Investors Turn Cautious as AI-Related Debt Shows Signs of Distress
Bond prices for data-center financing and neocloud firms have fallen, while credit-default-swap activity has surged, indicating growing investor skepticism about the sustainability of the AI boom.
Bond yields for financing data-center construction and for "neocloud" companies that supply AI computing power have risen sharply, pushing several issues into distressed territory. CoreWeave, the largest neocloud provider, recently saw its bonds classified as distressed, and nearly half of the bonds issued this year to fund data-center projects by major tech firms have been rated junk.
At the same time, trading volume for credit-default swaps linked to AI firms has more than quadrupled between July and September, a market signal that investors are seeking insurance against potential defaults.
Analysts cite a combination of factors behind the shift, including the sector’s current lack of profitability and projections that U.S. consumers and businesses would need to generate about $3.5 trillion in AI-related revenue by 2032 – roughly 9 percent of GDP – to justify the capital outlays. A Brookings Institution paper notes that this scale of spending would exceed historic infrastructure booms, such as the railroad expansion.
Recent corporate developments have reinforced concerns. Character.ai filed for bankruptcy after reporting monthly losses of $8-9 million, while OpenAI postponed its planned initial public offering amid doubts about investor demand. Anthropic’s prospectus disclosed an $80 billion loss for the year and escalating infrastructure costs.
Environmental opposition to new data-center construction has also intensified, with activists blocking projects worth an estimated $198 billion in the first half of the year. Restrictions on data-center siting could further limit the sector’s ability to meet its revenue targets.
Economists warn that the financing model for AI projects resembles the subprime mortgage market that preceded the 2008 financial crisis, with lenders relying more on the sponsor’s capacity to absorb liabilities than on the underlying economics of the projects.
While some experts argue that a full-blown crisis is not inevitable, the convergence of high-risk debt, mounting losses, and regulatory uncertainty suggests that investors are preparing for a potential correction in the AI market.