Chris Wood, Global Head of Equity Strategy at Jefferies, warned in his latest quarterly report that the AI industry is facing structural risks similar to the 2008 subprime mortgage crisis. He pointed out that hyperscale cloud providers are extending massive "unsecured credit" to continuously loss-making AI labs (OpenAI and Anthropic), which could lead to significant capital destruction in the US, with market share shifting to cheaper Chinese open-source large language models.

Wood's analysis indicates that approximately half of the $2.34 trillion in remaining performance obligations for the four major cloud platforms—Microsoft, Google, Amazon, and Oracle—as of the end of Q2 2026, is owed by OpenAI and Anthropic. Furthermore, the five largest cloud providers' off-balance sheet liabilities (data center lease commitments and implicit debt) have reached trillions of dollars, and they have issued $229 billion in bonds this year, competing with the federal government for financing. Wood also questioned whether the cloud providers' profit growth primarily stems from mark-to-market gains on equity stakes in AI startups, creating a circular financial engineering. He believes that the AI narrative "will experience a major reset at some point in the not-too-distant future" and advises investors to focus on credit markets rather than equity markets, holding "picks and shovels" (chip manufacturers) instead of the hyperscale cloud providers themselves.