Arthur Hayes: Why He’s in Cash & Gold (& What Turns Him Bullish) | Power, Printing and Price Action

Prices are down a lot and this is the best time to start deploying capital. We don't give a what inflation is. You start with the wrong expectations, you're going to get the wrong result.

Thanks for having me.

I'm waiting for the big print or some sort of market event that generates the impetus for central bankers, especially at the Federal Reserve, to restart the liquidity engine, print a lot more money. I don't think that we've created enough base or broad money over the last, let's call it, 12 months and that's why the price of Bitcoin and a lot of other cryptos have struggled. And so, while I am very, very long Bitcoin and a lot of other crypto assets, with the little bit of fiat cash that I have sloshing around in our accounts, the question is, would I be adding more risk to my crypto pile today? And the answer is no because I'm still waiting for the all clear that it's time to make the money printer go burr and crypto prices to accelerate to the upside.

So, at a broad level, US authorities are very concerned about market volatility as it relates to US Treasuries. And so, right now there's obviously a war going on in the Middle East. And another very deflationary aspect which is this agentic economy and the displacement of knowledge workers, especially in economies like the United States. And so, either one of these situations could produce an outcome where the central bankers have to print money. Either something happens in the war, whether it's foreign governments selling their Treasuries or a need to print money to expand the presence of the US military in the Middle East, that could trigger a large printing effort by the Federal Reserve and the US Treasury. And then the agentic side, if the continuation of knowledge workers losing their jobs at high-tech companies and service providers continues, which it has accelerated since the launch of the new Claude models in late 2025, we're going to see an impact on the ability of banks and other finance companies to receive the service of their debts that they have booked as assets on their balance sheet. What are these consumers going to do who lose their high-paying tech jobs or lawyers or accountants or what have you if they no longer have the ability to service their car loan, their credit card payments, or their mortgage? This is going to impact the banking system and I think that the Federal Reserve is going to have to print a lot of money to make sure that the fiat financial system survives the transition between the old style human-centric economy to this dual economy between humans and agents.

Well, if it doesn't happen, then we have the situation that we're in today. Bitcoin's fallen from 126,000 to around 70,000 where it is today. So, call it a 50% or 40% decline. That's the situation that happens when you don't have enough broad money created. There's a lot of other actors within the economy who are hoovering up capital, especially if you think about AI hyperscalers and the capex requirements that they have given the dwindling free cash flow piles at some of these large companies. So, there's a very big suck of cash out of the system. If there's not enough liquidity created, then Bitcoin is the early mover. It tells us that there aren't enough fiat units around in the world and that's why we've seen a decline of about 50% in Bitcoin and obviously a lot more in some other cryptos. What are we looking for? We're looking for that announcement of a money printing scheme, of an emergency meeting of a central bank, and, you know, throwing in the towel saying we're just going to print money. We don't give a what inflation is. That's what we're looking for. It'll be very clear when it happens. And obviously, again, it hasn't happened and that's why we're down 50%.

So, obviously we recently hit 60,000. We bounced off of that level. I think that we're going to test that level again. And if we want to have a real resumption of the bull market, at least exiting from this down and sideways movement, we need to hold 60,000 again. I have no idea whether that's going to happen or not, but I think that's the key level, at least in the short term. You know, how much money they're going to print? I have no idea. It'll be a lot of money. It needs to be in the trillions and that can reinvigorate the crypto market. And if the first stop is to get through 100,000 and then the next stop is through 126,000. And then we'll see what the quantum of fiat money created is and that'll determine how high we go.

There has been not enough money created. Central banks have not printed enough money. There hasn't been a situation that has caused them to panic and hit the burr button. Obviously there's been many things that could have happened between now and then. They haven't happened and that's why we're at the levels that we're at today. So, that's it. If the money's not printed, then we don't go higher. It's pretty much that simple.

So, obviously there's been a lot of noise about AI agents or Claudes. And really sort of kicked off in late 2025 with the newest Claude bot model. And this has allowed, essentially, especially on the development side, people to have an orchestrated team of agents to accomplish a particular task. I have an interesting anecdote of an OG crypto friend. He's been in the game as long as I have. He runs a gaming company out in Asia. We sat down for a meal. We were talking about how AI affects our businesses. And he said, you know, he's a trained engineer, been so for over 20 years. And he said he tried out the new model late last year and he was super impressed about the quality of code he was able to generate very, very quickly. And so, he brought together his most senior engineers at an offsite a few weeks ago and he tasked them, okay, let's try to automate our entire workflow using AI agents. Let's make sure that they can code overnight, that code review happens, testing happens all autonomously. And when we wake up in the morning, we as senior engineers review the code and see what we want to merge into production. And what he found was that his rockstar engineers were extremely productive with these new agentic tools. In one case, somebody was able to ship their entire six-month roadmap in four days all by himself. The upshot from that offsite was that he needed to completely change how his business developed code and developed things. And so, he's firing half his workforce. And we've seen a similar trend, whether it's at financial services companies. Workday only announced a very small reduction in headcount, but they framed it as AI optimization. I think it was 3 to 5%. Facebook's firing workers. You have the Block and a lot of other tech companies who are saying our agentic workflow in terms of how we ship code has improved to such a degree that we're able to do with much fewer, very, very highly paid engineers. And when these engineers lose their jobs, especially in the United States, there's a thing called unemployment insurance. Now, if you're making a couple hundred thousand dollars a year and you go on unemployment insurance, you don't necessarily get that same amount of money from the government. It's a state-by-state program and I think about the maximum in most states is between 25 to 40,000 US dollars. So, you're going from making 250, 300, 400, 500,000 dollars down to making 40,000 dollars. And so, what happens in that you save your money, you stop spending, and then you go to default on your debts if you have them. Credit card, car loan, mortgage payments. And so, in the first case, we see the highest propensity spenders in the most levered economy in the world to consumer spending, the United States, do not have the ability to, number one, consume at the rate that they were consuming. So, that puts a downward spiral in all consumer focused companies. And then you can't service the high levels of debt that you carry because the average American carries something like 20 or 30,000 dollars of debt across these various verticals. And this debt is an asset on the bank's balance sheet. And if these people can't service their debts, then how is the bank going to make sure that the asset is worth what they say it's worth? Obviously it's not in this deflationary agentic rollout. And the point is that not everyone has to lose their job. I'm only talking about the 50th percentile and below workers getting culled. This isn't a mass doomsday event. But even 10 to 20% job losses in an advanced economy like the United States generates a huge banking crisis and I wrote about this and I put the number at something like 15% losses to banks and that essentially could cause a financial crisis which the Fed will bail out with printed money. At least that's the theory.

I've been a hater of real world assets and sort of this tokenization game for a very long time. I think at the end of the day what does a financial services company want to do? They have a very expensive back and middle office that are prone to make lots of errors. If you use a fully on chain or permissionless whatever you want to call it system, you can essentially eliminate your back and middle office and purely focus on how do you generate revenue? And so I think that's the first step what these companies really want to do. Okay, we have these things called custodians whether they're shifting around paper PDFs or some sort of digital certificate of a stock, we can eliminate that. We can make settlements instantaneous instead of passing around PDFs between different teams. So there's all these things you can do when you move on chain that have nothing to do with sort of trading crypto assets or a lot of these protocols that people are shilling out to retail investors. So I think that's the first thing that a financial services company does is go on chain whether that's permissioned or permissionless, fire their entire back and middle office, become a lot more profitable and nimble. I don't think that necessarily translates into gains to most of crypto. It doesn't really matter. We've basically proven that you can essentially run a 24/7 highly performant financial system, all in a decentralized fashion over the internet. And now financial services companies have seen this experiment proved out for 15 years and are ready to implement it in their own workflows and essentially eliminate a lot of cost.

I think it does absolutely nothing for crypto. I think this is really about a cost optimization. Will banks allow people to trade crypto? Of course, because there's fee revenue in it. Does a bank want a decentralized network where they lose control over their workflows and their clients? No, they want a walled garden with low cost and high ability to charge fees. So I don't really see how this tokenization thing is really going to impact a lot of the high value cryptos that we have today. Maybe it will take away trading volume because now it's easier to trade stocks 24/7 with a bank. And you get a better trading experience and they're all well maybe I don't want to trade on a perp dex or whatever. I mean, I don't believe that's going to be the case, but this is the situation what happens when a bank makes themselves a lot more competitive to a 24/7 digitally native customer. But again, I don't really see how this is going to positively or negatively impact Bitcoin. It's a completely separate thing. It has nothing to do with whether or not a bank is going to buy crypto or put it on their balance sheet or what have you.

Well, I mean, I think we've seen over the past 24 months majority of new projects the tokens been down only. The obviously the bright star in that is Hyperliquid. And why was Hyperliquid super successful whereas a lot of projects were not? The first thing is actually have product market fit. They have a customer that's paying them real money in terms of fees for a real product, trading. And they take that money and they pass it back to token holders in the form of buybacks. And there's no big VC overhang. Now obviously I run a venture capital firm in crypto so I'm partly guilty of this, but at the end of the day what do we do as venture capital firms or family offices? We have a pool of capital and we provide it early to a project in return for tokens at a very, very cheap price. Then the project launches and we're locked for a period of time which makes sense to make us align with the long-term viability of the project, but eventually we need to get a return on our money because we were the first money in. We provided the riskiest capital to help this team build this project. Now the problem with that is, we've made our, call it 5 to 100x on our money and we're incentivized to sell it because we have to return it to our investors because that's what we do. We're asset managers. And so there's nothing wrong with what venture capital firms and family offices are doing with respect to providing capital early, getting a token allocation, selling that token when they're allowed to after the unlock. But from a retail perspective, because none of these projects actually solicited the support of the retail investor at the beginning, the only way that the retail investor can get involved is much later in the project once it's proven that it has some sort of traction to launch which means a higher price. But at that particular point in time, the early capital wants to earn a return and is selling in size because they've got something like a 10 to 30% of the supply of this particular token. And so this is why these projects all mathematically most of them will have to go down because they have not generated enough retail interest in their product or service to escape the drag of VC selling. And so essentially what it means is we have to change the model if we want these prices to go up which is go to retail people early saying, 'Hey, you want to take some risk and get in this project early? Subscribe to an ICO.' I love the ICOs. It was fun back in the day. Subscribe to this ICO, most likely we're going to fail, but if we're really successful like a Hyperliquid example, then this is a super valuable project. You've gotten in early, you make a great return and there's not this overhang of VC investors who are selling because they have to because they're fiduciaries of other people's money.

I think as I said before most altcoins have no product market fit. They have no customers. If they do have customers and making revenues they're not passing that revenue back on to token holders in the form of staking emissions or buybacks. The only coin that I'm excited about is Hyperliquid because I think it has again real clients paying real money and that real money makes its hands into me and us the token holders. Nobody else is doing that in the magnitude that they're doing it in and so that's why it's been very, very resilient in this sort of 50 to 60% decline in Bitcoin and other broader crypto prices and especially during the very volatile period of this US and Iran war.

I love prediction markets and I know this is probably not a weird but aggressive statement. I think insider trading should be legalized. At the end of the day a free market incorporates all relevant information. Insiders have relevant information. If they put it into the market, then we know what's actually going to happen. So the best thing about Polymarket and Kalshi are that political insiders are trading on this information on these decentralized markets and we're seeing trending markets before major geopolitical events happen. I know some of the media have been skewering certain political families and certain politicians or insiders for trading on Polymarket. I think we should encourage this more because at the end of the day as a people, we want to know as quickly as possible what is actually going to happen. If we have markets that are allowing everyone to trade on the information that they have on Polymarket or Kalshi or now HIP-4 on Hyperliquid which will be launching soon. And this is a great evolution in the transparency of governance around the world. So I love prediction markets. I love the insider trading that they're generating and I think they should be encouraged.