CoreWeave bond rout tests market’s math on AI buildout

CoreWeave’s bonds experienced a significant decline this week, prompting investors to question the market's willingness to finance the rapid expansion of AI infrastructure. The sharp drop in bond prices has raised concerns about broader risks tied to financing AI projects, which often take years to generate revenue. CoreWeave bonds, initially issued at a yield of 9.675% in June, have seen their yield rise to 15% in the current week.

Max Frumes, the global head of distressed and leveraged finance at 9fin, believes that CoreWeave serves as a "canary in the coal mine" for the sector, which is facing increased stress. He expects that other AI-linked debt will follow CoreWeave into distressed territory, leading to higher borrowing costs and more scrutiny from investors.

Portfolio managers are becoming more selective in their investments, scrutinizing their exposures in the AI infrastructure space. John Han, a portfolio manager at F/m Investments, noted that bonds with high yields, such as CoreWeave's, now warrant closer examination. Traditional sources of collateral and lower risk are increasingly preferred over high-yield securities.

Larry Holzenthaler, a senior portfolio manager at Catalyst funds, views the market's current reaction as an overreaction, anticipating a shift from optimism to bearish sentiment. He attributes the initial optimism to overlooking risks like the sheer size of funding needs and reliance on outside financing.

Deutsche Bank's global head of macro and thematic research, Jim Reid, anticipates a new phase where debt financing becomes increasingly crucial, given the growing emphasis on capital expenditures funded by debt rather than free cash flow. However, he warns that AI financing costs, while not dramatic overall, present worrying themes at the edges.

Vanguard bond portfolio manager Brad Collins highlighted the increasing challenge in assessing AI-related debt, as issuers introduce diverse offerings that require rapid decision-making. The rising inflation has pushed financing costs higher, adding to concerns about the future of AI capital spending plans. Additionally, political controversy surrounding data center construction further complicates the landscape.

Analysts estimate that 5% to 10% of planned projects may ultimately not be built due to permitting issues and power price concerns.