"The Fed Can't Print Moore's Law" - How the AI Crash Sends Bitcoin to $1M | Arthur Hayes

TRANSCRIPT

Arthur:
[0:00] You can't change the fact that a chip gets better every two years if you pump

Arthur:
[0:04] $10 trillion into the economy. So what's the first response going to be by the financial authorities to try to save the banks? They're going to say, oh, we just need to shovel fiat money in. And at that point, investors are going to say, I do not want to put any new capital, whether it's free or not, into AI because it does not meet its cost of capital. So therefore, this capital goes straight to crypto. And this is the implosion of the AI bubble and the follow-on money printing that's going to happen, especially in the United States, to try to stick shave the system from this gross capital misallocation, that's happened over the last, let's call it six to seven years by that point, is going to dwarf subprime and is going to take us to Bitcoin a million or whatever.

David:
[0:52] Bankless Nation, I'm here with Arthur Hayes. Arthur, welcome back to the podcast.

Arthur:
[0:56] Thanks for having me.

David:
[0:58] Arthur, I originally asked you to come on back in May because I saw you were getting bullish on some of the tokens that I was also bullish on Hype near Zcash. So I was like, sick, let's get Arthur on the show. We can be bullish together. We can talk about tokens. It's going to be great. And then you sold them. So now I don't have any tokens to talk to you about, about why you're bullish because you sold them all. What happened?

Arthur:
[1:18] So I'm always analyzing my thinking, especially on this AI narrative. And, you know, I felt that the risks outweighed the benefits of being long, especially in these things and time to take some profit. There wasn't the same asymmetry there in terms of the awareness of the potential, of things like near and hype. On the Zcash front, a bit of a different issue. The way I view things, if I think that there is a risk of like pretty much a complete wipeout in my capital due to some technical issue, I can't in good conscience keep my capital in that particular situation. Obviously, the Zcash team is fixing this issue, adding formal verification. When the new shielded pool launches, hopefully this thing will be fixed, and I can evaluate then and see if I want to come back in. And I might come back in at higher prices, but again, I've been doing this for a very long time, and I still have most of my money. So I'm more concerned about capital preservation than appreciation.

David:
[2:15] What was giving you the jitters? what's causing you to what did you see in the AI trade that caused you some jitters

Arthur:
[2:23] So obviously we have the SpaceX IPO it's done very well in the first two trading days.

Arthur:
[2:28] That's a low float, high FDV shit coin, in my opinion. And we know how those go when you start having unlocks. And we start having unlocks in September. We have Anthropic and Opening Eye coming to market as well. And also we have this Iran war energy situation. You know, if you believe the authorities, they say it's supposed to be fixed. And then I guess read some headlines this morning that Israel says we're going to keep bombing Lebanon and supposedly Iran cares about Lebanon. We don't know. So that's like a whole thing. We'll see if this thing actually materializes anything. But at the end of the day, we have this underlying bid for physical commodities that isn't going away. And so there could be a situation where we wake up in three, four months time and oil is at $120, but the strait is supposedly open. And what do you do then? Essentially, energy is the input for our entire civilization, including the AI, because AI is just transforming energy into intelligence the most efficient way possible. So, you know, all these things were weighing in my mind. Like, okay, I think the summer is going to be sort of a topping period for a lot of the markets. I've done really well on these calls. Let me just take the W and go chill. And, you know, if I need a buyback in other things and let me find some asymmetry in other places, right? Like I got into hype in the high 20s, low 30s. It's now whatever they got out at like 70s, right? That's a decent trade, right? I think to go from the 60 to the 150, my original call, not that it's going to be impossible.

Arthur:
[3:54] I think it will happen at some point, but it's going to be a harder slog than the first part of this run. And so, okay, as an investor, at least in how I view things, is I want to find those asymmetric trades where I can deploy a little amount of capital and get a big return, and then I can put the rest in sort of low-income, yielding, you know, bond-like structures or Bitcoin.

David:
[4:14] So how are you positioned right now? Are you in cash looking at opportunities, trying to find the opportunities? How are you positioned?

Arthur:
[4:22] So obviously very long Bitcoin, as always, perennially long. Bitcoin, I got a lot of cash.

David:
[4:26] When you sell these tokens, like sell hype and near and Zcash, which I imagine are shorter term positions than Bitcoin otherwise would be. Do you sell into Bitcoin?

Arthur:
[4:34] Depends. So I haven't this time. I'm just sitting in a lot of T-bills, just earning some yields. And, you know, if I'm able to find some asymmetric opportunities before I really check out for summer vacation, then I'll do that. Otherwise, you know, I'll just be compounding my rates at whatever, 3.5%. Like not exciting, obviously there's a big bull market going on, but thankfully on a decently large capital base, it doesn't really matter to me. I more care about my mental sanity. Right.

David:
[5:01] Okay, so if you look at the crypto markets, there are pockets of dispersion everywhere. You have some of the majors, Bitcoin, Ether, which, you know, Bitcoin is sitting at its 200-week moving average. Ether is like 30% below its 200-week moving average. You know, Mr. Market is coming to us today with a deal on these assets. Meanwhile, some of the other assets that we were just talking about, Zcash, VVV, Nier, Hype, are breaking through at all-time highs. You know, Hype especially is at all-time highs. And that's not really something that we've ever seen before in the crypto world. And so while there's like some, in some parts of crypto, there's a bear market. In some parts, there's applications and protocols that are breaking through

David:
[5:44] all time highs. Have you seen this before? And like, what do you make of this? What does this tell you about the state of the industry?

Arthur:
[5:49] Well, obviously, there was a great trade, conceptually, at least back in the day when I really went long ETH back in like, you know, when it was to trade $100 or $200, which was there was some chart that was going around about the value of all applications on top of Ethereum were worth more than Ethereum. I think it was either one, Ethereum is undervalued or two, the applications are overvalued. And we were in this sort of a, you know, pseudo bull market at the time. So, let me just go along ETH. So, I guess fucking back the truck up in ETH and 20x, right? Or 10x, whatever it was.

David:
[6:18] That was in 2020 to 2021?

Arthur:
[6:20] Yeah, around that time period.

David:
[6:22] That era? Yeah.

Arthur:
[6:23] That era. And so, yeah, I mean, speaking of ETH, I think ETH has one of the probably the cleanest setups for something that could really rally off of you know a very big, safe crypto safe market cap right like i don't it has a lot of issues but, ain't going to zero overnight at least i don't think so right actually i could but the risk of going to zero overnight is much less than hype or any of these other tokens not that i don't love hyper liquid or some of these other projects but, It has a certain Lindy effect and a certain largeness and stickiness that I could just deploy a lot of capital in the ETH and turn my phone off for a year and a half and be happy about that. I would not do that in the same notional with anything else in the ecosystem other than probably Bitcoin.

Arthur:
[7:10] So I think that's probably one of the best setups I see out there for a mega cap coin because ETH is trading well below its 2022 all-time high of 5,000. Whereas pretty much every other token of value that has existed at that time period, a major, at least broke through that. You know, Bitcoin did, Solana did, you know, pick your, you know, bigger cap dino coin and it got through those levels. Might not be trading there now, but at least it got there. ETH hasn't even gotten to that point yet. So I think there is, if ETH starts to move, people are really going to jump on that train because they so badly want ETH to just like, why isn't ETH at 15,000 or some level? And there's obviously structural reasons why it's not, but people would love to believe that it could get there.

David:
[7:56] What do you make of ETH trading like well below its 200-week moving average? Do you pay attention to the 200-week moving average that much?

Arthur:
[8:02] No. No, you don't. I mean, I don't really look at technicals too deeply. I'm more of a narrative gut feel kind of guy. I just look at the chart, right? And if you look at the troughs of the last five years, we're trading near those levels. I mean, yeah, it'd be great to get an ETH at $1,000, but like, okay, cool. You're great at Bitcoin at $5,000 too. Like, yeah, sure, sure.

David:
[8:20] Sure, sure, sure. Okay. So, well, I'll kind of present the question again. Bitcoin is above the 200-week moving average. ETH is 30% below it. Does that, do you like Ether as a better deal than Bitcoin in this moment?

Arthur:
[8:37] Yeah, so if you gave me a dollar fiat capital, and so would choose one, I'd choose Ether over Bitcoin, purely just on chart perspective.

David:
[8:44] And is that what you're doing?

Arthur:
[8:46] Is that what I'm doing? I'm evaluating it. We'll see.

David:
[8:49] Okay. Okay. So one of the things I want to ask you about is like, I know you're a big hydrocarbons guy. You're a big energy guy. We just got the peace deal in Iran. As you said, it was like, sure. You know, Israel can do its thing. But like WTI and a lot of the oil tickers are trading down at the lowest they've been since the start of the Iran war.

David:
[9:13] Is this durable? Well, how is this input into your trade and into your attitude towards the markets?

Arthur:
[9:18] I believe that this is a fake out and we're going to be uncomfortably, uncomfortable levels of hydrocarbon prices. Now, I'm not saying oil is going to be at $200, but it's going to get up to uncomfortable levels. And I basically subscribe to the theory that, you have all of this restocking effect. So let's say that this deal holds and the strait opens for, what do they say, like 30 or 60 days, whatever it is, and you have to pay your environmental fee, but it's not a toll, you know, of course. You know, love the environment.

Arthur:
[9:50] And these ships go through, volumes get back to some level close to where it was pre-war, but you have two issues, right? So the governments that spent down inventory to keep the price not going crazy, like the United States and a lot of other countries, have to rebuild. You have commercial inventories that have to rebuild. And then you have everybody else who is caught short or caught unprepared is going to be forced to restock their inventory to acceptable levels because let's say that Trump and Netanyahu decide that they want to bomb Iran again after the US elections in November, then you can't be a politician and go back to the people like, yes.

Arthur:
[10:30] The Americans bombed Iran and we didn't have enough oil supplies because we trusted that America would keep sea lanes open and all those things. Fine, fine, you can be wrong once. You can't come back to the same people and go, oh, well, they did it again 12 months later and we didn't learn our lesson. We still kept holding treasury bonds and big cap tech stocks. And for that reason, you guys can no longer get on a plane and go on vacation because we have no jet fuel. We have no diesel. We have none of these things because we refuse to sell them these financial assets and buy these real assets. I think that's over. I think countries are going to start looking more like Japan and the U.S. and China and having 100 plus days of imports ready of, you know, hydrocarbons and these refined products. And that additional demand is going to slowly push up prices to levels that we have, you know, maybe we get $320 on WTI oil in 6 to 12 months. So I think this is a great buying opportunity for companies.

Arthur:
[11:28] Exxon Mobil, you know, oil services funds, natural gas companies, Latin America energy plays that are within sort of the U.S. orbit. So I like these investments because if I'm correct, then you're getting them at great prices for something that if you believe the commodity math is almost, not a certainty, but higher probability outcome than the market is assigning it.

David:
[11:51] There are a couple of things that I think align with what I understand to with what you're saying. First we have 60 days to figure out what a lot of the nuances of this deal actually look like to be honest like some of the some of the deal is like we're just going to punt on the details because we both want a deal signed now to signal to the world that we're signing this deal but like what happens with nuclear material tbd what happens how does this straight up from use actually open tbd like what that environmental fee actually is and how it's enforced tbd so these are all like jittery surface area for the deal to fall through, for this to actually not actually manifest into a long-term truce. And we kind of just punted the problem into the future. So that's one problem. And we also have 60 days for that to figure out. So the strait isn't going to necessarily be open for that time.

David:
[12:41] The other time is that we've been drawing down the United States, China, Japan, as you said, have been drawing down on their oil reserves. And so they need to bring them back up. And that also takes months of buying. And so like, you know, oil is cheap now, but it's not going to immediately go away. We have the inflation problem to deal with. So inflation is here today. It just started showing up after the oil spiked up three months ago when the Iran war started. And so we have at least three months of inflation to actually show up in the market because of the higher oil prices. And again, this problem isn't going to be solved for another one to two or three months before this rate can actually open, if it does actually open, and all these countries do restock. And that's all that's going to happen if Trump doesn't decide to reverse course and bomb Iran again, which I actually think should be in consideration of any investor. We could just be at this thing again in one to two months. Did everything I say check out? Is that aligned with you?

Arthur:
[13:44] Yeah, and I think, think about it from the IRGC's perspective, right? Every time that Trump has done a deal with them, it's either been a ruse to strike them again, whether it's, you know, Trump or his attack dog, Netanyahu, and Israel. So, yes, there's this truce for 60 days, but you would be, you should be fired as a war general if you believe that somehow you're safe. That this means anything other than Trump wanted a deal to stick shave the markets, so that maybe Elon's deal goes well with SpaceX and you can float these mega IPOs without any issues, But, yeah, to your point, nothing's actually been solved. It's just like the tariff thing. There was initial salvo. The market did not receive it very well. You know, Trump had to backtrack. And then U.S. and China basically engaged in these, like, nothing burger extensions, extensions. And, you know, there's not been sort of a grand bargain, if you will, a sort of a status quo understanding of where things are. But I think something similar could happen with the Iran war. You know, they probably have some strikes between Iran and Israel and Lebanon and whatever. However, the deal is still nominally in place at prima facie, but underneath, the flows are nowhere near as much as they were, you know, February 26th.

David:
[15:00] Yeah. I do agree that high oil prices could be the simple answer to like, why be bearish? The world cannot sustain like $110, $120 oil prices for more than a few months before we already are seeing inflation kick in. And United States, we have it way easier than everyone else. And so like the bear case, the recession trigger is high oil prices. And so we know that there that is a possible future. So we have to be wary about that. Granted the stock market knows this and the stock market has been doing great it shrugged off the iran war so easily you know it got a little bit jittery but then the ai trade you know and the memory stonks all just went through the roof and so like one of the other reasons why you cited as like taking like a risk off position is because we have open ai we have anthropic spacex coming into the market at very high valuations low float high ftb shit coin as you call it

David:
[15:54] what do you what do you make of like the current health of the equity market? Like, is it too euphoric, too skitterish? What do you think?

Arthur:
[16:03] Well, the assumption that it's all AI, obviously, the equity market is literally just an AI story. And any second or third derivatives from this, you know, CapEx build out that is going to be probably bigger than the railroads in terms of the percentage of world GBDP spent on building this infrastructure, whether or not it's useful, we'll see. And that's why all these stocks are rallying. And that's why, you know, if you believe that this is bigger than the railroad build out, does oil at 100 or 150 really matter in the next one or two years? Is Google going to stop their CapEx build out because oil is higher? If they believe, you know, every single major U.S. Tech CEO has basically said, I'm OK with my company going bankrupt if I don't risk that I lose the AGI war or, you know, risk being not at the table for AI. If that is the prevailing consensus among these eight guys, that this is why they need to spend $100 plus billion a year on a CapEx, the oil prices are relevant.

David:
[17:00] The oil prices are irrelevant up to a certain point. True, up to a certain point. We saw $90, $100 oil, oils are relevant. $120? Not so sure.

Arthur:
[17:09] We'll see. Again, if you listen to the rhetoric about, oh, we're investing for an S-curve in technology, oh, this is going to be the biggest build-out since the railroads, and we're not there yet in terms of the amount of money spent as a percentage of GDP. So that would tell you that we're in the early part of the bubble. And given the sort of manic, utopic rhetoric out there about what AI can do and what it's going to mean for the world, I don't know if materially higher oil prices are going to dent that enthusiasm. We'll see it's all about does it do these hyperscalers scale back on their capex commitments because they believe that either A they can't charge a high enough price to earn a return or you know this this.

Arthur:
[18:00] They just think they're not going to make money, enough money. Or, you know, the market says the banks stop lending to them or the U.S. Government stops giving them preferential treatment. So, like, all these things are the variables. Oil does affect those. I think if I'm going to be wrong, it's that people don't care. $150 oil, cool, but AI is going to usher in utopia where none of us do any work anymore. So what do we care about oil? And Elon's building in space where there's infinite sun. Yeah, I mean, there's orbital mechanics that come into play there that you haven't really understood, but okay, maybe. So if this is what you believe, then does terrestrial prices of dead dinos really matter? And so I think that would be the pushback against my bearish thesis on AI.

David:
[18:44] Yeah. What are you looking at? What indicators are you looking at to kind of

David:
[18:48] measure how much time left we have in the AI trade, in the AI bubble? Is that how you think of it?

Arthur:
[18:56] Yeah, so I see that this is a bubble. We are either in the late stage or the early stage of a bubble, whatever, right? So when I put my hat on as a trader, I'm like, okay, I got to think about the exit from this thing. I fundamentally believe that humans are always over optimistic at the beginning of any technological buildout. I believe that this could be bigger than the railroads in terms of percentage built. Okay, what happened with the railroad build-up? You had two financial crises in sort of the mid to late 19th century, 1873 and 1893, basically predicated on credit driven to railroads and sort of an over-indexing about the return that these things could generate. So, you know, I think that we could be, this could be the first little test, maybe this summer, we'll see, and then, you know, you rally off of that and we're still believing the hype in terms of what AI can deliver. And, you know, I think, you know, you read Michael Burry and a lot of these guys about what's wrong with this situation. It's all these circular revenue deals. It's that, you know, Google invests in a lab or NVIDIA invests in the lab. The lab takes an investment and buys the chips.

Arthur:
[20:04] Google is underwriting the credit of the data center that is going to provide them the compute, right? So it's all just circular accounting. The depreciation is pushed off in the future, right? You're taking a chip that depreciates every two years, pretty much, which we've seen over time, but you're giving it a six-year life or five or six-year life. That doesn't make any sense to amortize a loan based on that. You're saying that this CapEx will continue indefinitely, even though you have Chinese competitors with models that are good enough charging 1% of the price. And so I think this, a lot of investors have sort of bought into the.

Arthur:
[20:45] The America of, you know, is it the baddest, biggest bad-ass country, best tech, dah, dah, dah, dah, dah. It's the same fallacy you thought about cars, right? Do you give a fuck if you were to drive a Ford or a BYD? Give me the BYD. It's cheaper, right? It's just as good, if not better. So I think people have bought into the brand of the USA, but they forget that like the brand value erodes when a customer is faced with a product that's good enough and 1% of the price, which is what, you know, a deep seek or Kimi model can, can deliver. And so I think these are some of the things that eventually the market will recognize. I don't know when that is. And then they'll say, okay, I don't believe that I should be paying this multiple for this company. Or the executives at these hyperscalers say, I don't believe I'm earning the return I need to earn. I need to scale back my CapEx projections. And that's when the whole edifice starts to break.

Arthur:
[21:39] And you can add on top of that anti-AI politics. TBD on whether or not that becomes a thing in this election and.

Arthur:
[21:47] I've been reading, you know, different things and some publications that I like to get a mood on what's going on in the U.S. Unfortunately for my thesis on sort of Trump turning on AI bros, I still think that the Democrats haven't leaned into this issue enough because they're just bought and paid. They're just as bought and paid for as Republicans from, you know, big energy companies, AI data center people and all that sort of stuff. But there's anger there. I mean, I don't live in the U.S. I'm just reading magazine articles, but you live there. You know, there is this pent up, I'm afraid of losing my job. Why is this data center coming into my town, making stuff more expensive, polluting my water? Fuck these guys. I don't own any of these stocks. I'm not up 50x in my portfolio. I'm struggling.

Arthur:
[22:32] Why should I essentially underwrite this expansion when I get nothing from this? Where's mine? And I think there's going to be a whole class of politicians who, that are going to tap into that anger. And that'll be the number one issue that Americans are voting on, you know, maybe in this midterm, maybe not, but definitely for the 2028 presidential election. And that should scare the shit out of investors. Yeah. Because you can take a look at what happened in the, you know, late 19th century, early 20th century with all of the action against sort of the industrial revolutionaries, the robber barons. And yeah, they all still were wealthy after the fact, right? John D. Rockefeller is still the richest man in the world after they split up his companies. But like, if you were just an ordinary secondary market holder of these shares, you know, you could have gotten battered and bruised as sort of capital and labor came to these sort of agreements about how we're going to police this new economic productivity that has been created.

David:
[23:27] Maybe I'm reading tea leaves here, but what I was thinking about while you were saying that is like, look at what happened to the crypto industry after we had our big run-up and bubble and like rise to fame in 2020 and 2021, like we were on top of the world, like crypto bros, you know, ran the world. The whole world thought crypto was cool for a brief moment. And then in 2022, interest rates came, the bubble popped.

David:
[23:52] And then the regulators came at crypto when we were weak, you know, when we were down, bad and out. Not while we were on top of the world in 2021, but Elizabeth Warren built her anti-crypto army in late 2022. And Gary Gensler came in 2023. And it was when we were kind of like, just like, you know, licking our wounds from the market is when the regulators came. And so maybe we could project forward into the AI world where you have a lot, a lot of You see Tom Shaughnessy's tweet about like one possible way the AI bubbles pops, which is a little bit what you were talking about, which is everyone shifts away from like this, the paid plan where you pay $50 a month and you get to type into chat as much as you want, which just doesn't appropriately price these things. But where you go is you go to the API model or you just pay per token and the paper token, you're going to be charged a lot more because the whole plan, a subscription based model is just not at all reflecting the actual costs of the company. So everyone is going to be, it's like a pay-per-use thing. And all of a sudden usage is going to dry up because it's actually way more expensive and all of this stuff is getting subsidized. And so you do have like a couple like dominoes getting lined up of like ways to pop the actual revenue metrics of these like super scalar AI companies followed by just like exactly what you're saying, just like AI politics, just like being the second of a one-two punch into the whole industry.

David:
[25:13] What do you think about that?

Arthur:
[25:14] Yeah, for sure. and it's going to take time. This is not an immediate thing.

David:
[25:19] Yeah, it's a two-year plus thing.

Arthur:
[25:21] And I don't think people really appreciate that the effect that China will have on this. It will commodify the token market just like it's commodified every other market. Like China builds a gigawatt of electricity at one-sixth the cost of the U.S. So they can essentially, when they want to, well, they're already doing it, right? You can use DeepSeek. You can use, you know, those other Alibaba, Tesla, whatever. Baidu models. And as long as you don't talk shit about the Chinese Communist Party, you can pretty much get all the same answers that you want, if not more.

David:
[25:53] Yeah, their tokens are going to be free compared to ours.

Arthur:
[25:55] So it's like, okay, I'm trying to find out the cheapest rental car. Do I need to use the ChatGBT, Claude, Sonnet, Opus, Grok, whatever the fuck, model that's 100x more expensive? Or do I just go, To the Chinese one, that's super, super cheap. Maybe they are practically free. And I get the same answer, right? I just care about rental cars, right? Most people aren't doing super duper secret, like I'm going to build the next missile, destroy the world kind of shit, right? Like they're just trying to optimize their lives. Same with businesses. Do you really care if you use the US or Chinese model? At least the Chinese model is open source. You could theoretically like put that open source model on your own bare metal server racks and run this. Or you can give all your data to Anthropoc, OpenAI, GenMI, Elon, whatever, right? I'm like, do you really want to do that? So I think it's an interesting thing and it's the same sort of mental fallacy that a lot of people felt with a lot of other things outside of luxury or.

Arthur:
[27:00] High-end luxury, which is the end consumer doesn't care about brands when the competing product is good enough and is a fraction of the cost. Like your brand value goes to zero. Do you care what your brand of car is if you're not driving a Lambo or a Ferrari? Fuck no. Get me from A to B. Tesla's fine. BYD's fine. Zeker's fine. I don't give a fuck what the car brand is. I want to be safe. Get me from A to B. Cheapest price possible, right? And I, but yeah, okay, fine. The Birkin bag, I'm going to go pay Hermes a bunch of fucking money. But apart from that, there is no brand value. So stop talking about like USAI if you're going to say it's 100X more expensive than the Chinese version and does the same stuff.

David:
[27:41] Right. Yeah, ultimately at the end of the day people are just going to want the Toyota Corolla models. And that's going to be supplied by the Chinese.

Arthur:
[27:48] Yeah, absolutely.

David:
[27:50] Yeah. I suppose like in this moment right now, like AI company equity value is up only. Like you can go trade Anthropic on Hyperliquid. It just goes up and they just raise it a higher and higher round. Everything related to AI just goes up. In addition to that, they're subsidizing the actual usage. And so the cost of a token isn't actually really the cost of a token. And that only really works when, you know, number goes up, when there's some free money like flying around. I'm sure being in the crypto, in the AI industry kind of feels like being in the crypto industry in 2021, like all money's fake and you just have to like build and execute as fast as possible. And so as soon as that kind of stops and the marginal consumer or consumers broadly start becoming price sensitive, then a lot of this effects will start to actually like show up and people will start to look at like fundamentals and reality rather than kind of the hype and the narrative. But we're not in that phase in the market right now. People aren't price sensitive to tokens. AI labs aren't even price sensitive to tokens.

David:
[28:49] And so everyone is in this kind of like euphoric state. How do you know how long we are in this state? Like what do you look for as an investor?

Arthur:
[28:56] I think the biggest tell is going to be this depreciation schedule, right? So a lot of these financing deals, like you got to care about the debt because usually it's the debt and the credit that's going to force the, you know, mark fiction to reality, right? And close that gap. And right now you have people underwriting amortizing gpu deals with a five to six year period when things are getting better in a two to three year time frame.

Arthur:
[29:23] So that is mathematically going to start really biting us in you know late 2027 2028 time frame if you think about the the pace of which capex and loans were issued 2025, to 2026 for the banner year so add two to three years under that, you really get into this 2028 timeframe, which is the critical period where either demand has so ramped up so high and, that, you know, that doesn't matter, or, disappoints and just the math of if I amortize an asset at six years, but it has a useful life of two, then all of a sudden, all those H100s and Blackwells that you underwrote are essentially worthless. And it's not as if, The Frontier Labs can train on anything bad because if it's all about, I need to have the best model and a one or two or 5% incremental improvement could lead to a compounded permanent moat of my competitors. I cannot use the NeoCloud with H100s and Blackwells. I need to have whatever the newest NVIDIA or Huawei chip is. And I have to do that.

Arthur:
[30:34] I cannot use these other chips. And I think that's when the road meets the road and that's when we mark to market on this. And we see, okay, where's the revenue? Where's the capital returns? Does this make sense? And I have a feeling that that'll be the test. And the greatest part for crypto is that if we do get an AI credit event, number one, it'll be bigger than 2008 because the whole world is on this dilution that AI is the most, the biggest technology ever. We're going to build more CapEx per percentage of GDP than we did with the railroads, much bigger than subprime. And the Fed can't print Moore's law. I don't care how much money that you throw at this thing. You can't change the fact that a chip gets better every two years if you pump $10 trillion into the economy. So what's the first response going to be by the financial authorities to try to save the banks? Oh, we just need to shovel fiat money in. And at that point, investors are going to say, I do not want to put any new capital, whether it's free or not, into AI because it does not meet its cost of capital. So therefore, this capital goes straight to crypto. And this is the implosion of the AI bubble and the follow on money printing that's going to happen, especially in the United States, to try to stick shave the system from this gross capital misallocation.

Arthur:
[31:56] That's happened over the last, let's call it six to seven years by that point. Is going to dwarf subprime and is going to take us to Bitcoin a million or whatever, right? This is the big print. If you get this, if this thesis is correct and you time this well, you'll never work again. This is the John Paulson subprime trade. This is the big short Michael Burry trade of this particular epoch. And that's kind of how I'm thinking about this. I don't know when this is going to happen, but I know I got to get this right. Because if you're able to be liquid at this situation and when this fiat money floods into this system, it can't go to the AI trade anymore because again, the Fed can't print Moore's Law. Can't make a fucking depreciation schedule change no matter how much money you print. And I think that's the best argument for why crypto is going to do so fucking well when AI completely implodes. And that's what I believe. I don't know when this is going to happen. And I'm going to continue updating my mental model, but that's kind of where I'm at right now.

David:
[32:57] Huh. I find this incredibly compelling. I love that. What I love about this is that it is so counter cyclical. It's like we have two opposing waves happening. You have the AI at the top and crypto at the bottom. And a lot of the timing is for the future escaping that we are doing right now. Like we don't know, we don't know, but we're future escaping. The timing of everything where if you give a one to two year time horizon,

David:
[33:19] that kind of aligns with the political fears that we might have in about two years. So your window for opportunity is between zero and two years for the AI bubble to pop. It's like it could happen tomorrow, it could happen two years from now.

Arthur:
[33:35] If you go back to Subprime, Subprime started in 2007 with the failure of those two hedge funds, right? We started to mark housing prices stopped rising in 2006. So you had the first cracks in 2007 and then the crisis in 2008. So, you know, and you had essentially a panicked bailout where... Bush was not elected as a Bush was in office. He rolled out TARP. Obama came in this, you know, supposedly more equitable Democrat, whatever, continued the bailout, right? Because he had no other ideas on how to solve this thing. And they did it anyways. Didn't put any beggars in jail, all this kind of stuff, right? So, like, if you think about the perfect, perfect storm for this is marked to reality in terms of underwriting GPUs at six years versus two. That chickens come home to roost starting in 2027, but probably will limp on until 2028. And then you have a presidential election, which could be a referendum on AI.

Arthur:
[34:37] You know, what has AI done for me lately? Okay, quit. I can book a travel. I don't need my assistant to book my travel anymore, but my water's dirty. And, you know, I didn't have enough money to buy Tesla, or sorry, SpaceX or Anthropic or, you know, SK Hynix or whatever the fucking stock you want to pick that's up. 20, 30, 40 X from 2022. I'm just some poor American person living paycheck to paycheck, you know, with dirty water and hearing a buzz outside my house because of this fucking data center. And I don't own any of this shit. So fuck these guys. Then, you know, I could be a Democrat. I could be a Republican. And there'll be some smooth talking politician who's going to tap into this anger. And again, I think the 2020 presidential election will be a referendum on AI. And it doesn't really matter who wins or loses. But just the fact that that is a referendum on AI, you as an investor are afraid that this, you know, guy or gal is going to be successful and then tax you to oblivion. And all of a sudden there goes your all your returns. Right. And so you should sell ahead of that event. So I think if this is going to happen, the perfect storm for a a big print yield curve control back the fucking truck up situation would be sometime in the fall of 2000 and 28. So, you know, Obviously, I don't want people to lose money, but at least for my portfolio, if this thing were to happen in the way that I see it, that would be my perfect storm of a situation.

David:
[36:05] Right. And the time to have sold your AI stonks would have been well before that because by the time a big print comes along, it's like AI stonks are already in trouble, correct?

Arthur:
[36:15] Yeah, because the S&P, what, like, so the TARP happened in whatever it is, September of 2008. S&P didn't bottom until March 2009. Right. And they were printing and printing and printing. QE was announced in December, printing and printing and printing. It didn't do shit.

David:
[36:31] You need to already be into Bitcoin, into Ether, well before this.

Arthur:
[36:37] Not necessarily, because again, everything's correlation one on the down draft, right? Investors are going to sell everything.

David:
[36:42] Oh, you're trying to ride the AI bubble into cash, let everything correlate to one on the down, and then buy the bottom? Is that a pseudo plan?

Arthur:
[36:51] That is the ideal plan. Oh my God, you're the fucking best trader ever. Right, right, right. You just fucking crushed it.

David:
[36:58] We'll see if you do that. Because while I enjoy this story, I am looking at the crypto, Bitcoin, and the blue chip prices, Bitcoin and Ether. They are strictly deals in this moment. They are off their highs. Again, Ether is 30% below the 200-week moving average.

Arthur:
[37:14] To your point, what catalyst brings them higher? If the incremental unit of fiat is going to fund CapEx are going to some bank that's going to lend to companies, where is the capital that's going to be available to push Bitcoin to 200,000 or ETH to turn to 15,000? They may rally a bit, but I think this is the reason why people are so disappointed. It's like, okay, yeah, Bitcoin took out its all-time high. It was from 69 to, 125. It doubled from the all-time high to all-time high. That's probably one of the most, shitty bull markets ever. It sucks.

David:
[37:51] It sucks. Yes. Why? No one was thrilled.

Arthur:
[37:54] Because AI took all the money. Yeah. We printed a lot of money globally. And what happened? All that capital was funneled into AI. The only thing that, investors want to allocate to is AI, whether that's on the debt or the equity side. And I don't see that changing if this AI story continues to be the only theme that you're investing in, especially if AI can survive higher oil prices. You're definitely going to be, oh, wow, AI survived. Oil's at $150 and the NASDAQ's at $40,000. Then energy doesn't matter. It's all about AI. NASDAQ, $100,000, right? And so I think, yeah, Bitcoin could rise because it's a shittier version of a tech stonk, but it's not going to be that explosive rally that you expect given the trillions of dollars and euros and yen and yuan that had been printed. And so that is why I think investors are still going to choose AI as the fastest horse to sort of exchange a infinitely debased unit of fiat for some financial asset that will outperform the debasement.

David:
[38:58] In order for crypto to kind of just have the vitality and energy just and the funness that we once had in the markets, do you think like AI needs to just stop sucking all the oxygen out of the room?

Arthur:
[39:11] I don't see why, I don't see how on a broad sense you can get Bitcoin at $500,000, with this AI story in play. Because again, okay, if Bitcoin went to $500,000, let's call that a 10x return, right? Was it Sandisk at 50x this year or something crazy? Yeah, something crazy. That's a high next, went up 10x, right? So you could find some second or third derivative down from some component that's.

Arthur:
[39:42] Necessary for data center buildup, put your entire net worth into that, and you could be 20x in six months. Why would you buy Bitcoin? I think that's how you have to think about this. It's always an opportunity cost. Again, it's all the same trade. Everyone knows that this unit of fiat will be worth less in the future. What can I buy now that's going to outperform that? and I want the fastest horse. AI is the fastest horse and it's proven itself to be the fastest horse, and therefore an investor with this you know marginal unit of fiat is going to continue to choose AI especially if SpaceX at 180 Anthropoc does well OpenAI does well, the oil price goes up and the rally in the party continues, it's only going to make it harder for some of these cryptos to do extremely well. They could still do okay, but then you're like, oh, I made 50% on my crypto, but this other tech stock was up 20x. I messed up.

David:
[40:41] I do remember doing some podcasts, end-of-year podcasts, in December or January, six months ago, and there were a lot of consensus takes at the time of like crypto is just not going to be that exciting because the AI trade is going to suck all the oxygen out of the room. And I remember doing a handful of podcasts like that. And that's exactly what happened. And like here we are and Bitcoin is down $40,000. Ether is down 50% from where we started this year. I think that math is correct. And the only tokens that have made all-time highs are like somewhere between $500 million and $2 billion market cap coins with the exception of Hyperliquid. And really the only activity has really been down market in like the one to two billion market cap range.

Arthur:
[41:25] Yeah, I mean, it makes sense.

David:
[41:26] Yeah, yeah. Let's talk about Hyperliquid. What's it like? Did you create perps or were you just the first person to implement them?

Arthur:
[41:35] So there was a paper written by Robert Schiller about a perpetual thing.

David:
[41:39] Right? That's from like 1993, right? That's like the early one?

Arthur:
[41:42] Something like that, right? Yeah. And we didn't actually cite that paper and I had never read it before, but when we came up with the idea, but we kind of arrived at the same conclusion. And initially it was, we took the Bitfinex Bitcoin and dollar rates and then created this never ending thing, but it wasn't responsible enough to the market. And so the real innovation that we invented was this self-correcting mechanism, the funding rate, how you look at the past X period, the premium rate discount, apply that going forward as the funding rate. That is what really made the perpetual swap work in a crypto, high-paced, highly leveraged environment because you had this very quick resetting of what the rate should be in the real market. And it's completely self-contained. You don't need to reference outside prices to arrive at this mechanism. So that is a new invention that myself and the team at BitMEX created starting in 2016. So I would say, yes, we invented the perpetual swap. This is not a Robert Shiller product.

David:
[42:36] Yeah. What's it like to be in 2026 and perps are just the narrative? What's it like to be the creator of perps? I love it.

Arthur:
[42:44] I mean, I think I can't, I love that like Terry Duffy and, you know, the Nisey guy are like freaking out about Hyperliquid and, you know, all these perps like go fuck yourself.

David:
[42:55] What does it mean to be freaking out?

Arthur:
[42:56] I'm going to stare. Oh, regulator, come Hyperliquid. They're taking all my money. Like get the fuck out of here. Like Jeff and his team of, you know, 11 engineers have completely whooped your ass. Like, good on them. I love it. And so I'm really proud of Jeff and what they've done and all the other entrepreneurs that have taken this perp thing into a new level. And now we have perps on SpaceX, perps on Anthropic, like perps on oil, perps on whatever. Like, it's great. I love it. I think it's great.

David:
[43:27] Can you paint the bull case for the perpetual as an instrument? Like, why will perps, I don't know if you believe this, but like, why will perps eat the world?

David:
[43:36] Why will like Wall Street eventually get perpified? Like, why are perps good?

Arthur:
[43:41] Because retail investors love them. At the end of the day, Wall Street is about extracting value from dumb retail to smart institutional investors. And so if the dumb retail says, I want a 24-7 product with 100x leverage on whatever financial price there is around the world, and I don't want to have to understand futures curves and all these sorts of things, the perpetual swap is a product for them.

Arthur:
[44:04] Combined with socialized loss mechanism, which is basically what we have invented in the crypto ecosystem. That's Hyperliquid, that's Binance, that's BitMEX, that's all of us. Right. We've done, you know, starting in 2014 up until the present, we sort of refined the system, everybody doing their part to sort of like make this the perfect product. So if all of a retail, if the marginal retail person says, OK, I want high leverage, trade 24-7. And the only way to have that happen is on a perpetual swap with socialized loss. I'm not going to open my Robinhood account and trade there or trade some leverage features contract on the CME. And without the dumb retail, then you just have Ken Griffin hitting against Don Wilson and that's not the game that they want to be in. They're going to figure it out. Okay, fine. I'm going to deal with whatever compliance complexities there is and the tech issues and the custody issues. I now need to trade on Hyperliquid or whatever the dominant perp decks is and they'll get there because the clients are there. The people that they want to make money off of have migrated and so that's why it's a structural trend And I think the problem that a lot of these traditional exchanges have.

Arthur:
[45:12] They don't understand why they're not going to be successful with the purpose that they're launching. Number one, a lot of them don't do 24-7. So it's automatic, like, what the fuck are you even thinking? This is a 24-7 market. Number two, you haven't attacked custody. I don't know, clearing. The clearing situation has to be socialized loss. That's the only way this is going to work for you as an exchange to offer the high leverage and the notionals that are trading. If you continue to cling to this guaranteed settlement clearinghouse model that's been around since, you know, probably, you know, 50 to 100 years, you will not be able to, from a risk perspective, get to the point where you can offer a compelling product to the average retail trader, Hyperliquid, Binance, everyone else will just be a better compelling product because they're able to offer the high leverage. Yes, you do have this issue where you could get your ADL or your profit cap and all that sort of stuff. And I know there's a lot of institutional investors who don't like that uncertainty about their profit payoff, but that's the trade-off to get the retail person excited about trading on this. And the fact that Triified doesn't understand that it doesn't get why these products are successful means they're going to launch all these perps, they'll trade a bit. But again, the marginal retail punter is not going to trade that. They want the high leverage. They're perfectly happy giving up some of their profit in adverse scenarios, and they'll continue to trade more and more on hyperliquid.

David:
[46:32] Can you talk about the socialized loss mechanism and why that's so fundamental to the perv? Why do you emphasize that so much?

Arthur:
[46:38] Because the exchange cannot lose money, right? You can't offer a product where the exchange is put at risk. And when you do guaranteed settlement, the exchange is at risk. Because you're saying at any price, at any velocity of getting to that price, I will make sure that the person in profit, even if all the other traders are bankrupt, gets paid. How do they get paid from my balance sheet? So to rectify that, we built all of these clearing organizations. And the clearing members put up bonds, and they got a little bit of a fee for this risk capital that they put up. And that worked fine in sort of a bounded timeframe market. And you have sort of a central banking government that can bail you out if things get really, really bad.

David:
[47:11] I see.

Arthur:
[47:12] Now, So in crypto, right, we have highly volatile assets, bearer assets, you can't print crypto, and you're not going to see your clients. And so if you have 100x leverage, you have to protect yourself of the exchange. The only way to do that is socialized loss. And so unless you put a socialized loss mechanism into your clearing situation or your margining, then you, from a risk perspective, cannot offer this product.

David:
[47:36] And this is the same mechanism that makes these things like a protocol rather than needing to have like illegal entities in KOIC because hyperliquid is not going to come and knock on your door saying, hey, you owe money. We're just going to ADL you instead. The same mechanism. Correct.

Arthur:
[47:50] You have to say whatever you put in as a max, you can lose. Your initial margin is a maximum loss.

David:
[47:56] And if you get ADL'd as a trader, all that really means is like you won as much as possible and there are no more chips to give you because you won them all. And so no one can really be like, are you really upset? Like, sorry, other people didn't take this out of your trade.

Arthur:
[48:09] Well, you're really upset if you didn't understand how the mechanism worked or maybe the implementation wasn't as advertised. But any sophisticated trader by this point, like ADL has been around since 2014. Right. Like you don't understand how this works.

David:
[48:21] You should want to get ADL'd in the sense that like that means you're you.

Arthur:
[48:24] You should never want to get ADL, but what I'm saying is you should understand how it works and have bottled this out. And if you lose money and you didn't know what was going on, then you deserve to lose the money.

David:
[48:34] Okay, there's Hyperliquid, there's Leiter, and then there's Coinbase and Robinhood as a centralized end. There's all of these different players, and the Polymarket and Kalshi are in this race. There are so many players racing for the U.S. perps market, and with a lot of the statements, we just had Mike Selig from the CFTC on the show. So there's a lot of just like positive perp statements from like coming out out of the CFTC. What I'm trying to get a grasp on is like how valuable the U.S.

David:
[49:01] Market is or is the perp inherently like an offshore thing? Like where do you think the perp thrives better, onshore or offshore?

Arthur:
[49:10] I mean, I'd say that probably most U.S. people will still trade perps. I mean, we'll see the incremental uplift on what any of these platforms get if they have a, you know, they can properly... Market these things. But I don't know. It's big. Obviously, the U.S. Retail trading market's a massive market, biggest in the world, or one of the biggest in the world. So it's a massive market, and the ability to advertise freely will be a great boon for whoever is allowed in that space.

David:
[49:35] Yeah. I think one of the questions I asked is like, what you said earlier was like, the perp is just fundamentally a retail instrument because it just works well with what retail traders like, and that's why everyone else is going to come and center around the perp because retails there. But like if we put the PERP inside the United States, there's going to be regulations on it. Like you're not going to get 100x leverage. Maybe you get 10 because that's just what it means to operate inside the United States. And so if retail really wants all the bells and whistles that make the PERP a good product for retail, don't they need to go offshore and therefore isn't the PERP a fundamentally offshore instrument? That's my thought process.

Arthur:
[50:09] Maybe, but at the end of the day if you have distribution and you can market, like people love marketing getting marketed to. And so that's something that an offshore platform can't do effectively. And so I think that is one thing. Okay, they might have lower leverage, but you're able to reach a different user that was not going to be finding Hyperliquid on the web, right? They need to see the ad on the bus station or the train or whatever and that kind of thing before they're going to onboard.

David:
[50:35] Do you think there'll be an offshore winner and an onshore winner?

Arthur:
[50:38] Sure, for sure. I mean, Hyperliquid's already going to be an offshore winner. Binance is already an offshore winner. you know well the onshore the problem with onshore is it's going to be highly commoditized right it's going to be you know 5-10 licenses given out or whatever it is and everyone's just going to blow a bunch of money in marketing and you know it's going to be very difficult to make money.

David:
[50:59] Do you think Hyperliquid flips Binance?

Arthur:
[51:01] Eventually. Eventually. It'll take time.

David:
[51:04] Why will it do that?

Arthur:
[51:06] Because it's a better product. And, you know, you don't have to deal with all the other things of operating an exchange. And at the bare minimum, you outsource security to the layer one, where Binance has to secure their environment, which is extremely expensive. And it's only to get more expensive. But the client doesn't understand that. They just see the rate. They just see the fee. Right. And that is the fundamental reason why decentralized exchanges are better is because a client pays for the security of the L1, but then pays again for the trading fee.

David:
[51:37] Do you trade on LIDR at all? The ZK L2 perp tax?

Arthur:
[51:41] No, I mean, I was an investor in it, or I think we invested in some of it. I don't trade on Hyperliquid either. I don't trade using leverage.

David:
[51:49] You invented the perp, and you don't use perps? Yes. Should we take a lesson from that? Are you spot only?

Arthur:
[52:00] Yes, spot only. If you're not into trenches, you know, living this shit day in, day out, alarms on your phone, don't use leverage.

David:
[52:09] Okay, so why are you a spot guy just because?

Arthur:
[52:12] This thing is already so fucking volatile. You don't need leverage to make good returns.

David:
[52:18] I find it so funny the guy who made perps doesn't use perps. But, okay, but you can lend to the perp market. You can be the, you can do like the automatic A&M.

Arthur:
[52:28] Yeah, Athena is essentially lending to, you know.

David:
[52:31] Do you like doing that as like a barbell strategy where you.

Arthur:
[52:34] I mean, when the rates are high enough, yeah. I love, you know, staking USD. Right now the rates aren't that high, so I don't do it.

David:
[52:39] Sure. Arthur, I got no more questions for you, my man. Thanks for coming on the show. I appreciate it. Thank you. I always learn a lot. Bank of the Station, you guys know the deal. Crypto is risky. You can lose what you put in, especially if you use a perp. don't use leverage. You heard it from Arthur Hayes. This is Frontier. It's not for everyone, but we're glad you're with us on the Bankless journey. Thanks a lot.