I’m out of Markets Until… | Arthur Hayes

Great. How are you doing, man? I had this whole interview with you planned out and then I woke up this morning and there was another Substack and it was like, wow, we're going down a big rabbit hole today.

All right, let's go.

Yeah. That and I went back and I challenged my process that got me to this conclusion of thinking that there was abundant liquidity and Bitcoin should rise. And obviously I was wrong about that. From October of last year to the present, Bitcoin is down 50% yet the Nasdaq is up. There's been money creation especially in the United States. So why has Bitcoin not done so well? So I said, okay, let me figure out where did the money go? I think it went to AI. So I went on Perplexity and I asked it can you give me an estimate of how much debt was issued in various forms to AI or AI adjacent companies from the start of ChatGPT commercialization November 30th 2022. The headline number is about one and a half trillion dollars is the estimate versus one and a half trillion dollars of global US dollar M2 increase. Again, very rough numbers, I'm not trying to be 100% correct. It's more about the directionality and getting the magnitudes right. So at a high level, all the money went to AI and therefore Bitcoin didn't get any love from all this printed money. Now obviously that's not the case because Bitcoin rallied from the FTX lows in November 2022 to the present, up 7x. But the reason why it was able to rally even though AI is sucking all this capital is the intensity at which AI demanded capital in dollar and other terms increased dramatically starting in 2025. So from 2025 to 2026, the present, about $1.3 trillion of debt has been issued out of $1.5 trillion. So this is why Bitcoin had the ability to rise as AI was just revving up in terms of its suction of all fiat capital around the world. And so if that's the premise, then as long as the AI bubble's going, that's what you're investing in because you can look, Korea, Taiwan, Japan, America, pick your stock that's in the AI stable. Some of these things are up like 50x in 12 months. You're not getting those sorts of returns in crypto right now. And so you have to be in AI. If you're saying, 'Hey, I've got a unit of fiat. I'm worried about inflation. I know there's more fiat coming down the pipe. Well, I should be playing the AI game because that's the game that's generating the returns necessary to escape the trap.' And that's where it's happening.

Maybe could be. I mean, I guess focus on liquidity. So, I mean, that's a perfectly valid thought process.

Correct.

So if there are two issues that animate the American public, both Republicans and Democrats. Number one is affordability. People are worried that things are getting more expensive. And of course they rightfully blame the Republicans even more so after the Republicans started this war that nobody voted on. And so there's not really much Trump can do about that. Even if the war ended yesterday, this oil and commodity inflation has been building through the supply chain for months. You can't just stop it. And so inflation is baked in. There's not really much you can do about that situation. Gas prices are going to be elevated, food prices are elevated due to the war and other things that have been happening through the first year and a bit of the reign of Trump too. So I almost think you got to kitchen sink the affordability issue. So if you can't really do anything about that issue, what's the next thing that people care about? And I think that they probably care about this more because yes if a price of good goes up 10% but you have a job, okay it's bad but you still have a job. If AI takes your job, you have no income and shit's going up in price. So, what are you afraid of more? The AI taking your job or the inflation. And so both rich and poor, red and blue Americans are scared shitless about AI, especially because every day the media is running these sound bites from Elon, from Sam Altman, from pick your AI tech bro, who has to come out and say AI is going to be the most transformational technology that ever was because they want you to buy their overpriced company or they want the government to give them more subsidies or they want the government to say this is a national security so we must support you against China so they have to paint this doom and gloom picture that AI is coming for everybody but that works against them politically because now everyone's like oh shit AI is coming for me it's coming for my job I'm not going to make any more income and it could happen as quickly as next week and I was making 250 now I'm on unemployment. And so what's happening around America people are saying hey I don't want that data that's built there that's causing inflation. They're becoming receptive to these messages from Democrats like Elizabeth Warren and Bernie Sanders and AOC talking about fairness or redistribution. Why is it that there's only 15 people who make all the money in AI plus their stockholders yet they took all of our data to make these things? Where's yours, American person, red or blue? Where's your stuff? And I think that's a very seductive message. And I think this is the only issue that Trump could lean into and reverse what is surely to be an absolute shellacking in the House by the Republicans and the Democrats. There's a path to victory if he goes anti-AI and I think that is Trump as someone who has no ideology. He's all about winning. If all he has to do is say a few words and change the direction of people's attention towards I am all about responsible AI. I am all about redistributing this productivity miracle to all Americans. Just like he gave the biggest handout since the New Deal during COVID. No, this is not something a Republican president usually does is hand poor people money. He handed everybody money. Rich, poor, doesn't matter. No means testing. Biggest handout since the New Deal was a Republican president. So don't tell me that Trump can't say, 'Okay, if I need to go anti-AI for six months to get reelected so that my team stays on top.' It's just words. Remember, he doesn't have to do anything he said. He's a politician. People expect it anyways. But at least you have to come out and at least try to meet the people where they're at. And so I think this is the only way he can win.

If the war, I mean, if he stopped the war right now, it's just too late.

I think it was a miscalculation, but again, it almost doesn't really matter at this point. Who cares? It's not as if he can turn back the clock. He did what he did. Like it or not, now we're here. What are you going to do about it? And so now there's high prices. The high prices will definitely be reflected in the inflation numbers up until the elections. The only way to fix the elections is to go hey pro responsible AI. That will swing the vote and then win the election and then go make good with your tech brothers down the line. Because at the end of the day, having a job is more important than whether or not inflation happens. Because if you don't have a job, you're way more impacted than if inflation is 5, 10, 15%, whatever it is. So, I think people are more afraid of AI and it doesn't help that every day the tech bro has to sell you that AI is going to transform everything because they need to sell you something that's ridiculous multiples.

I think we're in a bubble. Are we in the early stages, late stages? I don't know. I do know that we have a company coming to market. SpaceX at 100 times sales. OpenAI is not a profitable company. Anthropic obviously is a profitable company and we're basing all these assumptions on the fact that token usage to the price per token will come down and the usage will go up. That's literally the story. So the only problem with that is if we're in the middle of this war which is taking 20 to 30% of supply of hydrocarbons, energy, which is used to run the turbines to create the electricity to create the intelligence. So theoretically you could have the price per token going up and so what happens the usage goes down. And so if your model is predicated on some ridiculous expansion of usage but the price goes up because energy goes up that doesn't work. Even if the earnings are good and we're not saying that these companies are making actual money. Nvidia has an amazing business, amazing earnings. Google, all they have amazing earnings, but we're seeing, well, I want to pay even more for those earnings and I think the token price is going to decline and usage is going to go up like Jevons paradox says in such an exponential way that I just have to be long all these stocks. But tell me if you still believe that $200 oil. Can you still believe that story of $200 oil? I don't know. Can you still believe that story? If the commander-in-chief of the country that's most responsible for the explosion of these AI companies says, 'I want to tax you 50, 60% because it's not fair.' Great. You have great earnings. It's going to the government. Do you want to buy that stock?

And we could be in a short pause and it could be a very painful correction or not. But I'm worried about what's my positioning. What's my positioning between now and November, knowing what I know about what could happen with the oil price and the politics of AI? Do I want to be long this at this price or is this a situation where I should take some shares off the table, wait for these things to resolve? If I'm wrong, fine. AI is still going. I'll buy back in at a higher price. Who cares? To your point, these are companies making great earnings. Token usage is exponentially increasing. So what if I miss a few X's? If I'm in the middle of the bull market, I'll just buy back in and I'll be fine. But the last thing you want to do is invest and get owned 50, 60% in your portfolio, people's forward outlook fundamentally changes because they're afraid of taxes or higher energy costs not withstanding what their belief on token usage. That to me is worse. So I can't obviously, I don't have a crystal ball. I have to solve for one of these things. And now I'm going to solve for capital preservation. Live to find another day. Let some of this stuff pass. If November comes around and nothing that I said happens, okay, cool. I'll be back on the bus.

Why the Senate? I don't think any people are saying that the Republicans are going to lose the Senate. I think the reason why he cares about losing the House is more personal than anything else. If the Democrats have the House, what are they going to do? They're going to subpoena him, his family, his lieutenants until the death. If they learn anything that was unearthed and they come into power in '28, guess what the DOJ is going to be very busy doing? Going after Trump and his lieutenants using information gleaned by public testimony in front of Congress. So, I think that if he wants to relive 2020 to 2024, don't lose the House of Representatives or, you know, get the people where they're at, be the populist that you are, and go and call Elon and Jensen and Sam into the room and look guys, don't cry. I'm coming at you guys for four months. Don't worry. I got your back. Now, go out there and don't be a little bitch and just take it like a man. Take this ass beating like a man because the Republicans got to win and then I'll come back or we can get back to partying after November. That's what I would do if I was Trump.

Yeah. Because, okay, I put it into crypto terms. I have to tell a lot of our founders, I'm like, 'Guys, you need to paint the chart green. If you launch it at this ridiculously high valuation, the law of large numbers makes it very difficult to outperform that.' So irrespective of whether or not SpaceX is a good business or not, offering it at the seventh largest company in human history, does that make it very easy for you to make money for the secondary investor if he launched it at 100 billion and said okay cool and did the same thing, much easier to make that chart green versus at 1.8 trillion or whatever it is. If he goes up 50%, he's now the fifth largest company or whatever the number is, larger than Amazon, one of the best run companies in human history. And this is a space company with an unproven thing saying, 'Okay, we're going to build these data centers in space and overcome all these issues with physics and orbital mechanics, all these things. We're going to overcome all that and we should already be given the benefit of the doubt and the seventh largest company in human history.' I don't know. I think that's a bit ridiculous. Are you worried about if SpaceX was a hundred billion in offering? I'd say, 'Oh, this is going to rip. It's going to go up 2x on the first day. People are going to be really good about AI.' Like, look, this thing went up 200% on day one. I need to continue buying SK Hynix and Samsung and whatever the fuck I was buying. But like, Facebook comes out and does 10% or 20% over the IPO price, people like, 'Ah, I'm not in this for 10 to 20% based on what I believe the future is going to be. I need some fucking explosions.' And it's hard to generate explosions at $1.8 trillion. That's the point. It's regardless of the multiples. It's hard to move a number that big. And everyone's looking at this IPO as a bellwether on whether or not the animal spirits are still in AI or not.

I mean this would be very interesting to see the perverse impact of passive investing. I really don't think that's the biggest issue with it. It's the fact that if you take a look at the unlock schedule for SpaceX a massive amount of shares come unlocked at the same time Anthropic and OpenAI are scheduled to do their IPO in September. And you have Google and Facebook has just announced stock sales. So they're saying, 'Okay, I can't issue the amount of debt that I want to issue.' Or maybe I'm still going to do that as well, but I'm also going to tap the equity markets. So I'm no longer buying back my stock. Now I'm issuing stock. And if Facebook and Google are going to do it, then why shouldn't every other hyperscaler start issuing stock as well if they get favorable terms? So everybody's just like, I need this capital. I'm borrowing it. I'm issuing that stock. All these different ways. And I just think it's a lot of supply to digest. Now if the market digested it in a way that we look at the future like oh great this is amazing we got this opportunity to buy these things at this price. Okay but financial history doesn't really point to that being the outcome.

Yeah, and I still don't think it's really going to happen until post the election. I really don't see the Fed having the political cover to do that until after the election, especially if you take a look at the two-year rate is trading what 50, 60 basis points over effective Fed funds telling the Fed you should be hiking rates. Whether or not Powell does it or not, we'll see. But that's what they should be doing.

Maybe Trump is like hey I'm going to try to win some ground back on this affordability thing, you know, hike. Maybe could be right. It's one way to attempt to solve the problem. I don't know. I think all things are on the table here. There's an election to win. Politicians are very flexible with what they will do to win an election and I think people think that Trump has all these ideological constraints which are not really constraints. You just got to win.

I mean, it goes for the narrative. I am doing something about inflation. If his guy who's going to cut rates, rates cuts, inflation starts accelerating, bond yields spike, that's front page news. Certain type of voter cares about those sorts of things. How is that Trump addressing this problem? That is the number one issue of Americans. That is the reason why people think his party is going to get their asses kicked.

I mean I don't... what was his... so he had this whole, I don't know when he said this. I'm paraphrasing here. You know the war inflation is we can look through the war inflation because it'll be over at some point. We have this AI productivity miracle. It's going to generate less inflation, therefore we can lower rates. I mean, I think that's the argument that the people who believe Warsh should cut rates will use. Whether or not he believes it, I don't know.

Nothing. Then that comes down to the press conference. What does he say? I think he does nothing. He holds and it's either a hawkish hold. Either, hey, we see inflationary pressures building in the economy or he talks some mumbo jumbo about productivity blah blah blah. And people say, 'Oh, okay. That means that eventually we'll get rate cuts.' So it's a hold plus what does he say? I don't know. So, but I don't want to hope, but I don't want to be in the situation where I'm hoping he says something dovish because that's going to continue this AI thing given we have all these issues coming to market, all the supply and very high valuations. Whenever I think about that there's a lot of hope in my investment and I don't want to be hoping for things in my investment thesis. That is usually a recipe for disaster.

Basically, or let's see what rhetoric that Trump is going to put out to try to win this election and if my thesis is correct and what is the price of oil. If oil goes up to $150, $160 because this war thing continues to grind on and stocks of hydrocarbons have been drawn down below levels and there's restocking and all these things with the Jim and Groom that you listen to, the commodity experts, if that comes to pass, this is an interesting take. We have this AI thing, the most transformative technology in the world, is it immune to higher interest rates and higher energy prices? I say no, but maybe not.

I'll place your bet on that. There already is companies complaining that they're spending too much on it. There's Uber. There's a few news clippings basically saying...

I mean it's really expensive at the frontier model at the top end is very expensive. And so now you're seeing people say okay well did I get enough utility out of whatever the work is I was giving the best model at that price. If no then I'm going to use the cheaper model or I'm just going to reduce my spend. So there definitely is price consciousness amongst consumers and if you want to grow your ARR like Anthropic did, you best be selling them expensive shit. So I think this does affect earnings big time because you can't have a situation where now you've got to try to compete with DeepSeek at the lower end and you're spending however many hundreds of billions on capex. That's a recipe for disaster. If you've got to go try to compete with the Chinese, then why are we investing in these companies?

Yeah. And again, there's just too many ifs and I think the probability of hope not being a strategy is very very high in this particular juncture. And so, yeah, it's risk off for me and how I think about the world.

I've always been in the back of my mind of okay this is a great ride I need to get off when am I going to get off what is the intellectual framework that tells me it's time to get off the train as I want to, you know, I see this great number on the screen and I want to realize that in actual profit not just like roundtrip this shit. So I'm always thinking about what am I getting wrong right now? What am I not appreciating? And so, you know, I listen to a lot of different macro people around the horn and some consistent themes start coming up. Number one, this war continues to go on and the unfortunate thing is that if the price of oil stays at these levels, there's no incentive for Trump and the IRGC to come to a deal because the deal means actual concessions and both domestic populations, Americans and Iranians will be upset with whatever their leader comes back with. If Trump comes back with handing money to the Iranians and the strait is still closed or there's a toll or whatever, like what the fuck did you start this war for? It's worse than what it was before. If the IRGC comes back and did a deal with the great Satan, they're like, 'Why are you disrespecting the Republic? He just assassinated the leader and his whole family. What the fuck are you guys doing?' So, both of these sides and the only thing that gets them to come together is a high price of oil. But every time they look like they're about to sign a deal, the price of oil plummets. And so, you have this reflexive back and forth. And at the end of the day, we're four months in and the strait's still pretty much closed. And this inventory issue is slowly ticking away. And so day after day after day gets more acute, more acute, more acute. We get closer and closer to that X date, whatever that is. And so this problem becomes more top of mind of can my portfolio rally at $200 oil? Now, I don't think the oil is going to get to $200, but that's how I have to think about it at a high level. Okay, if I believe that there's a situation where oil could get to 200 because this reflexive back and forth between Trump and the IRGC continues in this fucked up way where nothing actually gets solved. I'm getting closer and closer to this situation where I believe my portfolio cannot do well. And then on the AI situation, you know, I've been listening to a few folks and they're like, it's the nominal amount of data center buildup is not important. It's the second derivative of it. It's the acceleration or deceleration. So AI capex accelerated, AI credit accelerated massively from 2024 to 2025 into 2026. But again, law of large numbers. Is AI capex going to go from one trillion to two trillion next year? No. It's going to go up from maybe a few hundred billion. That's the estimate. So that's a deceleration. Yes, it's still growing. Yes, the earnings are great and all that sort of stuff, but I'm paying for accelerating earnings. That's what's generating 50x on SanDisk and SK Hynix at 1 trillion, acceleration of earnings expectation that continues into the future. But the law of large numbers tells me that this will decelerate even if the earnings are strong then we get multiple compression. I'm not saying that AI is not profitable. All I'm saying is that we were paying you 50 times earnings and maybe we'd pay 40 times earnings...

And so that means that...

Got to get off the train.

Yeah. So initially one of my analysts put it in the chat saying you should take a look at this and I didn't. I took a look at it, I was like, 'Oh, I took a look at it, saw the price, there wasn't really much movement.' I was like, 'Oh, they fixed it. We're all good here.' Next morning, wake up, it's down 30%. Okay, obviously I got something wrong here in my mental thinking. I read it again. I'm like, 'Okay, well,' and then I reached out to some very smart people than me and I asked him a very simple question. Is there any way to formally prove that no one minted coins cryptographically, mathematically, etc.? And everyone I asked said no. Some people gave me a bunch of reasons why it didn't matter and da da da. And I was like, okay, well, the reason why I'm in Zcash is if fundamentally I believe in privacy. I believe that a certain percentage of my Bitcoin holding should be in this private money. It's very, very good. And yes, I'll make a lot of money going on paper going up, but that's not the point. The point is to have this privacy. And so if you're telling me that an AI agent in the hands of a very smart human was able to find and fix this bug and you cannot prove to me that there was no minting. Not that I believe that they did it. I 100% believe that there was nothing nefarious that happened. That's not the point. I demand the amount of money that I had in Zcash demands perfection against big tech, big government, and big AI. And without perfection, then what do you have? You're worried that you got something wrong that you wake up one morning and there's another bug and you're down 50%. So if Zcash was at a thousand I wouldn't buy it back in unless this issue is fixed with formal verification that if there's a bug detected you fix it and you make sure that no one used it in an improper way and it doesn't matter what the price of Zcash is until you have that I think it's something that I cannot invest in in the size that I would like to invest in. So in the new upgrade, which is the upgrade that makes Zcash quantum resistant, they actually I think the problem is that you've got the shielded pool and the unshielded pool and no one knows what's going on in the unshielded pool. But in the new upgrade, it's called the Tachyon upgrade. Effectively, what happens is you can audit the shielded pool as well. So that problem then becomes resolved. Now there's a little bit of time between now and the upgrade. The upgrade is I think scheduled for Q3.

Yeah. Okay. Q3, Q4 is that upgrade and then hopefully this problem is sorted and then that goes away. That's why I stayed in Zcash because I was like, 'Okay, they caught the bug. Chances are that no one minted. We'll never know. Well, we will know. We'll know when they move to the new pool.' And hopefully if we can get through that hurdle, there is a fix and the fix is a permanent fix and then we're back. We're back on the bike. But yeah, I mean...

There's another zero day bug. Game over. See you.

No, that was more just it's a liquidity thing, right? I'm taking profit on everything. I don't believe that the macro is favorable to my positioning and my belief about the future and therefore I should take my chips off the table and see if I'm right or wrong. But do it in a position of strength and a position of not actually giving a fuck what happens and it's a summer time. I'm going to be on vacation so I'm not really going to worry about this shit.

It's not going the right direction.

Worldcoin is a proxy for OpenAI. The better that SpaceX does, the better that people assume OpenAI will do and therefore the better that Worldcoin does.

Awesome. Thank you.