The Federal Reserve's November Financial Stability Report put artificial intelligence at the top of its risk list, with 87% of surveyed market contacts naming AI a salient risk to the US financial system. That is up from 50% in the spring survey and well ahead of the next most-cited concern, a correction in risk assets at 52%. Higher long-term interest rates followed at 48%, while cyberattacks and fiscal debt sustainability were each cited by 43%. The New York Fed polled 23 professionals at broker-dealers, banks, investment funds and advisory firms for the report, which covers the period through October 28.
Respondents flagged three interlocking AI worries: stretched equity valuations, AI-enabled cyberattacks, and complex financing tied to data center construction. The Fed said debt issuance linked to advanced computing reached about USD 680 billion in October 2026, roughly one-third of nonfinancial investment-grade corporate bond issuance, against less than 6% three years earlier. Its estimate of bank holding company exposure to the wider AI infrastructure ecosystem sits near USD 1 trillion, and hedge fund leverage is at its highest since 2013. Several contacts warned that an unexpected slowdown in AI earnings could trigger a broad market correction and tighten credit across the economy.
For founders, the signal is less about model capability than about the cost and availability of capital. If lenders and investors reprice AI infrastructure risk, the knock-on effects land on cloud pricing, GPU rental rates and the financing terms available to startups building on top of that capacity. The Fed stressed that the survey reflects respondents' views rather than the Board's own assessment, and it described the broader system as resilient. Still, two consecutive surveys moving AI from a mid-table concern to the outright top risk is a shift worth planning around: diversify vendors, avoid single-cloud lock-in, and keep runway assumptions defensible if compute pricing stops falling.
