Fitch reports that an AI correction has become a significant global credit risk.

      Fitch Ratings has encapsulated several underlying market concerns in a single statement, cautioning that a potential correction in the AI market is emerging as one of the most significant credit risks to the global economy. This assessment is part of the agency’s Global Risk Outlook for the third quarter, released this week, and it coincides with a time when the investment in artificial intelligence has escalated to such an extent that any setbacks would impact sectors beyond just technology.

      The essence of the concern revolves around interconnectedness rather than the financial state of any particular company. Over the past year, equity markets, corporate bond issuance, and even overall economic growth have become increasingly reliant on AI, so a reevaluation of long-term returns would not remain isolated.

      “The magnitude of investment in AI is such that the economy’s and the capital market's exposure to a potential correction is considerable,” the agency stated. It added that the intertwining of capital markets and economies with AI has “created a vulnerability for credit.”

      This concern has been circulating among policymakers for some months, and it resonates with warnings from the BIS that a downturn in AI could impact credit markets as severely as the crisis of 2008.

      Fitch highlights specific figures to underscore this concentration. It mentions that the cyclically adjusted price-to-earnings ratio of the S&P 500 is nearing levels last observed during the late-1990s dotcom bubble, and that US corporate bond issuance surged by 26% in the first half of 2026. Much of this borrowing is attributed to a small number of companies, with Amazon, Alphabet, Nvidia, Meta, Oracle, and SpaceX collectively issuing approximately $182 billion in investment-grade bonds, a trend that has extended Big Tech's AI debt into European markets.

      With spending showing no signs of diminishing, the exposure continues to rise. Capital expenditures for Alphabet, Amazon, Meta, and Microsoft are expected to increase by about 75% to around $700 billion annually, a rate that currently exceeds the cash generated by these businesses.

      IT investments alone contributed 1.4 percentage points to US GDP growth in the first quarter, reflecting the same interconnectedness in another context. When investments on such a scale are increasingly financed through debt, the question of whether AI revenues materialize in a timely manner evolves from a purely equity-market issue into a credit concern, which is precisely the point at hand.

      This strain is not merely theoretical. Earlier this month, S&P downgraded Oracle to BBB-, just one level above junk status, as its data center expansion consumed cash rapidly, and Fitch's outlook suggests that the downgrade of Oracle may be less of an anomaly and more of an early indication of a broader trend. However, the agency underscores that a moderate decline would be manageable, but a more significant, prolonged correction could have “broader market, macro, and credit implications depending on its magnitude, duration, and spread.”

      Fitch’s caution also extends to the real economy. It has pointed out that the efficiency gains from AI might weaken job markets and diminish tax revenues in developed nations, even as the technology enhances productivity. The agency has also noted private credit as an area to monitor, concluding that this segment is unlikely to pose a systemic risk by itself.

      Across its recent analyses, a consistent thread emerges, as the advantages of AI for corporate credit remain largely incremental and difficult to quantify, while the potential downsides are becoming easier to project.

      This does not imply that a correction is imminent, nor does Fitch predict that it will happen. What the agency conveys is a burgeoning exposure, indicating that a considerable portion of the market's recent strength hinges on a singular bet yielding positive results. Uncertain revenues, historically high valuations, and unprecedented borrowing coexist, and whether this proves to be cautious foresight or an undervalued risk ultimately depends, as Fitch acknowledges, on future returns that are yet to be determined.