The Federal Reserve said Friday that risks tied to artificial intelligence have become the primary concern for financial stability, according to its latest semi-annual Financial Stability Report.
A survey conducted in September and October, and summarized in the report, found that 87% of contacts identified AI-related risks as a meaningful threat to the financial system - up sharply from 50% reported in the May survey. Respondents singled out several channels through which AI could create vulnerabilities, including heightened cyberattack risk, the emergence of complex financing structures connected to AI projects, and the possibility that an AI-driven slowdown in corporate earnings could precipitate a broader market correction.
Other risks were also prominent in the survey. A broad correction in risk assets was the second most commonly cited concern at 52% of respondents. Higher long-term interest rates worried 48% of contacts. Cyberattacks were specifically flagged by 43% of those surveyed, and 43% also pointed to fiscal debt sustainability as a concern.
The Fed noted the survey took place before several high-profile developments that have occurred since September and October. In particular, the survey preceded recent events at chipmaker Nvidia. The report records that Nvidia's CEO, Jensen Huang, stepped down on a Thursday after revelations of extramarital affairs, including one involving a Chinese businesswoman who was reported by the New York Times to be under suspicion as a Chinese spy. The timing of the survey also predated recent midterm elections during which public concern about AI featured prominently as an issue.
The report underscored the rapid growth in borrowing tied to advanced computing and data center investments. By October 2026, advanced computing-related debt issuance reached $680 billion, which the Fed described as roughly five times the total for all of 2023. That level of borrowing now accounts for nearly one-third of all nonfinancial investment-grade corporate debt issuance, compared with under 6% three years earlier.
The central bank drew a parallel between the current wave of investment and the late 1990s broadband telecom investment cycle, observing that the present cycle of debt issuance and infrastructure spending for AI-related computing could soon eclipse the scale of that earlier expansion.
What the survey measured
- AI-related risk: 87% of contacts viewed it as a significant financial stability risk (up from 50% in May).
- Broad risk-asset correction: 52% of respondents cited this as a major concern.
- Higher long-term rates: noted by 48% of those surveyed.
- Cyberattacks and fiscal debt sustainability: each flagged by 43% of respondents.
Context and timing
The Fed report makes clear that the survey responses reflect conditions and perceptions gathered before several subsequent developments, including leadership changes at a major chipmaker and election-season debates over AI. The report also quantifies the shift in corporate borrowing patterns tied to advanced computing projects and data centers, important details for understanding how new technology cycles can affect credit markets.
Implications noted by the Fed
Officials highlighted the speed at which investment-grade corporate debt has shifted toward financing advanced computing infrastructure. The Fed compared the current expansion in AI-related capital formation and borrowing to the broadband telecommunications investment phase of the late 1990s, indicating the possibility that current investment and debt issuance could surpass that earlier period.