−25%Infinity TerminalGet it now
The Federal Reserve Bank of Cleveland, host of this week's Financial Stability Conference.

Business · 20 November 2026

BusinessInvestigation

Central bankers warn AI debt threatens financial stability

Regulators meeting in Cleveland this week argued over whether debt financing the AI boom could turn an investment surge into a financial shock.

By the MarketPedia Business Desk · 4 min read

Cover · IllustrationThe Federal Reserve Bank of Cleveland, host of this week's Financial Stability Conference.

14thedition of the Cleveland Fed-Yale Financial Stability Conference
2 dayslength of this week's sessions in Cleveland
July 2,026when hedge fund Situational Awareness nearly collapsed on AI-linked leverage

The Federal Reserve Bank of Cleveland and Yale University opened the 14th annual Financial Stability Conference in Cleveland on Thursday, with two days of sessions built around one question: whether the debt financing the artificial intelligence boom could turn an investment surge into a financial shock.

Keynote speakers Paola Boel of the Cleveland Fed and Cornelia Holthausen of the European Central Bank opened the sessions, according to the conference agenda.

The timing was pointed. Two days earlier, on November 18, the ECB published its Financial Stability Review along with a focus note on the debt financing of AI, warning that leverage tied to data centers and chips has grown faster than regulators can track.

Debt that outlives the hardware

The ECB's note did not lean on a single number to make its case. It leaned instead on shape: the maturity of debt issued by hyperscalers often exceeds the economic lifetime of the physical capital, such as advanced chips, that the debt is meant to fund, the central bank said.

That mismatch is the crux of the systemic concern, in the reading of analysts who track corporate leverage. Chips depreciate within a few years. Bonds financing them can run for a decade or longer.

If the hardware goes obsolete, or the AI business case falters, before the debt matures, lenders are left holding claims against assets that have already lost most of their value. That is the scenario regulators in Cleveland spent Thursday trying to size.

Circular financing

A second worry raised in Cleveland concerns the structure of the financing, not just its duration. The IMF's October Global Financial Stability Report described a circular pattern in which chipmakers and cloud providers put capital into AI labs that then buy the hardware those same chipmakers and cloud providers make.

The same IMF analysis found that hyperscalers have increased their share of global investment-grade bond issuance as their free cash flow shrinks under the weight of capital spending.

The likelier reading, analysts say, is that this circularity makes it harder for outside investors and rating agencies to judge where the real risk sits. Revenue, capital and collateral move within a small cluster of related companies rather than through arm's-length markets, so a shock to one node travels fast to the others.

The tower of a Federal Reserve regional bank, framed against a November sky.
The tower of a Federal Reserve regional bank, framed against a November sky.Illustration · MarketPedia

A near miss in July

The conference did not have to reach far for an example of what can go wrong. In July, the US hedge fund Situational Awareness came close to collapse after high exposure concentration and leverage left it unable to absorb a reversal in AI-linked positions, the ECB said in its Financial Stability Review.

The episode echoes a concern raised in the minutes of the Federal Reserve's most recent policy meeting, cited in the ECB's review. "Almost all participants commented on risks associated with a scenario in which the returns on AI investment turned out to be disappointing to investors," the minutes said.

Cornelia Holthausen, speaking on November 16 on risks and resilience in the euro area financial system, said regulators are watching for a reassessment of the AI narrative that could trigger fund redemptions and force asset sales, amplifying any initial shock.

Her speech's title, "Risks and resilience," was not accidental balance. Holthausen also pointed to resilience built up elsewhere in the euro area financial system, even as she flagged AI-linked exposure as a growing vulnerability to monitor.

A data center complex of the kind now financed increasingly through corporate bond markets.
A data center complex of the kind now financed increasingly through corporate bond markets. Illustration · MarketPedia

Tools regulators do not yet have

What to do about it is where the conference's two days diverge most. Supervisors broadly agree that opacity is the first problem: a large share of AI-linked financing runs through private credit vehicles and off-balance-sheet structures that do not show up cleanly in bank capital ratios.

Beyond that, consensus thins out fast. Academics at the conference are weighing whether macroprudential tools built for bank-centered crises can do anything against what some call mechanical amplification: the tendency of leveraged funds to sell into falling markets, pushing prices down further and forcing still more selling.

Whether credit rating agencies have fully priced the risk of AI debt whose maturity outlasts the hardware it funds is unresolved. November figures from the rating agencies are not out yet, and the research circulated ahead of the conference does not settle the question either way.

A further question raised in conference materials, and not yet answered on the record by any speaker, is how a shift toward agentic AI handling routine banking tasks might affect deposit stickiness, and with it the funding stability that commercial banks have relied on for decades. If customers move money at the instruction of software rather than habit, deposits that banks have always treated as stable could prove far more mobile than models assume.

A technician walks a server aisle inside a data center financed partly through corporate debt.
A technician walks a server aisle inside a data center financed partly through corporate debt.Illustration · MarketPedia

The conference runs through Friday, November 20, with sessions continuing in Cleveland.

No central bank has announced new rules on AI-linked leverage. The ECB's Financial Stability Review carries no policy recommendation beyond flagging the risk, and the IMF's October report stops at description rather than prescription.

The argument in Cleveland has moved past whether AI investment is too large. The open question, still unresolved after two days of sessions, is whether the debt behind it was built to survive a disappointing quarter.

Sources

More from the Newsroom

We (www.marketpedia.eu) and selected third parties (1) collect personal information as specified in the privacy policy and use cookies or similar technologies for technical purposes and, with your consent, for measurement and “marketing (personalized ads)” as specified in the cookie policy.

With respect to advertising, we and 1237 selected , may use precise geolocation data, and identification through device scanning in order to store and/or access information on a device and process personal data like your usage data for the following : personalised advertising and content, advertising and content measurement, audience research and services development.

You can freely give, deny, or withdraw your consent at any time by accessing the preferences panel. Your choices apply only to this domain. Denying consent may make related features unavailable.

Use the “Accept” button to consent. Use the “Reject” button to continue without accepting.