Rate Brief ENDE

Fed's Barr: AI firms' off-balance-sheet obligations now exceed $3 trillion

Federal Reserve Governor Michael Barr spoke on 16 November in Milan, at the Association for Computing Machinery's conference on AI in finance, about financial stability risks from artificial intelligence, drawing on the Fed's Financial Stability Report issued the previous Friday. He said he is an optimist about AI's long-term promise but that it can only strengthen the financial system if it also supports financial stability.

On the investment boom, Barr said the report looks at what it calls "advanced computing debt" through three channels: the rapid growth of that debt, circular financing networks and other complex obligations. On-balance-sheet debt of the seven largest AI firms has grown significantly in 2026 and is expected to exceed $800 billion before year-end. Their other obligations - long-term contracts, lease commitments and residual value guarantees - have grown faster, and the report estimates the potential maximum nominal value of these off-balance-sheet liabilities at more than $3 trillion. Hyperscalers and large GPU producers also hold over $100 billion of equity investments in AI labs and cloud start-ups, which then use the money to rent or buy compute from the investors. He also described collateral guarantees, in which larger firms guarantee the resale value of data centres or GPUs backing asset-based financings.

Barr said that, given the still-limited debt behind the boom, he is less concerned about direct losses. He flagged that GPUs may depreciate quickly as newer chips overtake them, although lease rates for older chips appear to be holding up, and that substations and grid connections may take years to be used if demand disappoints.

On trading, he said generative AI is not yet used for autonomous trading at large scale, as he understands it, but that reliance on a small set of similar models trained on similar data risks herding and synchronised reactions, and that a flaw shared by many models could have systemic consequences. He cited Markus Brunnermeier's 2026 Jackson Hole paper on prices becoming less informative when AI agents understand humans better than humans understand them.

Fed's Barr: AI firms' off-balance-sheet obligations now exceed $3 trillion
Fed's Barr: AI firms' off-balance-sheet obligations now exceed $3 trillion — Rate Brief

Why it matters

The figures Barr cited put the bigger exposure off the balance sheet: more than $3 trillion of potential obligations against roughly $800 billion of reported debt, much of it guarantees by strong-credit firms for weaker ones. That is the part that would transmit stress if AI investments were repriced.

Written by Victoria Shinder.