BitMEX cofounder Arthur Hayes has updated his Bitcoin price prediction to $1 million, now linking the target directly to an expected AI credit bust rather than just monetary policy. In his latest essay titled 'This Is Fine', published on February 17, 2026, Hayes argues that AI-driven white-collar job losses will ignite an AI credit crisis severe enough to force the Federal Reserve into large-scale money printing. As reported by BeInCrypto, Hayes expects central banks to print enough money to eventually rescue over-leveraged AI lenders, with this flood of liquidity carrying Bitcoin toward seven figures. This represents an escalation from his previous $126,000 Bitcoin target made in May, when he called AI spending history's largest fiat credit bubble. Bitcoin traded near $64,100 at the time of writing, up roughly 1% over 24 hours but down 0.5% over the past week with daily trading volume at $22.7 billion.
Hayes' latest analysis uses Bureau of Labor Statistics data to estimate the credit damage from AI-driven job displacement. According to his calculations, a 20% reduction in US knowledge workers would inflict approximately $330 billion in consumer credit losses and $227 billion in mortgage losses, generating a combined $557 billion in credit damage that represents a 13% write-down of US commercial bank equity. Hayes notes the distribution problem, explaining that the eight Too Big to Fail institutions are adequately capitalized, but thousands of smaller regional banks are not. The market will identify weak balance sheets, crush stock prices, trigger regulatory capital breaches, and spark depositor flight, creating a sequence Hayes compares directly to the regional bank collapses of early 2023 but at greater magnitude.
During his appearance on the Thinking Crypto podcast, Hayes highlighted critical structural flaws in AI infrastructure financing. GPU loans for AI hardware are amortized over five to six years, while GPUs become obsolete in about two years, creating a fundamental mismatch between asset lifespan and loan duration. As Hayes explained, 'As a credit event, this will be bigger than subprime. If it's all about if we go to the China price, then all these assumptions in terms of the cash flows that these GPUs are based on become kind of spurious, and it becomes a credit event.' He warns that if cheaper Chinese AI models commoditize inference, the financial assumptions supporting GPU loans could collapse entirely, potentially triggering a crisis larger than the 2008 financial crisis.
Hayes raises significant concerns about the scale of AI-related financial commitments, with about $1.5 trillion in AI-related debt issued between 2022 and mid-2026, predominantly in 2025. Major tech companies have made substantial data-center lease commitments totaling around $1.09 trillion, which is nearly four times their existing lease liabilities of $285 billion. S&P Global analyst Andrew Chang points out a critical maturity mismatch for Oracle, with lease durations of 15 to 19 years versus customer contracts lasting only five years, heightening credit risk if contract renewals fail. Hayes believes 'AI sucked all the capital out of the room' and continues to absorb capital that would otherwise support Bitcoin's market rally.
Despite his bullish long-term Bitcoin prediction, Hayes maintains a cautious near-term outlook for the cryptocurrency and outlines a two-scenario framework for traders. He projects Bitcoin could range between $60,000 and $70,000, with a potential dip to $50,000 before a credit-cycle recovery. During his Thinking Crypto interview, Hayes expressed uncertainty about timing, suggesting the AI bubble unwind could happen 'this fall' or take 'years'. Hayes warns that the worse the initial credit-destruction event, the larger the eventual monetary response, and the more violent Bitcoin's recovery from whatever lows it hits during the dislocation. He explicitly advises traders to wait for a confirmed Fed pivot before deploying aggressively into risk assets, as the next major entry signal comes from the Fed's balance sheet rather than price action alone.
