AI Investment Boom Compared to 2008 Housing Bubble

The author argues that the artificial intelligence investment boom increasingly resembles the housing bubble that preceded the 2008 financial crisis. The author describes AI as rapidly becoming the biggest investment in U.S. history. The author warns that substantial investment may fail and that the consequences could match those of 2008.

The author argues that too little is known about the financing supporting the boom. The author compares the known arrangements with practices used by Wall Street firms, including Lehman Brothers. The author cites the 158-year-old bank's collapse in 2008 as the trigger for a panic. The author puts the resulting taxpayer bailout cost at nearly $700 billion in today's dollars and the employment losses at nearly nine million Americans.

The author argues that companies are concealing substantial debt outside public balance sheets. The author says firms entrusted with pensions and insurance are risking customers' money on complex financial products backed by assets whose value is uncertain. The author argues that a small set of companies trade large quantities of a popular product among themselves, exaggerating demand and making systemic risks harder to assess. The author criticizes accountants and credit-rating firms as financially dependent on debtors and using weak justifications to endorse the arrangements.

The author argues that an AI crash could erase more wealth than the 2008 crisis because more U.S. funds are invested in AI-heavy stocks than in homes. The author says the Trump administration is introducing changes that increase opacity and risk. The author describes the administration as reportedly keen to avert a collapse.

The author argues that signs already suggest the emerging AI industry may not soon earn the trillions of dollars needed to repay the industry's enormous debts. The Financial Times reported earlier this month that OpenAI's projected annual revenue at September's end was about $50 billion, 28 percent below what had previously been believed. Figures published this month by The Wall Street Journal suggested slowing revenue growth at Anthropic, OpenAI's main competitor. The technology-heavy Nasdaq has declined more than 5 percent during the past two weeks.

The author argues that OpenAI and Anthropic are sufficiently intertwined with the economy and national security that Washington will need a plan if the companies fail. The author argues that a properly designed rescue would require meaningful safety protections, safeguards against surrendering control of the technology to China and a genuine commitment to the public interest.