Federal Reserve Survey: 87% Now Call AI a Financial Stability Risk as Data-Centre Debt Passes $680 Billion
Artificial intelligence is now the risk Wall Street worries about most, according to the Federal Reserve's Financial Stability Report released on 13 November 2026: 87% of the market contacts the New York Fed surveyed in September and October named AI as a salient risk to US financial stability, up from 50% in the spring survey. The same report says debt issued to finance advanced computing infrastructure passed $680 billion by October 2026, about five times the $130 billion issued in all of 2023.
The Financial Stability Report is the Federal Reserve Board's twice-yearly assessment of the resilience of the US financial system, covering asset valuations, borrowing, leverage and funding risks, plus a survey of market contacts about near-term risks. The November 2026 edition reflects data as of 28 October 2026 and was released on Friday 13 November.
The survey: 87%, and what changed since spring
Staff at the Federal Reserve Bank of New York asked 23 market contacts — professionals at broker-dealers, banks, investment funds and advisory firms — about risks to US financial stability in September and October 2026. Box 5.1, "Survey of Salient Risks to Financial Stability," says plainly: "AI was the top-cited risk." The report's summary table puts the numbers beside the spring survey, when 20 contacts were asked in March and April:
| Risk cited | Fall 2026 (23 contacts) | Spring 2026 (20 contacts) |
|---|---|---|
| Artificial intelligence | 87% | 50% |
| Risk asset correction | 52% | 35% |
| Higher long-term rates | 48% | 20% |
| Cyberattacks | 43% | 20% |
| Fiscal debt sustainability | 43% | 30% |
Two things about that table deserve emphasis. First, AI did not just rise — it rose the most, from half of respondents to nearly nine in ten, overtaking the geopolitical risks and oil shock that topped the spring survey. Second, the survey is a summary of what respondents said, and the report states explicitly that it "should not be interpreted as representing the views of the Federal Reserve Board or the Federal Reserve Bank of New York." The 87% is Wall Street's worry, recorded by the Fed — not the Fed's own forecast. The report flags vulnerabilities; it does not predict that any of them will crystallise.
Three worries inside the AI number
The report breaks the AI concern into three distinct areas. The first is valuation: elevated equity valuations and the possibility that lower-than-expected earnings could result in lower valuations and a market correction. The second is cybersecurity, "stemming from either a malicious attack by sophisticated state actors or damage caused by frontier AI models themselves." The third is "risks associated with complex forms of AI-related financing" — the borrowing structures behind the buildout, which the report examines in a dedicated box.
The correction worry gets its own section. Respondents "widely cited the risk of a sharp correction in AI-related equities, driven by uncertain returns on large capital expenditures. Many noted that stretched valuations, circular financing arrangements, and heavy reliance on debt to fund the AI buildout could amplify losses, with the potential for a deterioration in broader risk assets to spill over to business investment and household consumption." The fear is not only that AI stocks fall — it is that the fall spreads, because the buildout is debt-funded and the financing links big technology firms to smaller, riskier ones.
The cyber worry cuts the other way. Respondents said advances in AI are lowering the barrier to sophisticated attacks by state and non-state actors, while others pointed to the risk of unintended damage caused by frontier AI models themselves. Cyberattacks and fiscal debt sustainability were also seen as carrying the most severe potential consequences of any risk cited.
![]() |
The $680 billion buildout
The financing numbers sit in Box 2.2, "Leverage in the Data Center Buildout." "Advanced computing–related debt issuance rose to more than $680 billion by October 2026, about a fivefold increase from $130 billion in all of 2023," the report says. Investment-grade technology companies issued most of it, and advanced computing–related borrowing rose from less than 6 percent of nonfinancial investment-grade corporate debt issuance in 2023 to about 31 percent in 2026. Outstanding public debt related to advanced computing is estimated at more than $1.1 trillion as of October 2026.
The composition of that debt is changing, and the report says that matters as much as the size. Issuance is evolving "from mostly corporate bonds to a combination of public and private debt, structured credit, and project finance." Large technology firms are also "rapidly increasing future lease commitments, backstop guarantees provided to other firms, and other complex and contingent obligations." Those other obligations are large: not-yet-commenced leases whose undiscounted values exceed $1 trillion, and take-or-pay contracts and residual value guarantees whose undiscounted values exceed $2 trillion.
That is where the "circular financing" phrase comes from. Hyperscaler firms enter long-term leases for data centre capacity with third-party developers, and those leases let the developers borrow to build. GPU makers guarantee the residual value of the GPU collateral to support purchases by AI firms. "These debt financings linking hyperscalers and GPU designers with smaller firms are often supplemented with equity investments by the hyperscalers and GPU designers, which create a 'circular financing' dynamic," the report says. The largest US bank holding companies have close to $1 trillion of exposure to large technology firms, data centre developers, advanced computing–related special purpose entities and private credit vehicles, and utility companies.
The report draws one historical comparison. The current data centre buildout "often draws comparisons to the late 1990s and early 2000s broadband telecom investment cycle." Five hyperscalers' cumulative capital spending and debt issuance since 2020 have been tracking the telecom industry's 1995 cycle closely, the report says, "but advanced computing investment has been accelerating of late, suggesting that the pace of the current cycle's investment may surpass that of the 1990s telecom cycle relatively soon." The telecom cycle ended in a bust. The Fed does not say this one will.
The rest of the report is not calm either
AI dominates the survey, but the underlying vulnerability assessment gives it plenty of company. The forward price-to-earnings ratio of the S&P 500 "remained well above its historical median," partly driven by market enthusiasm for AI. Corporate bond and leveraged loan spreads "stayed at the low end of their historical range." Hedge fund leverage "reached the highest level since 2013." Treasury yields "rose markedly," and term premiums climbed above their median level since the late 1990s.
None of that is a prediction. The report's framework is explicit: shocks are hard to predict, vulnerabilities can be monitored as they build, and the job is to assess where stress would hit hardest if it arrived.
![]() |
Background: the market context is the same story this blog has tracked all autumn — is the AI bubble bursting after one brutal week for Alphabet, the Microsoft Google Amazon cloud earnings Q3 2026 that fund the buildout, the AI data centre electricity costs landing on consumers, and the Anthropic IPO filing.
What to watch next
The next Financial Stability Report arrives in about six months, and the survey will be the number to compare: whether AI stays at the top, and whether the percentage keeps climbing or starts to fall. Before that, watch the two things respondents actually named. First, earnings against capital expenditure — the report's correction scenario starts with lower-than-expected earnings from the companies spending the most. Second, the financing — whether the private credit, project finance and structured deals behind the buildout keep finding buyers.
![]() |
Key Takeaways
- The Federal Reserve's November 2026 Financial Stability Report, released 13 Nov, found AI the top-cited risk in its survey of 23 market contacts: 87% named it, up from 50% in the spring survey of 20 contacts.
- Advanced computing–related debt issuance passed $680 billion by October 2026, about five times the $130 billion of 2023, rising from under 6% to about 31% of nonfinancial investment-grade issuance.
- Beyond the debt itself, not-yet-commenced leases exceed $1 trillion and guarantees and contract commitments exceed $2 trillion, in what the Fed calls a "circular financing" dynamic.
Conclusion
The November 2026 Financial Stability Report does not say the AI boom will end badly. It says something narrower and harder to dismiss: the market professionals the New York Fed asks now rank AI as the top risk to financial stability by a wide margin, and the borrowing behind the buildout — more than $680 billion this year, five times the 2023 total, wrapped in leases, guarantees and circular financing — is large enough that a disappointment in earnings or in the debt markets would not stay contained. None of that proves a bust is coming. It proves the Fed is watching the same numbers the sceptics are.
Sources
- Board of Governors of the Federal Reserve System — Financial Stability Report, November 2026 (PDF)
- Board of Governors of the Federal Reserve System — Financial Stability Report index
- Reuters (syndicated) — AI rockets to top of worries about financial system in latest Fed survey, by Pete Schroeder (13 Nov 2026)



Comments
Post a Comment