The biggest risk of AI is not a bubble, but "borrowing money and racing ahead"! The IMF issues a warning.

```

The biggest worry in the current AI boom may not be valuations, but the way funds are being raised.

On June 30th, Tobias Adrian, Director of the IMF's Monetary and Capital Markets Department, stated that the valuation of AI-related stocks hadn’t necessarily reached bubble levels. What truly warrants the attention of financial regulators is that global tech giants are increasingly financing through medium and long-term debt to invest heavily in rapidly evolving AI infrastructure. This mismatch between asset and liability maturity is the real potential source of future financial stability risks.

Speaking at the European Central Bank’s annual forum in Sintra, Portugal on Tuesday, Adrian said the model can be sustained as long as AI business profits keep growing. But if future AI profits fall short of market expectations, companies’ debt repayment ability could quickly come under pressure—this is the real issue to focus on.

His remarks also responded to recent debates about an "AI bubble." Previously, the Bank for International Settlements (BIS) just named AI as one of the four key risks threatening global financial stability, and ECB Executive Board member Isabel Schnabel opened the forum by questioning whether there is a bubble driven by AI.

IMF: Valuations not as dangerous as imagined

Adrian believes there are still distinct differences between the current market and traditional bubbles.

He said that determining whether stocks are in a bubble fundamentally depends on whether valuations are disconnected from earnings, and the recent market correction has eased earlier valuation pressures. On the one hand, AI-concept stocks have pulled back, and prices have fallen; on the other hand, corporate earnings keep beating expectations, reducing price-to-earnings ratio pressures.

Although investors previously held "extremely aggressive" expectations for AI commercialization returns, actual corporate profits continue to surprise the market, meaning the current rally is not entirely detached from fundamentals.

Adrian also pointed out that the typical "rising tide lifts all boats" phenomenon seen in bubbles has not appeared within the market.

Over the past year, what has really surged are the foundational hardware firms such as those making AI chips, while the software sector has seen notable corrections. If the market had entered bubble territory, investors would indiscriminately chase all AI-concept stocks, not this clear structural differentiation. In his view, this shows that the market is still pricing based on profitability.

The real issue: AI investments are "borrowing long, buying short life assets"

Adrian is more concerned about changes to tech giants’ financing structures than rising stock prices.

He said that large cloud computing companies (Hyperscalers) are increasing leverage by issuing medium and long-term bonds to raise funds, buying AI chips and building data centers on a large scale. The issue is that the lifespan of these AI infrastructure assets is far shorter than the maturity of their debts. Especially for GPUs and other AI chips, which update and iterate rapidly—a generation may become outdated or obsolete within a few years, while bonds issued by companies often mature over several years to ten years or more.

This means companies are essentially using long-term debt to invest in assets that depreciate rapidly, causing a classic maturity mismatch.

Adrian believes that as long as AI models keep generating profits, and corporate customers and consumers are willing to pay for cutting-edge models, this funding model won’t present problems. But if AI commercialization falls short of expectations and profit levels decline, companies’ cash flow may not cover long-term debt costs—then financial stability risks could emerge quickly. He said:

"The truly crucial issue is whether AI can sustainably generate profits, not how high stock prices are today."

Risk DisclaimerThe market involves risks; investments require caution. This article does not constitute personal investment advice and has not considered the special investment objectives, financial situation, or needs of individual users. Users should consider whether any opinions, views, or conclusions in this article meet their specific circumstances. Investing based on this is at your own risk. ```