Arthur Hayes, co-founder of BitMEX and managing partner at Maelstrom, published an essay on February 17 arguing that Bitcoin’s recent price drop is an early warning sign of a coming US credit crisis caused by AI-driven job losses.
Bitcoin hit an all-time high in October 2025 and has since fallen sharply. The Nasdaq 100 has remained relatively flat over the same period.
Hayes says this divergence matters because Bitcoin has historically moved in line with US tech stocks. When the two split, he argues it signals stress in the broader credit system.
His core thesis is that AI tools will eliminate a large share of white-collar jobs within the next few years. He points to statements from AI company executives who have predicted that most knowledge work done by humans could be automated within two years.
According to the US Bureau of Labor Statistics, there are currently 72.1 million knowledge workers in the US out of a total workforce of 164.5 million.
Hayes built a model estimating what happens to US bank balance sheets if 20% of those workers lose their jobs. He used Federal Reserve data showing total consumer credit at $5.1 trillion, stripping out government-backed student loans to arrive at $3.76 trillion held by banks.
He also factored in mortgage debt, noting that knowledge workers earn an average of $85,000 per year, placing them around the 70th income percentile. Around 74% of that group own homes, and 42% of homeowners carry a mortgage, with an average balance of $250,000.
Running those numbers through a 20% job loss scenario, Hayes estimates banks would face $330 billion in consumer credit losses and $227 billion in mortgage losses, totaling roughly $557 billion.
After accounting for existing loan loss reserves held by US commercial banks, he estimates the event would cause a 13% write-down of total US commercial bank equity.
Hayes says the too-big-to-fail banks would likely survive, but smaller regional and community banks with higher leverage or riskier loan books would face collapse. He compares this to the regional banking crisis of early 2023, when three banks failed within two weeks.
He points to current market data as early confirmation: the iShares Software ETF (IGV) has underperformed the Nasdaq, private credit lender Blue Owl has sold off alongside it, and the ratio of consumer staples to consumer discretionary stocks has risen, suggesting people are spending less on non-essentials.
Hayes argues the Federal Reserve will be slow to respond. He says Trump’s decision to launch a criminal investigation into Fed Chair Jerome Powell has created internal resistance at the central bank.
Powell’s term as chair ends in May. If Trump’s reported pick, Kevin Warsh, is confirmed, Hayes expects him to face a hostile board of governors that could block policy changes.
Hayes believes the Fed will eventually be forced to print money once enough regional banks fail and credit markets seize up. He draws a direct comparison to March 2023, when Signature Bank’s collapse led to an emergency joint announcement from the Treasury and Fed.
He says Zcash (ZEC) and Hyperliquid (HYPE) are the two assets Maelstrom plans to buy when that moment arrives, and has stated he plans to publish a model showing HYPE could reach $150 by July.
The post Arthur Hayes Warns of AI-Driven Credit Crisis: Is a 2008-Style Crash Coming? appeared first on CoinCentral.


