Gartner Says $3.5 Trillion in AI Infrastructure Spending Could Be at Risk

what happened

Research firm Gartner stood by its forecast that up to $3.5 trillion in artificial intelligence infrastructure investment could be in jeopardy. Chief forecaster John Lovelock told CNBC on Monday that current industry returns do not justify the scale of capital being deployed. He expects most of the neocloud market to disappear by 2030, with the frontier model business narrowing to a duopoly of Anthropic and OpenAI.

Anthropic's quarterly revenue growth slowed to 45% in the third quarter from 140% in the second, and its $16.8 billion quarterly revenue is small next to the more than $5 trillion hyperscalers are projected to spend through 2030. Lovelock said computing capacity currently exceeds demand, and many services hand out tokens for free.

In the week ended October 16, the S&P 500 lost 2%, the Nasdaq Composite fell 4%, and the Dow Jones Industrial Average gained 0.7%. CoreWeave (CRWV) dropped more than 25% over the week, while IREN, Nebius and Applied Digital closed higher on Monday.

why it matters

The main market risk is a rethinking of companies building AI infrastructure. Capital spending is rising faster than revenue, and compute providers face possible asset impairments. The pressure falls mainly on technology stocks in the Nasdaq and on neocloud shares rather than the broader market.

market impact

Nasdaq and AI infrastructure: doubts about returns weigh on tech stocks and neocloud providers

instruments