DoubleLine’s Robert Cohen warns AI debt will reach bubble levels

Artificial intelligence spending is increasingly becoming a credit-market story, and Robert Cohen of DoubleLine is warning that the wave of borrowing tied to AI could eventually look like a bubble. Speaking at Bloomberg’s Global Credit Forum in New York, Cohen said AI-related debt will “almost certainly” reach bubble levels, even as companies continue to tap both debt and equity markets to fund growth.

Cohen’s comments fit into a broader cautionary view about how much credit is flowing into the AI buildout. According to Bloomberg’s reporting on his remarks, he argued that the demand for debt issuance remains strong even as corporate credit supply has shrunk since the pandemic, creating conditions that could encourage overleveraging if the AI boom slows or disappoints.

In a separate Bloomberg interview, Cohen also said that the current fundraising environment remains healthy for many companies, especially higher-quality borrowers with strong balance sheets and steady cash flow. That distinction matters because it suggests he is not calling for a broad collapse in credit, but rather warning that the risk is rising in the more speculative corners of the market.

His concerns extend beyond AI itself to the structure of private credit. Cohen said the private credit market is heavily weighted toward lower-rated borrowers, with about 90% of it in the B3-or-lower range, according to a YouTube excerpt of his remarks. He described those as risky credits and said that any economic slowdown, or any disruption from AI-related competition, could put those loans under pressure.

The warnings come as other prominent investors are also flagging possible excesses around AI. Bloomberg reported separately that Bridgewater founder Ray Dalio sees an AI bubble that could burst, underscoring growing skepticism among some market veterans about how long the enthusiasm can last. For lenders, the issue is not just whether AI technology succeeds, but whether the companies funding it can continue servicing the debt taken on to build data centers, chips, and related infrastructure.

For investors in credit markets, the key question now is where the risks are accumulating. Cohen’s message, as reported by Bloomberg, is that the current environment still supports borrowing for strong issuers, but the combination of aggressive capital spending, heavy debt funding, and concentration in lower-quality private credit could leave parts of the market vulnerable if growth cools or expected AI returns take longer to arrive.