Hayes Shorts the Crash, Longs the Printer: $BTC and $ETH as the AI Bailout Bet 🖨️

BitMEX co-founder Arthur Hayes has advised traders against wagering on an AI crash, instead recommending long positions in Bitcoin ($BTC) and Ethereum ($ETH) to capture the liquidity he expects from a multi-trillion-dollar government rescue. Sharing the view on X on October 15, 2026, Hayes wrote: "Is AI finally going to dump? IDK but I wouldn't bet either way. Instead just trade the eventual multi-trillion dollar bail out. Long $BTC $ETH." He paired the call with a chart of CoreWeave's declining stock price, signaling where the risk may originate.

Hayes anchored his thesis in a historical pattern in which major technology rollouts become overbuilt, collapse and are subsequently backstopped by authorities. He projected the AI data center buildout would meet that moment in late 2027 or 2028, after which he expected policymakers to print more money than they did following the 2008 financial crisis. In that scenario, he argued, Bitcoin and other crypto would absorb the new liquidity, with Bitcoin potentially reaching $1 million or higher. He pointed to the 2020 pandemic response as a comparable case, noting that the Federal Reserve more than doubled its balance sheet as Bitcoin rose from under $5,000 to above $60,000.

That relationship has weakened in recent months. US money supply reached a record in August, yet Bitcoin trades roughly 35% below its October 2025 peak, a gap mirrored across global liquidity measures. Both Bitcoin and Ethereum have also trailed most AI stocks in 2026, complicating the case that crypto automatically catches the next wave of liquidity.

The debate has since shifted toward the financial plumbing behind AI. CoreWeave, the AI cloud provider, dropped 15% a day after CNBC host Jim Cramer identified it as his only trusted AI cloud pick, drawing attention to AI-related debt across the sector. In a September newsletter, Hayes wrote that major US insurers were insolvent at market value once AI debt was marked to market. The unresolved question for traders is no longer simply whether AI shares fall, but which entities absorb the losses when they do.

Not all market participants share Hayes's timeline. Fundstrat co-founder Tom Lee has compared current AI conditions to 1997 rather than 1999, placing a potential bubble peak closer to 2033. Hayes, for his part, has framed his positioning as a bet on the rescue rather than a prediction of the crash.