Debate over whether there is a bubble in the artificial intelligence sector is intensifying. The bond market and the stock market appear to hold different views on the sustainability of AI’s prospects.
Significant stress signals from credit default swap (CDS) pricing show that the bond market is concerned about the debt situation of companies like Oracle (ORCL). Its CDS prices surged in November last year and are currently near four times the level from last September, indicating that bond market worries about Oracle’s debt issues remain persistently high.
These concerns primarily stem from the sharply rising costs of building data center infrastructure to support AI growth, particularly the massive $300 billion partnership agreement between Oracle and OpenAI. According to reports, internal forecasts from OpenAI project it will burn through over $100 billion in cash before it begins generating cash flow starting in 2030. While this is the company’s own projection, the market remains skeptical.
However, upon closer examination, the surge in CDS prices appears to have a limited scope of impact, and the stock market seems to agree. Among the major cloud computing giants, the relatively weak stock performance of Oracle and Microsoft is no coincidence—Microsoft recently disclosed that 45% of its backlog comes from OpenAI-related demand. In contrast, the outperforming Alphabet (GOOG) has minimal exposure to OpenAI, and its financial position is better able to support its AI-related capital expenditures.
Historical experience shows that during technological revolutions, companies often overextend themselves, and capital inevitably flows, driven by a herd mentality, into areas that ultimately prove inefficient. This is an inherent characteristic of human behavior and market cycles. However, predicting the timing of a bubble burst is extremely difficult: U.S. housing prices peaked in July 2006, about two years before the 2008-2009 financial crisis, and today’s prices are far above those levels. Selling property during the crisis proved unwise in the long run.
If bond and stock markets become more cautious about investing in AI-related companies—favoring high-quality firms like Alphabet—this could curb some unrealistic development plans. Markets need to function through effective risk pricing, and this screening mechanism may lead to more sustainable long-term growth for AI.
