Fed Warns AI Investment Boom Is Emerging as Top Threat to Financial Stability
November 15, 2026
The Federal Reserve’s semi-annual Financial Stability Report found that financial-market contacts increasingly see AI-related risks as the biggest threat to stability: 87% named them, up from 50% in the spring.
Concern also rose across the other risks surveyed, including a correction in risk-asset prices (52%, up from 35%), higher long-term rates (48%, from 20%), cyberattacks (43%, from 20%) and fiscal-debt sustainability (43%, from 30%).
The Fed said the scale and pace of AI infrastructure investment could soon exceed the late-1990s broadband boom, with contracts, leases, guarantees and equity stakes tying major technology firms to smaller developers and financiers.
Large technology companies and graphics-chip producers have invested more than $100 billion in AI labs and cloud-computing startups, which often spend that money on services or chips from their investors; Barr said customer revenue will eventually need to replace this supplier funding.
The IMF estimates that AI companies’ circular financing—where investment flows back to investors through purchases—has topped $1 trillion, with about two-thirds funded by debt; Georgieva said the pace of growth is especially concerning.
Some obligations are difficult for investors to identify and assess because they sit off balance sheet, including energy contracts, data-center and computing leases, and guarantees on resale values.
Whether the spending will generate adequate returns remains an open question: 24.6% of U.S. companies reported using AI in October, and 57% of adopters used it in three or fewer business functions; 37% of companies in a McKinsey survey said AI was boosting operating profit.
If investment pulls back, the downturn could feed on itself: falling valuations and tighter credit could stall data-center projects, cut demand for chips and infrastructure, and spread losses among companies, investors and lenders.
Barr said limited debt in the boom and strong capital at major banks help contain direct-loss concerns, but complex financing, higher borrowing costs, hedge-fund leverage and counterparty exposures could amplify a sharp reversal and forced selling.
If expected revenues fail to materialize, falling AI asset values could trigger equity and debt losses, defaults or bankruptcies, and weaker investment and consumer spending; a BIS working paper estimates investment is 1.5 times the efficient level.
The Hyundai Research Institute expects U.S. AI investment to keep growing through 2027, supported by cloud revenue and existing plans, but says projects could be delayed and spending reduced around 2028 as returns, financing and infrastructure constraints converge.
Meanwhile, data centers are driving up electricity and freshwater demand while new power-generation capacity lags; Elon Musk has proposed putting data centers in space and powering them with solar panels.
Summary based on 26 sources
Get a daily email with more Macroeconomics stories
Sources

The New York Times • Nov 16, 2026
Opinion | The Financial Bubble Cycles Are Getting Worse
CNBC • Nov 13, 2026
AI risks are the biggest threat to the financial system, Fed survey finds
Axios • Nov 16, 2026
The AI boom faces a dot-com reckoning