Arthur Hayes: AI Bubble Burst Could Push Bitcoin to $1M
Financial commentator Arthur Hayes thinks the collapse of the AI boom could send Bitcoin to $1 million. My take: the mechanism is more convincing than the target.
Hayes, who describes himself as a “genius trader,” expects trouble to surface in 2027, with the full damage arriving in 2028. His case is simple. Extreme, too. When the AI boom breaks, governments and central banks may flood the financial system with enough money to trigger a massive Bitcoin rally.
In his latest essay, “Situationship,” Hayes argues that investors have misread AI infrastructure. He sees less of a technology business and more debt-heavy real estate. That shift matters. Viewed through a property-and-credit lens, current valuations suddenly look shaky.
Most investors treat AI spending as a wager on better software and faster technological progress. That is only half right, Hayes says. Underneath the software sit data centers, power plants, land and costly hardware—all financed like property, often with borrowed money. These assets cannot be sold quickly. They also face the same credit cycles that crushed earlier construction booms. I’ll be honest: this is the strongest part of his argument. The leap from distressed infrastructure to $1 million Bitcoin? Much harder to buy.
Hayes thinks reckless lending for AI construction will cause the collapse, not falling profits at AI companies. His preferred comparison is the 2008 mortgage crisis rather than the dot-com crash of 2000. The distinction is crucial.
AI demand could keep growing while the financing underneath it falls apart. Sounds contradictory? It is not. New chips are becoming faster and more efficient, allowing existing data centers to process more work without companies maintaining the same construction pace. Building slows. The debt from the earlier spree remains. Hayes expects markets to notice in 2027; by 2028, he says, the damage will be impossible to ignore. That timetable is strikingly precise. Personally, I like forecasts with dates: at least we will know soon enough whether he called it correctly.
Hayes also thinks banks will keep funding AI projects as the danger grows. The loans generate revenue. AI spending supports US policy goals, and lenders may assume Washington will bail out the largest companies if the trade goes wrong.
Conventional advice says lenders pull back when risks rise. In practice, banks do not often walk away while the fees keep rolling in. Hayes expects them to lend until the damage becomes obvious, then ask the government for help. He doubts authorities would tolerate widespread bankruptcies among large AI companies or infrastructure owners. The response could include emergency loans and direct aid. Authorities might also purchase troubled assets. Why does that matter for Bitcoin? Because Hayes’ case depends on central banks printing more money to contain the fallout. That happened in 2008. Different companies, familiar bailout playbook.
Hayes says AI has absorbed money that might otherwise have flowed into Bitcoin. Once the trade unwinds, bailout spending and easier monetary policy could redirect some of that capital toward BTC, eventually pushing its price to $1 million.
Investors have spent enormous sums on AI infrastructure in recent years. Hayes thinks this partly explains why Bitcoin has not climbed as much as growth in the global money supply might imply: capital that could have entered BTC instead funded data centers. When the cycle reverses, investors will need somewhere else to park that money. Hayes expects Bitcoin to find a floor before beginning a long bull market. He also predicts that the next bailout will exceed the response to the 2008 crisis. That is his bridge to $1 million. My reaction? It is a huge leap. The forecast needs both a brutal credit crash and an equally aggressive rescue. One without the other does not get him there.
What this means
Hayes expects the AI credit boom to end with investors scrambling for scarce assets such as Bitcoin. The sequence is familiar: cheap debt funds too much construction, then losses spread through the financial system. Governments inject more money to halt the panic. We have seen that script before, though never with AI infrastructure at its center.
Most bubble arguments assume the underlying technology must fail. Counter to that view, AI does not have to become useless for Hayes to be right. Companies merely need to build more data centers than the market requires and borrow more than they can repay. If those projects begin failing, governments may intervene to protect banks and large businesses. Hayes thinks the extra money entering the system would power Bitcoin’s next major rally. It could also strengthen the case for holding BTC as a hedge against currency dilution and financial turmoil, much as some people own gold. But let’s not soften the risk: Bitcoin swings far more violently than gold. Calling it a safe haven does not make holding it comfortable.
Watch AI investment in 2027 first. Lending for data center projects could expose trouble early, especially if banks tighten their terms or reject deals they readily accepted the year before. Central bank policy comes next; new quantitative easing or emergency lending programs would fit Hayes’ prediction. Then comes the market test. Does Bitcoin find a clear bottom and remain above previous resistance? If so, another bull cycle may be underway. Still, $1 million would be far from certain. Yes, that sounds cautious after laying out Hayes’ bullish chain—but caution belongs here. Such a move would show only that he got the opening stages of the story right.