HomenieuwsAI nieuwsArthur Hayes: 'AI has taken the marginal dollar'

Arthur Hayes: ‘AI has taken the marginal dollar’

Cryptofocus bron in google
Bybit - sponsored
Ontvang nu een welkomstbonus bij Bybit, maar liefst 30 USDC!

A month ago Arthur Hayes said the Bitcoin bottom was in. Then he sold his altcoins and walked it back, which made him the easiest target on crypto Twitter. So I asked him to walk me through the flip on New Era Finance. His answer is uncomfortable for anyone waiting on the old playbook: the money crypto needs went somewhere else, and it is not coming back until the AI trade cracks. Meanwhile the one position he does want is the coin this industry loves to hate.

What you need to know
1Hayes counts about 1.5 trillion dollars of AI-related debt since ChatGPT launched, against roughly 1.5 trillion dollars of new M2: "AI has taken the marginal dollar. There's no cash to go to crypto."
2He sold winners like NEAR because the asymmetry is gone. Now he wants a hated large cap: "I think Ethereum might fit the bill," the only major coin still under its 2021 high.
3The Clarity Act changes nothing in his view. Traders route around rules, and DeFi that needs a US law to work "has failed at its job."
4My read: Hayes talks his book and admits it freely. But his liquidity math explains this market better than any story about Strategy selling 32 Bitcoin.

AI took the marginal dollar

My first real question was the one every reader here has been asking since the fourth quarter: why did crypto simply stop mattering? Hayes did not hesitate. “Because AI has taken all the money.” Then he laid out the plumbing behind that sentence, and it is worth reading twice.

“Commercialization of ChatGPT, November 30th, 2022 until the present moment, we have about one and a half trillion dollars of various forms of debt issued for AI-related stuff. And about one and a half trillion dollars of M2 was created. So essentially AI has taken the marginal dollar. And so there’s no cash to go to crypto.”

Arthur Hayes, New Era Finance

Every new dollar the system printed found a data center to fund. The timing fits: Bitcoin peaked just as the capital spending wave ramped, and the 2026 spend is nearly double last year’s. In his words: “If it’s not AI, it doesn’t matter.” Investors are not being irrational, he argues. They are doing exactly what you would do. “You have a unit of fiat, you know it’s being debased. And you go, okay, what’s the fastest horse? Right now it’s AI.”

The same rotation explains the mood gap between Wall Street and everyone else. The top few percent own the assets that AI is inflating. The rest watch SK Hynix do 10x in a year while their paycheck buys less. “If they ask me what I think about the economy, I think it’s terrible,” Hayes said, playing the part of the average voter. Sentiment is not broken. It is accurate.

You can watch the full conversation here, and this one rewards the full 54 minutes:

This interview is an episode of my podcast New Era Finance. Subscribe there if you want these conversations the moment they drop.

Why Arthur Hayes is buying the most hated coin in crypto

Hayes was early and loud on NEAR and Hyperliquid, months before either was fashionable. He has now harvested those trades, and his explanation of the original NEAR entry is a small masterclass. The coin sat near a dollar, down 99.99 percent from its high, fully diluted, with no wall of venture capital waiting to dump on him. “It was a perfect setup for doing a 3x in a few months. All you needed was a little bit of narrative.”

Today those names are consensus, and consensus is precisely what he refuses to pay for. So what is he hunting now? His filter is three words long.

“Things that people hate. Is there a large market cap crypto or altcoin that’s been hated, been forgotten, and just needs a little bit of narrative juice? That’s what I’m looking for.”

Arthur Hayes

He also wants to be able to ignore it. “I’m more concerned about capital preservation than capital accumulation. I want to be able to take a large position and turn my phone off.” That rules out every small cap that requires living in a Discord server. It leaves one obvious candidate, and he named it without me pushing.

“I want a large-cap crypto that has not done very well, that you might feel left out. I think Ethereum might fit the bill. It’s probably the only major coin that has not taken out its 2021 all-time high.”

Arthur Hayes

His case is deliberately unromantic. Ethereum has been hated “for years, for good reason.” The L2 economics are, in his words, “kind of fugazi.” But it is also “the most secure decentralized computer,” the chain where the developer talent still lives, battle-tested enough that he can size the position without fearing it goes to zero. That combination, hated plus indestructible, is exactly his kind of asymmetry. Full transparency from my side: I still hold NEAR in my public portfolio and said so on the show. He sold his. Both can be right, because we are not running the same clock.

Lees ook:  Bitcoin-koersweerstand: breekt de uptrend van hogere lows?

The Clarity Act? “Irrelevant”

Half of crypto is positioning for the Clarity Act vote as the moment institutions finally flood in. I asked Hayes how the vote fits his Ethereum thesis and got the fastest answer of the hour: “Irrelevant. Doesn’t matter.” Then he went further than I expected.

“If DeFi and ETH need a US regulator to pass some act, then we have failed at our jobs as building decentralized systems. I’ve worked at a bank. I have friends at hedge funds, and I can tell you right now, nobody cares. If it’s making money, they will find a way to trade it.”

Arthur Hayes

Compliance departments will contort whatever entity structure is needed to touch a winning trade, he argues, and if a US law is your investment case, “just go trade some stocks. It’s way easier.” What actually matters to him is the capital formation model. Tokens should reward participation, not venture money. “Participation is what is needed, not money.” The current game of low float, high FDV listings that get dumped on retail is why nobody is excited anymore. His fix is blunt: “If you generate money, you have to give it back to the token holders. If you don’t do it, your tokens are irrelevant.” And his model citizen is the one everyone already knows: “Be like Hyperliquid and you’ll be successful.”

Stop blaming Saylor

Last week I published my conversation with Michael Saylor, recorded days after the market decided that Strategy selling 32 Bitcoin was the reason for the entire correction. I put that narrative to Hayes. He dismantled it in one line: “He’s a coincident indicator.”

The real story, again, is where the printed money went. Trillions in new fiat were created these past years, yet Bitcoin peaked in October 2025 around 125,000 dollars, not even a double over the 2022 high of 69,000. “The money creation went to a different sector. AI sucked it all up, and it will suck it all up until the AI bubble bursts, whenever that happens.” Saylor’s sales are the excuse, not the cause.

“If Saylor was selling 32 Bitcoin and Bitcoin was at 200,000, we’d have a different response.”

Arthur Hayes

I also tested the four-year cycle on him. The consensus trade right now says Bitcoin bottoms around 40,000 in October, which is exactly why I doubt it happens; consensus trades rarely get paid. Hayes refuses the calendar entirely. He is a quantity-of-money investor, and his one rule for cycles applies to every chart being shared on X today: “Everything rhymes, but isn’t the same.”

Lees ook:  Waarom dalen goud en bitcoin nu samen zo hard?

The next crisis gets the Big Print

If AI ate crypto’s bull-market/" data-term-title="Bull Market" role="button" tabindex="0">bull market, the obvious question is what ends the AI trade. Hayes sees two cracks in the foundation. The first is hardware: the entire buildout is financed on long amortization schedules, while “a GPU pretty much is obsolete within 2 years.” The second is pricing: the world is funding American AI at San Francisco prices while Chinese open-source models deliver a good-enough product at a fraction of the cost. In every industry China has entered, he notes, the China price eventually set the global standard. DeepSeek at one percent of the price is not a meme to him. It is the margin call waiting at the end of the boom.

“You can’t change Moore’s law by printed money. I think the next crisis is going to be met with printed money, and then it’s going to get even worse. That’s the big print.”

Arthur Hayes

His script for that moment is specific. The first move is down, everything sold together, Bitcoin included, exactly like March 2020. Then comes yield curve control and money creation at a scale that makes the COVID response look modest. “Being liquid for that moment will make your returns for the next two decades. I want to get that trade perfectly.”

“You want to be the Rothschilds”

Where most guests reach for 2021 comparisons, Hayes reached for 1873. Measured as a share of GDP, he expects the AI buildout to rival the railroads, and the railroad era produced crashes so violent that the 1907 panic gave birth to the Federal Reserve. “AI will be as big or bigger than the railroad buildout. You need to study 1873.” Same theme, different characters, different technology.

“You want to be JP Morgan. You want to be the Rothschilds. You want to be liquid when everyone else is illiquid, and that’s when you make your money.”

Arthur Hayes

One right crisis pays for a career of mistakes. John Paulson never repeated subprime and never needed to. Whoever bought Bitcoin at 3,500 dollars in March 2020 can be wrong for a decade and still be up. Hayes wants that trade again, and he is happy to be wrong on everything until it arrives: “I can get everything wrong from now until then. I don’t care. As long as I am playing that convexity right at the turn.”

That is what stayed with me after this hour. The loudest flip-flopper in crypto is, underneath, making one long argument for patience with dry powder. He changes his mind in public every week and does not care what you think about it, because none of those calls are the trade that matters to him. There is a lesson in that for everyone glued to the four-hour chart, and I include myself.

This is the second New Era Finance conversation I am publishing here within a week, after Michael Saylor. That is the level of guest this podcast brings every week, so do yourself a favor: subscribe on YouTube, or listen on Spotify and Apple Podcasts. Hayes writes his own essays at his Substack. Follow me on X via @CryptoMichNL for daily updates. See you at the next episode. Ciao!

Partnerdeal

Bybit - Welkomstbonus 30 USDC!

Ontvang nu een welkomstbonus bij Bybit, maar liefst 30 USDC!.

Lees hier de voorwaarden
Nu naar Bybit
Voorkeur cryptofocus google nieuws
Michael van de Poppe
Michael van de Poppe
Michaël van de Poppe is econoom, trader en ondernemer uit Nederland. Hij studeerde economie en werkte als trader aan de Amsterdam Stock Exchange. In februari 2017 maakte hij kennis met Bitcoin en altcoins. Begin 2018 stapte hij over naar fulltime trading in crypto. In 2018 richtte hij MN Trading op, een educatieve crypto-community. In november 2023 breidde hij dit uit naar drie takken. MN Capital investeert in en adviseert early-stage Web3-projecten. MN Academy biedt cursussen aan voor beginners en gevorderde traders. In 2025 evolueerde zijn consultancytak naar MN Fund, een beleggingsfonds geregistreerd bij de Nederlandse financiële autoriteiten, gericht op volatiliteitshandel en OTC-transacties. Michaël schrijft over Bitcoin, altcoins, macro-economie, technische analyse en portfolio management. Hij volgt de markt via cyclusanalyse, on-chain data en fundamentele ontwikkelingen binnen blockchain en de AI-economie. Van 2019 tot 2021 schreef hij voor Cointelegraph. Hij is host van de New Era Finance Podcast, waar hij wekelijks spreekt met toonaangevende experts over crypto, finance en de toekomst van de economie. Zijn analyses bereiken meer dan 815.000 volgers op X. Zijn YouTube-kanaal telt meer dan 160.000 abonnees. Bij CryptoFocus.nl schrijft hij over markten, zijn eigen portfolio en investeringsstrategie.
Lees ook