we had a halving in early 2024, and then you had all this hashrate come online [later]. We‘re pushing 700 exahash today. I think if anything, what you’re going to see is bitcoin is going to continue to push sideways. We’re seeing an institutionalization of bitcoin–ETFs have come on board; you now can do options on ETFs; you now have Morgan Stanley’s 15,000 wealth advisors soliciting for bitcoin. We’re still in the early days of this and you’re now going to have hopefully the repeal or the reversal or the number of exemptions to SAB121, and you’re going to see banks being able to custody bitcoin in addition to BNY Mellon, which means you can then start doing leverage, etc. You’re going to build a lot of liquidity. With all those, what should be demand drivers? You still see institutional buyers sniping to buy bitcoin. They’re sitting on the sidelines. They wait for a dip and they buy a dip. They are not in the market hurrying up, forcing the price up. So until there is a catalyst that drives the price of bitcoin up and maybe the election is the catalyst, I don’t think we’re going to see bitcoin price move up dramatically.
Forbes: Thank you.