BitMEX co-founder Arthur Hayes appeared on Kyle Chasse crypto, hosted by Master Ventures CEO Kyle Chasse, and explained why he sold several of his largest crypto positions, including HYPE, NEAR, Worldcoin and Zcash. The original podcast title was Arthur Hayes: Bitcoin's Final Dump Before The Pump, while the Odaily headline summarized the core message as AI draining the market and Bitcoin facing a difficult path above $100,000 by year-end.
Hayes said he had just published a roughly 5,000-word article titled Reality Check, and that the podcast version compressed the same argument into a shorter discussion. His reason for leaving those trades was not a project-specific view on crypto. Instead, he laid out a macro chain that runs from oil prices and the Iran war to Donald Trump’s midterm election strategy, a turn against the AI industry, the end of the AI bubble, and the inability of crypto markets to isolate themselves from that shock.
Oil, the Strait of Hormuz and the midterm election problem
Hayes argued that Trump and the Iranian Revolutionary Guard need some form of agreement to end the conflict. In his framing, oil prices control how angry different groups become and therefore how much pressure each side feels to negotiate. Trump has to worry about domestic voters, who dislike high gasoline prices and energy-driven inflation. Iran faces pressure from China and other developing economies that need oil and commodities to move through the Strait of Hormuz.
When oil prices rise, Hayes said, everyone becomes more eager to talk; when prices fall, the incentive to reach a deal fades. He described this back-and-forth as a dynamic that has already continued for around three months, or for as long as the war itself has lasted. During that time, the world has been drawing down commercial and national reserves of oil and other hydrocarbons. Energy analysts may use different charts, but Hayes said their conclusion is the same: inventories were ample before the war, helping keep prices low, but the surplus is now being consumed at an accelerating pace.
His worst-case scenario is that no agreement is reached, and by October the Strait of Hormuz remains effectively blocked, with only 25% to 30% of shipping volume getting through. A less severe outcome, in his discussion, would be a deal in one or two months that restores shipping to some degree. Even then, countries would need to rebuild strategic stockpiles and would likely hold more than before, because they had just experienced what it means for their supply of essential goods to depend on Trump and a group of Iranian generals. That rebuilding would create additional demand for oil, natural gas, helium and other inputs needed to operate a modern economy.
Trump’s anti-AI pivot as Hayes’ central political risk
Hayes connected the energy argument to the 2026 U.S. midterm elections. He cited Polymarket odds showing Democrats at 82% to retake control of the House of Representatives. In his view, Republicans are losing the cost-of-living narrative because voters see inflation directly at gas stations and in everyday prices. He does not believe Trump can easily make people think inflation is not there, because voters see it each time they refuel their cars.
That leaves AI data centers as the political issue Hayes believes Trump can flip on. He described the Democratic message as one built around no more data centers, taxing AI giants and regulating AI. The fear is not limited to lower-income workers, he said, because wealthier workers also fear their jobs being replaced by AI. A party that channels that fear can combine two messages: inflation caused by a Republican war, and an AI construction boom backed by Republican politicians.
Hayes said Trump could adopt that message himself. In his imagined campaign line, Trump would call for more scrutiny of data centers, an AI national dividend and taxes on AI companies. Hayes stressed that what Trump says before November does not have to match what he does afterward. He views this as the only issue on which Trump can quickly reposition himself and try to present Republicans as defenders of Americans against AI, even if the same political side had helped finance the boom.
According to Hayes, Trump’s willingness to attack AI depends on oil prices, which in turn depend on the reflexive relationship between Trump and the Iranian Revolutionary Guard. The longer the war drags on, the more commodity pressure builds for later price increases, and the greater the incentive for Trump to strike at AI in order to protect Republican chances in the House. Hayes said taxation and regulation are the most damaging threats to the AI narrative. He pointed to South Korea, where a politician proposed a national AI tax and Cosby hit its daily limit down that day. Hayes said that if such language is publicly adopted by the ruling party, especially by Trump, the AI bubble would top during the months leading into the election and crypto would be pulled down with it.
Why Bitcoin lagged AI despite the liquidity thesis
Hayes also revisited a core assumption he has often used: that more liquidity should lift Bitcoin. He said the assumption has been too lazy in this cycle. Since ChatGPT was commercialized on November 30, 2022, Bitcoin has risen, but Nvidia and other AI-related stocks have risen far more. Bitcoin topped last October at $125,000. Hayes said his model showed trillions of dollars in liquidity being created, yet Bitcoin did not reach $500,000 or $1 million. That forced him to ask where the liquidity actually went.
His answer is that AI consumed it first. Hayes said he usually does not use M2 because he considers it too crude, but used it as a broad example: since ChatGPT, U.S. M2 has increased by at least $1.5 trillion. He then asked Perplexity AI how much debt had been issued to AI and AI-related companies. The estimate he cited was about $1.5 trillion, with $1.3 trillion concentrated in 2025 and 2026. In his interpretation, the AI boom began in late 2022, but the capital-market debt pump became heavily back-loaded and only reached full force recently.
That timeline explains Bitcoin’s path in Hayes’ framework. From 2022 through mid-2025, Bitcoin had more room to benefit from new liquidity because AI had not yet absorbed as much of it. Reverse repo declines and other factors were favorable. But AI capital expenditure and lending accelerated sharply in 2025 and especially in 2026, which aligns with the period in which Bitcoin struggled, after topping last October and then falling 50% to 60%.
Hayes said that if all liquidity is going into AI and the trend has not stopped, an AI bubble correction would not automatically send investors rushing into Bitcoin. Instead, they would sell AI, Bitcoin and everything else. In a bubble break, he said, all asset correlations go to 1. Only after the dust settles do selected assets begin to outperform. For that reason, if he expects a major correction in the AI complex over the next six months because of oil prices and U.S. politics, he does not see Bitcoin or other crypto assets escaping the first leg down.
SpaceX, Anthropic and OpenAI as a liquidity test
Kyle Chasse then raised another concern: the S&P 500 is rising while most stocks are falling, and the index is being carried by a small number of technology names. He also pointed to upcoming IPOs from OpenAI, Anthropic and SpaceX, saying they could add more than $4 trillion in new market capitalization to equities. Hayes said such IPOs are difficult to satisfy because investors do not merely expect them to trade normally. They expect an IPO to surge 50% or produce some extraordinary gain that proves the market still believes in AI and in that star company’s future.
Hayes focused on SpaceX. He said its issuance valuation is around $1.8 trillion, which would make it the seventh-largest company in the world. For SpaceX to rise another 50%, it would have to become larger than Amazon. Hayes said that if one reads its S-1, the company is valued at close to 100 times sales. In his view, that is an absurd setup for a company that would already be the seventh-largest in the world while still not having proven the key thesis he believes the market wants to buy.
He used the language of crypto markets to describe the structure: a low-float, high fully diluted valuation coin. He cited a float of 4% to 5%, rising close to 25% by September, while insiders sell to the market from July through October. Hayes acknowledged that satellite internet and related businesses have been executed well, but said those are not the reason investors are buying the SpaceX story in this context.
Because the starting valuation is so high, Hayes said it creates a situation in which beating expectations becomes almost impossible. A 10% gain would not be enough if investors expected 50%, 60% or 70%. If SpaceX disappoints, Hayes believes Anthropic and OpenAI would face heavy pressure to cut their offering valuations or reduce deal size. That, in turn, would create a damaging precedent: the AI giants themselves would be lowering future expectations, cooling investor enthusiasm and weakening the AI bull-market story.
Local data-center backlash, Warsh and the rate problem
Chasse challenged the idea that Trump would turn against AI, noting that many AI leaders helped him politically, met with him privately, supported him financially and benefited from his public support. Hayes responded that he used Perplexity AI to explore whether Republicans had a path to victory if Polymarket’s House forecast was right. He said the reason Trump wants to keep the House is political self-preservation: if Democrats retake it, he and his family would face two years of congressional subpoenas, leaving little room to build the legacy he wants for a second term.
Hayes described Trump as someone without fixed ideology who cares about winning. He cited the 2020 pandemic-era fiscal transfers as an example, saying Trump sent checks broadly, with fraud and without strict income filtering, and that both rich and poor received money. Hayes then asked whether swing districts had local bans or legislation limiting data-center construction. His conclusion was that enough districts had bipartisan local resistance to data centers that an anti-AI message could flip enough seats for Republicans to retain the House.
He also emphasized that Trump would not need to actually pass policy before November. Hayes imagined Trump calling Jensen and other AI leaders and telling them not to panic while he attacks them for four months, because none of it will happen after November. Hayes compared this to tariffs, saying Trump’s hedge-fund friends lost billions of dollars while he tried to rewrite U.S. trade infrastructure, before he pulled back at a key moment. In Hayes’ view, if political strategists see an anti-AI stance delivering enough votes, there is no reason Trump would not use it rhetorically.
The discussion then moved to Warsh and the next FOMC meeting. Hayes said he did not remember Warsh’s most recent remarks in detail, but identified the narrative the market wants to believe: that the Fed can look through wartime commodity inflation and rely on an AI productivity miracle to deliver noninflationary growth, allowing rate cuts. Hayes rejected that as the comfortable market story. He said oil prices are higher and not coming down soon, while the two-year Treasury yield is roughly 60 basis points above the effective federal funds rate. In his words, the market is telling the Fed it needs to hike.
Hayes also said Trump may privately become less obsessed with rate cuts, because if he wants to address affordability, he should not push the Fed to cut while inflation is running at 3.5% to 4%. Hayes’ base case is that Warsh keeps policy unchanged, with the key distinction being whether the hold is hawkish or dovish. A hawkish hold that emphasizes building inflation pressure would lead markets to price in future rate hikes. Hayes said bubbles fear higher rates because a rising cost of capital pushes people out of the casino.
Crypto catalysts and what would bring Hayes back
Asked whether anything could give crypto a short-term relief rally between now and the midterms, Hayes said some investors could believe that MicroStrategy will continue to lift the market in some way, but he does not see many signs of money printing. Even if money is printed, he said, it is flowing directly into AI construction. He sees no major positive catalyst that would pull crypto out of its weak position or help it outperform AI.
He framed the issue through institutional incentives. In a high-growth, low-inflation environment, investors would rather buy Nvidia or Samsung than Bitcoin because those names have delivered extraordinary gains. If clients see Nasdaq up 50% while a manager is up only 10%, the manager must explain why they did not participate. Hayes said that logic consumes everyone and keeps capital chasing the strongest AI trade.
Hayes said he would consider re-entering if, by autumn, oil prices remain mild and do not rise sharply, and Trump does not turn on the AI tycoons. Even then, he set a strict condition: the SpaceX, Anthropic and OpenAI IPOs must open with exceptional success, because they are being offered at a scale he described as among the largest IPOs in human history. If reality fails to match those expectations, he said, the market has a problem.
On the next crypto bull market, Hayes said the sector needs more money printing, and that the printed money must not all go to AI. He does not know when that will happen and does not believe it is happening now. If the AI bubble breaks and financial institutions fail, a bailout could follow. That would be the moment when crypto can outperform again: AI would be discredited as a credit story, not gone as a technology, and investors would need something else to trade. Hayes said he hopes that something will be crypto, with liquidity returning to the sector.
He ended by stressing path dependence. Bitcoin has been the best-performing asset in human history over the past 15 years, in his view, but most people did not buy it at one cent. If someone entered during the ETF era, Hayes said the average buyer is down. Bitcoin does not owe gains to someone simply because they bought six months ago. In the quick-fire section, the show moved to questions about whether Bitcoin ends the year above or below $100,000, whether Hayes would buy HYPE back before year-end, and how he would allocate $1 million among Bitcoin, HYPE, short-term Treasury bills and gold.