AI-related risks have become the most frequently cited concern in the Federal Reserve’s latest financial-stability survey of market contacts, according to reporting released Friday. Eighty-seven percent of respondents identified AI as a salient risk to the financial system, up from 50% in the prior spring survey, CNBC reported.
The shift does not mean the Fed is forecasting an AI-led market crash or banking disruption. The survey measures what its contacts regard as important vulnerabilities. But the 37-percentage-point increase in AI concern—equivalent to a 74% rise from the prior survey’s respondent share—places the technology boom at the intersection of equity-market valuations, cyber resilience and the financing of the physical infrastructure needed to run advanced computing systems.
AI surpassed a broad correction in risk assets, cited by 52% of respondents. Higher long-term interest rates were named by 48%, while cyberattacks and U.S. fiscal debt sustainability were each identified by 43%, CNBC reported. The findings place concerns surrounding AI ahead of risks that have more routinely dominated post-pandemic market discussions.
What the survey result measures
The Federal Reserve’s financial-stability exercise gathers views from market participants rather than assigning a probability to any particular shock. Its ranking therefore is best read as an indicator of where investors, lenders and other financial-market contacts see concentrations of uncertainty—not as evidence that an AI-related disruption is imminent.
Responses reflected in the latest survey covered a period through Oct. 28, according to CNBC. The comparison with the spring reading is notable because the concern is not confined to one potential outcome. Respondents pointed to elevated equity valuations, the possibility of a market correction, cybersecurity threats and increasingly complicated financing arrangements associated with AI investment.
Those categories can reinforce one another without being the same risk. Expensive technology shares can be vulnerable if earnings expectations are reset. Cybersecurity worries concern the possible misuse of AI and the resilience of increasingly digitized financial and business systems. Financing risks arise when the multiyear cost of chips, computing capacity, power and data centers is supported through a mix of debt and contractual commitments spread among different companies.
Data-center commitments draw closer scrutiny
The scale of the build-out is a reason the financing issue has moved beyond technology-company capital-spending plans. Yahoo Finance reported that the Fed was monitoring leverage and contingent obligations connected to data-center construction, including debt, long-term leases and guarantees.
Yahoo Finance’s account of the report said advanced-computing-related debt issuance had exceeded $680 billion by October, compared with about $130 billion during all of 2023. On those reported figures, the newer amount is more than five times the 2023 total, although the comparison should not be treated as a forecast for annual issuance or a measure of losses. It instead illustrates how rapidly funding tied to the sector has expanded.
The same report put future leases and backstop guarantees related to the data-center build-out at roughly $3 trillion. That figure requires a different reading from the debt-issuance number. Debt is a present borrowing obligation. A lease is a contractual promise to make future payments, while a guarantee generally becomes a payment obligation only if the covered borrower or counterparty fails to perform. Such commitments can still transmit stress across companies and lenders, but they are not equivalent to $3 trillion of booked debt on corporate balance sheets.
That distinction is especially relevant to large AI infrastructure projects. A data-center developer may depend on long-term customer commitments; a technology company may rely on a specialized supplier; and lenders may assess their exposure against contracts whose economic value depends on future demand for computing capacity. Yahoo Finance described the Fed’s discussion as examining links among hyperscalers, chip designers and smaller companies that can make the large firms’ economic exposure less apparent from their standalone balance sheets.
Valuations and financing meet in public markets
For investors, the survey’s ranking joins two questions often considered separately. One is whether public-market valuations already assume sustained AI revenue growth. The other is how much of the infrastructure required to support that growth is being financed through debt and long-dated commitments. A reassessment of expected demand could affect both: equity prices through revised profit assumptions and credit markets through the durability of contracts supporting data-center projects.
The available reporting does not establish that those commitments are broadly distressed, nor does it identify a particular company or lender at risk. It does show that the Fed’s contacts increasingly regard the valuation, cyber and financing components as related sources of financial vulnerability rather than as isolated technology-sector issues.
The headline number also leaves an important practical boundary. An 87% response rate says AI is highly salient to surveyed contacts; it does not quantify the likelihood or size of a market loss. The more concrete figures in the report—the more than $680 billion in reported advanced-computing debt issuance and the estimated $3 trillion in future leases and guarantees—help explain why scrutiny has shifted toward the obligations underlying the data-center expansion, not just the companies selling AI products.