Artificial intelligence was the most frequently cited potential shock in the Federal Reserve’s latest survey of financial-stability risks: 87% of 23 market contacts named it in the fall survey, compared with 50% in the spring. The respondents pointed to AI-linked stock valuations, cyberattacks and complex financing for data-center infrastructure, as Reuters also reported.

The number is a warning about what a small group of market participants is watching—not the Fed’s forecast that AI will trigger a crisis. The Federal Reserve report says the survey summarizes respondents’ views and does not represent the views of the Fed Board or the Federal Reserve Bank of New York.

What the survey measured

The Federal Reserve Bank of New York collected responses from 23 contacts in September and October for the Fed’s November 2026 Financial Stability Report. Contacts included professionals at broker-dealers, banks, investment funds and advisory firms. They were asked which shocks, if realized over the next 12 to 18 months, could have the greatest negative effect on the U.S. financial system, according to the report’s survey section.

AI was cited by 87% of the fall respondents, up from 50% in the spring survey of 20 contacts. Other frequently named risks were a correction in risk assets at 52%, higher long-term interest rates at 48%, cyberattacks at 43%, and U.S. fiscal-debt sustainability at 43%, according to the Federal Reserve and CNBC. Respondents could identify more than one potential shock, so the figures show how often a risk was mentioned—not its probability of occurring.

Why market contacts are watching AI

The report groups AI concerns into three areas. First, respondents cited high equity valuations and the chance that weaker-than-expected earnings could prompt a market correction. Second, they pointed to cybersecurity risks, including attacks that use AI and possible unintended damage from frontier AI models. Third, they raised concerns about complex financing connected to the AI buildout.

The distinction matters because AI risk is not just about whether a technology company’s shares fall. A correction could affect firms investing in data centers, the lenders that finance projects and other businesses tied to the buildout. The Fed report says long-term contracts, leases, guarantees and equity investments can link hyperscalers and chipmakers with third-party infrastructure developers and lenders.

How large is the data-center financing buildout?

Federal Reserve staff estimated that debt issued to finance advanced-computing infrastructure exceeded $680 billion by October 2026, about five times the $130 billion issued during all of 2023. Reuters independently reported on the same data-center financing findings. That is an issuance measure, not the amount of debt currently outstanding. The report separately estimated more than $1.1 trillion in outstanding public debt related to advanced computing as of October.

The report also identified more than $3 trillion in other obligations reported by major technology companies as of the second quarter of 2026. This consists of more than $1 trillion in leases that had not yet started and more than $2 trillion in other obligations, including take-or-pay contracts and residual-value guarantees. Those commitments are not debt. They are generally undiscounted and should not be compared directly with on-balance-sheet borrowing. The Fed also says its tracked debt-issuance estimate does not capture every private placement.

Does the survey mean the Fed expects an AI-driven crisis?

No. The 87% figure represents the share of 23 respondents who named AI as a potential shock. It is not a forecast, an estimate of the chance of a crisis, or the official position of the Federal Reserve. The report separately says the U.S. banking system remained sound and resilient, while noting areas that warrant monitoring, including elevated asset valuations and high leverage at some hedge funds. CNBC also notes the report found the broader system fairly stable.

The survey does show that people in financial markets are increasingly focused on how AI investment is financed and whether future returns will support it. For technology businesses, lenders and investors, the key questions are whether data-center spending produces durable earnings and how financing links companies across the supply chain. The Fed’s findings describe concerns worth tracking; they do not establish that the AI buildout is already a systemic crisis.

Frequently asked questions

What percentage of survey contacts cited AI as a financial-stability risk?

About 87% of 23 contacts in the Fed’s fall 2026 survey cited AI, compared with 50% in the spring survey of 20 contacts.

What AI risks did respondents mention?

They cited elevated valuations and a possible market correction, cyberattacks or unintended model-related damage, and complex financing connected to data-center investment.

Is the $3 trillion figure all AI-company debt?

No. It refers to more than $3 trillion in other reported obligations, including future leases, take-or-pay contracts and guarantees. The Fed says these are generally undiscounted and are not directly comparable to on-balance-sheet debt.

Does the survey represent the Fed’s official view?

No. The report says the survey summarizes the views of 23 market contacts and should not be interpreted as the views of the Federal Reserve Board or the Federal Reserve Bank of New York.