IMF warns AI companies' debt is a bigger danger than overvalued stocks

Tobias Adrian, financial counsellor and director of the International Monetary Fund's Monetary and Capital Markets Department, said that the growth of debt financing in the artificial intelligence industry poses a more serious threat to financial stability than inflated valuations of tech company stocks.

Why debt worries the IMF more than stocks

For the past two years, talk of an "AI bubble" has mostly centered on the share prices of companies like Nvidia, Microsoft, Meta, and OpenAI's partners: if expectations for AI revenue don't pan out, those stocks could fall sharply. Adrian shifts the focus to the other side of the balance sheet — debt. In his view, what matters more for the financial system is not what the stocks are worth, but who borrowed money to build AI infrastructure, and on what terms.

The distinction is fundamental. A stock decline erodes shareholders' equity but doesn't necessarily trigger a chain reaction. Delinquent or restructured debt is a different story: it hits the banks, bond funds, and insurers that hold those obligations on their balance sheets, and it can spread through the system much faster.

The IMF regularly publishes global financial stability reports that assess risks building up across market segments, from sovereign debt to corporate lending. The fact that "AI infrastructure debt" has become a distinct topic in the rhetoric of the fund's senior officials suggests that the scale of this financing is already large enough to appear in such analysis alongside more traditional risk categories.

How the AI boom is financed with debt

Building data centers, buying Nvidia accelerators, and renting compute from cloud providers all require enormous capital expenditures. In recent years, some large tech companies and specialized data center operators have increasingly turned not only to their own profits but also to bonds, bank loans, and structured deals with private lenders to finance this buildout.

This model — capex paid for with debt rather than free cash flow — is exactly what worries IMF officials and other financial regulators. If demand for AI services or revenue growth falls short of the forecasts built into credit models, servicing that debt could become a problem not only for the tech companies themselves but also for their lenders.

How this differs from the classic dot-com bubble

The comparison to the dot-com bubble of the late 1990s, often invoked in discussions of AI, only partly holds. Back then, it was primarily stocks that were overvalued: companies raised capital by issuing shares, and the crash hit equity investors first and foremost. Today's AI buildout is largely financed with borrowed money, and the risk is held not only by shareholders but also by banks, pension funds, and insurance companies that bought the corresponding bonds. That is a structurally different channel for problems to spread if the cycle turns, which is why IMF officials insist that the debt side of the AI boom be monitored separately from the stock market.

What it means

The IMF's stance adds a new layer of regulatory scrutiny to the AI sector: what matters now is not just stock multiples but also the quality of the debt financing the data center boom. For investors and regulators, it's a signal to watch the debt loads of hyperscalers and data center operators as closely as their share prices.