Robert Cohen warned that artificial intelligence debt is likely to reach bubble levels in credit markets during a panel this week [1].
The warning highlights a growing divide among financial experts regarding whether the rapid expansion of AI infrastructure is sustainable or a precursor to a market crash.
Cohen, the portfolio manager and director of global developed credit at DoubleLine, spoke at the Bloomberg Global Credit Forum in New York on June 3 [2]. He said that the current trajectory of AI financing mirrors previous eras of intense industrial expansion.
"Artificial intelligence debt will almost certainly reach bubble levels eventually, given the history of periods of heavy investments in areas like railroads and the internet," Cohen said [3].
While Cohen sees a historical pattern leading toward a bubble, other industry voices disagree with the assessment. Robert Schiffman, appearing in a Bloomberg video, said there is no imminent crisis.
"There is no AI bubble," Schiffman said [4].
Other analysts suggest a more cautious middle ground. Sophie Huynh said that the possibility of an AI bubble is "something to look at," suggesting the risk is a point of observation rather than a certainty [5].
The debate centers on whether the productivity gains from AI will materialize fast enough to service the debt used to build the technology. Cohen's perspective relies on the precedent that massive capital expenditures in new technology often lead to over-leveraging before the market stabilizes.
“"Artificial intelligence debt will almost certainly reach bubble levels eventually."”
This disagreement reflects a fundamental tension in the credit markets: the struggle to value future AI productivity against current borrowing costs. If Cohen's historical analogy holds, the market may be ignoring the risk of over-capacity, whereas Schiffman's view suggests that AI's utility is fundamentally different from the railroad or early internet booms.
