Man Group Warns of 'Violent' Correction Risk in AI Credit Markets

Global asset manager Man Group is issuing a stark warning that while artificial intelligence (AI) technology is among the most significant technological advancements of the last century, the financial architecture supporting it is expanding at an unsustainable rate. The firm states that AI-related credit markets face the risk of a "violent" correction, urging investors to remain clear-eyed in this environment. According to Man Group, the AI industry has been built for a demand curve that may not materialize or whose economics may deteriorate even if it does.

Man Group's analysis suggests that while the AI boom is real, the financial structure built around it appears to be expanding more quickly than any credible adoption curve can justify. This phenomenon is likened to historical technological revolutions such as railroads, electrification, radio, and the dot-com era, where the technology itself endured, but the financing cycle broke as expectations outpaced the industry's ability to meet them. The current debt-fueled capital expenditure (capex) cycle, particularly for data centers, is described as one of the most intensive in modern business history. Private credit lenders are modeling AI infrastructure investments as 10-to-20-year assets, comparable to commercial real estate or telecom towers. However, GPU technology fundamentally changes every 12-18 months, rendering data center lifetime assumptions deeply incorrect. This leads to rapidly falling asset values, illusory collateral values, and fragile cash flow assumptions.

AI-related exposure is increasingly metastasizing throughout the economy and across the alternative investment landscape. Institutional portfolios now have direct and indirect exposure to AI, making it a portfolio-wide factor that affects equity beta, growth exposure, infrastructure demand, private-market valuations, energy requirements, and credit underwriting. A reversal in the AI narrative could ripple across multiple asset classes. While bond issuance in the AI and hyperscaler space has surged in public markets, a series of high-profile defaults in private credit has placed AI and software deals under scrutiny. Major technology firms like Meta Platforms (META), Alphabet (GOOGL), and Oracle (ORCL) have notably issued billions in bonds to fund their AI infrastructure.

In the broader economic context of this financial expansion, markets are reportedly driven more by a "fear of missing out" (FOMO) than by actual fear, with investors focusing on near-term gains rather than long-term risks. Man Group highlights that AI's spread through the economy constitutes a "quiet migration of risk," which is a distinctive feature of this bubble. Furthermore, Morgan Stanley estimates that approximately $1.8 trillion in hyperscaler liabilities are being kept off-balance-sheet. The cautious approach of memory and semiconductor manufacturers, who are limiting capacity expansion despite apparent insatiable demand, is also seen as a significant market signal regarding the sustainability of AI capex projections. Environmental factors, such as the electrical grid failing to keep pace with AI data center construction and the demand for freshwater for cooling, are also leading to delayed or canceled projects.

Analysts and market expectations suggest that investors must adopt a disciplined approach in this environment. Man Group advises rigorous credit selection, a clear-eyed view of borrowers, and portfolios diversified enough not to be held hostage to the AI story playing out as expected, rather than broad beta exposure. Experts emphasize the critical need to differentiate between businesses deploying AI to enhance their competitive position and those whose core products are being displaced by it. This indicates a period where markets will become more selective, and true winners will emerge in a less forgiving environment.