AI's $800bn Bet Is Starting to Crack
Big Tech committed $800bn to AI in 2026. Now earnings are missing, stocks are sliding, and investors want their money back.
By The energy-climate Desk
For two consecutive days in late October, tech stocks fell in unison. The Nasdaq ended October with its worst monthly slide since March 2025. [2]
The numbers behind the rout are stark. Meta, Alphabet, Oracle, Amazon, and Microsoft have collectively committed to spending $800bn on AI in 2026. [2]
Intel's position looks more precarious still. On Thursday, Intel's finance chief David Zinsner said the firm did not expect to bring in more cash than it spent this year, and forecast lower-than-expected revenue for its fourth quarter. The chip maker said it planned to increase spending on AI despite weaker demand for computers — the same computers that have historically driven demand for its processors. [2]
"Investors have turned ruthless, punishing any signs of disappointment in [third quarter] earnings," she said. "The AI trade has had a good run in recent years, but now investors want payback, and companies like Alphabet and Intel are in the crosshairs." [2]
There is a familiar shape to this kind of market reckoning: ambition funded on promise, then tested by the slow grind of quarterly reality. The question now is whether what is being built is worth what is being spent — and whether the answer will arrive before the capital does.
That anxiety is not limited to Wall Street. The agentic web — the emerging era in which AI agents triage inboxes, handle grocery orders, and access bank accounts on a user's behalf — is drawing on the same infrastructure budgets now under scrutiny. [1] As MIT Technology Review notes, whether the human web survives this transition "is anyone's guess." [1]
The AI bubble debate has been simmering for months. Chris Low, from investment management firm FHN Financial, told Reuters the companies were "showing some pretty significant growing pains as they aggressively pursue the AI build-out." [2] He added: "The thinking is that these signs suggest that the entire sector is going to slow down." [2]
For context: despite the October selloff, the Nasdaq remains up 7.8% over the past year. Amazon is up nearly 8% over the past 12 months. Alphabet is down 1.3%. [2] The damage so far is a correction, not a collapse — though Kevin Roose, a technology columnist at the New York Times, told NPR that an AI investment crash would be preferable to the technology permanently upending work and society. [1]
Whether the spending ultimately justifies itself depends on a question the market has always been bad at answering in advance: when does infrastructure become indispensable, and when does it simply become expensive? Data centres, like railways before them, tend to answer that question slowly and unevenly.
[1] The selloff may not mark the end of the AI investment supercycle — but it signals that the era of spending first and asking questions later is meeting its first serious resistance. The next earnings season will tell us whether this was a stumble or a turning point.
It would be great if all this was just a house of cards.Kevin Roose