“Big Short” investor Steve Eisman argues that the traditional playbook for reading the economy no longer applies. For years he lived by the rule “as the banks go, so goes the economy.” Now, he says, “Probably not this time.” His reasoning is that bank balance sheets look healthy while the real risk has migrated elsewhere. Non-accruals fell 5% year-over-year at JPMorgan ($JPM) and 4% at Bank of America ($BAC), and firms like Goldman Sachs ($GS) remain steady. Yet Eisman believes the economy’s fate now hinges on whether the AI buildout succeeds or stumbles. The danger zone, in his view, is private credit. These markets are heavily exposed to loans made to software companies, and they lack the transparency of regulated banks. If corporate budgets shift away from software subscriptions toward AI hardware, some of those borrowers could struggle to repay, and problems would surface in opaque corners investors can’t easily see. The market is pricing meaningful uncertainty. Polymarket puts the odds of a U.S. recession by the end of 2026 near 12%, while JPMorgan economists peg it around 35%. A separate contract on an AI bubble bursting sits near 17%. The backdrop matters too: the 10-year Treasury yield is pushing toward 4.6% and oil trades above $80, tightening conditions just as the AI bet grows larger.