Federal Reserve Governor Michael Barr said Monday that the debt and other obligations financing the artificial intelligence data center buildout, many of them kept off company balance sheets, could amplify a sell-off in AI investments, and that these off-balance-sheet structures warrant close monitoring.
Speaking at an AI-in-finance conference in Milan at 5 a.m. Eastern time, Barr said the on-balance-sheet debt of the seven largest AI firms has grown significantly this year and is expected to exceed $800 billion before year-end. Their other obligations, such as long-term leases, contracts and guarantees on the resale value of data centers and chips, have grown even faster: the Fed’s Financial Stability Report estimates their potential maximum value now exceeds $3 trillion. The report covers Amazon, Alphabet, Broadcom, Meta Platforms, Microsoft, Nvidia and Oracle, and notes that these obligations are counted before discounting, so they are not directly comparable with balance-sheet debt.
Barr drew on the Fed’s Financial Stability Report, released Friday, which focused on the AI investment boom. He said the biggest cloud companies and chipmakers also hold more than $100 billion in equity stakes in AI labs and cloud start-ups that then rent or buy computing power from those same investors, a pattern he called “circular financing.”
He said he is less concerned about direct losses because a still-limited amount of debt backs the boom, but that high leverage at hedge funds could turn a sharp reversal into rapid selling. Barr added that Fed policymakers do not target market levels, and he avoids commenting on market moves.
Barr also warned that traders relying on similar AI models could move in lockstep, and that AI is raising cyber risks. The speech did not address interest rates. Fed Chair Kevin Warsh is scheduled to speak at noon Eastern on Monday.
…
That’s the story. … Thank you for listening to Cedar Global Markets. We’re supported by our partners, and you can learn more about the partner featured on this page. We’ll see you next time.




