IMF

I.M.F. Warns of a Debt Spiral in the A.I. Build-Out

co-produced by AI16. October 2026, 07:57

The A.I. boom is increasingly running on borrowed money, and that has the International Monetary Fund worried: In its newly released Global Financial Stability Report, the I.M.F. ranks a correction in A.I.-related assets as the top risk to financial markets, ahead of rising bond yields and cyberattacks. There is “no room for complacency,” the I.M.F. economists Athanasios Vamvakidis and Jason Wu write in a blog post accompanying the report.

An A.I. Correction as the Top Risk

A.I. now props up the global economy and markets, through investment in data centers, chips and infrastructure and through the strong earnings of the companies driving the boom. If that investment turns out weaker or less durable than expected, or if profits fail to materialize, the I.M.F. says A.I. stocks and other A.I.-linked assets could face a correction. Rising leverage and concentrated exposures would amplify it.

Two figures from the report show how much leverage has built up. The big cloud companies, known as hyperscalers, now account for 10 percent of all global investment-grade bond issuance, up from less than 1 percent, as Unite.AI summarizes. At the same time, their free cash flow is shrinking as A.I. spending soars. Investors are also borrowing more: Single stocks now make up 28 percent of leveraged E.T.F.s, up from 16 percent. Hedge funds’ equity positions plus margin loans have grown from about 10 percent to about 15 percent of advanced economies’ G.D.P.

The recent slump in A.I. and chip stocks showed how quickly sentiment can turn. Over the summer, experts at the European Central Bank had already warned of an A.I. bubble bursting.

The A.I. Financing Loop

The I.M.F. is particularly concerned about circular financing in the A.I. industry. It examined 14 publicly disclosed deals since 2023. In them, hyperscalers invested about $300 billion (about €268 billion) in equity in A.I. companies and provided about $650 billion (about €580 billion) in debt. In return, the A.I. companies committed to spending $1.4 trillion (about €1.25 trillion) with their backers in the future, for example on data centers, computing power and cloud services. About $320 billion of the debt sits with the funded companies themselves, according to the report, partly backed by investor guarantees or held off balance sheet.

The problem lies in the mutual dependence. The recipients are usually financially much weaker than the hyperscalers funding them. If one of them runs into trouble, its backers’ expected revenue falls, and with it the value of their stakes and the security of the loans tied to these deals. The I.M.F. calls this “wrong-way risk”: The investments lose value at exactly the moment the revenue linked to them dries up.

The I.M.F. therefore recommends limiting leverage and liquidity risks in the A.I. sector, for instance through sector-specific capital requirements and concentration limits on bank lending.

The Counterpoint

The I.M.F. itself stresses that the hyperscalers earn strong profits in their core businesses, generate solid cash flow and have good credit ratings, and that their investors are generally financially sound. Because A.I. earnings have repeatedly beaten forecasts, price-to-earnings ratios have actually fallen. Many on Wall Street do not see a bubble either: Analysts at Goldman Sachs and Morgan Stanley consider such fears premature, and a JPMorgan indicator framework concluded that the A.I. boom does not yet meet the criteria of a bubble, according to Tech Times.