AI bubble concerns intensify ahead of earnings season

According to Ray Dalio, the AI bubble could burst soon due to rising interest rates and the heavy borrowing by tech firms for AI infrastructure.

Is a dot-com-style collapse on the horizon?

The outlook for the AI sector appears increasingly cloudy. OpenAI posted annualized revenue of about $50 billion at end-September, falling $20 billion short of the $70 billion forecast. Worries that AI demand is decelerating and that revenue won't support heavy infrastructure spending have put pressure on shares of Oracle, Microsoft, and Nvidia, as well as the Nasdaq-100 index.

This week, selling pressure extended to Asian semiconductor and technology stocks. Progress by Chinese chipmaker CXMT in developing a new memory architecture has sparked fears of heightened competition for global memory producers. Additionally, Broadcom's chief executive likened AI investors to casino gamblers, implying that some lack a full understanding of their bets, and cautioned that financing AI infrastructure might become a significant hurdle.

Going forward, elevated Treasury yields could intensify the situation by reducing the appeal of riskier AI stocks for investors.

Nevertheless, this does not resemble a classic bubble because the large infrastructure investments are supported by genuine increases in revenue, orders, and contract backlogs. For instance, Microsoft's Azure has crossed $100 billion in yearly revenue, Amazon Web Services recorded its fastest growth in years, and Celestica and AMD are also experiencing robust demand for AI infrastructure.

In contrast to a typical bubble, companies are not speculating on future demand that might not appear. Instead, they are striving to meet demand that already surpasses current capacity. Meanwhile, major tech firms continue to be highly profitable and possess sufficient cash to finance record spending without endangering their financial health.

Yet risks persist because these large investments have not yet generated returns, and lofty valuations—over $2 trillion for the upcoming Anthropic IPO—make the sector susceptible to growth disappointments.

Earnings season will provide the next major test for the AI sector. This week, alongside JPMorgan, Goldman Sachs, Wells Fargo, Citigroup, Bank of America, Morgan Stanley, and BlackRock, ASML and TSMC are also reporting results, which should offer clearer indications of whether the AI boom continues.