Wall Street starts asking when the Artificial Intelligence boom will pay
Alphabet lost about $500bn of stock-market value in a day and Intel more than a quarter of its own, in a week when the biggest technology companies showed how fast their spending on AI is growing.
The biggest technology companies reported their results this week, and the numbers were, by most measures, very good. Alphabet's revenue grew 25% to $127.6bn in the three months to September;1 Microsoft's grew 19% to $92.3bn;2 Amazon's cloud business grew 39%.3
Investors sold anyway. On Thursday 29th October Alphabet, Google's parent company, closed down 12.6%, which took about $500bn off its stock-market value. Only one company has lost more in a single day: Nvidia, in January 2025.4 On Friday Intel fell 28.6%, and the Nasdaq, the index where most big technology companies are listed, had its largest one-day fall since 5th June.5
What changed was not whether the companies are making money. It was how much they are spending to keep making it, and how soon that spending pays.
Spending that has doubled
The boom in Artificial Intelligence runs on data centres: buildings full of specialised chips that train and run AI models. The companies building them report that spending as capital expenditure, or capex.
In the three months to September, Amazon spent $77.3bn on property and equipment, against $35.1bn a year earlier.3 Alphabet spent $51.4bn, against $24.0bn.1 Microsoft's additions came to $39.4bn, against $19.4bn.2 Across the three, spending has more than doubled in a year.
Meanwhile Alphabet's search revenue, its largest business, rose 12% to $63.35bn, slower than the 17% of the previous quarter.4 And Alphabet spent more cash than its businesses brought in: its free cash flow, what is left after investment, was minus $9.1bn for the quarter.1 Amazon's free cash flow over the past year was minus $42.4bn.3
Neither company is slowing down. Alphabet kept its forecast of $195bn to $205bn of capital spending this year.6 Amazon expects about $220bn this year,3 and its chief executive, Andy Jassy, said it would "spend meaningfully more in 2027".7 TechCrunch put the four biggest spenders' plans for 2026, including Meta's, at more than $740bn, and said Big Tech had borrowed $326bn this year, three times as much as last year.8 Analysts expect the total to pass $1tn next year.6
Intel's different problem
Intel's fall had a different cause. Its results beat forecasts, but it told investors to expect $15bn to $16bn of revenue this quarter, against the roughly $17.1bn analysts had pencilled in.9 Its chief executive, Lip-Bu Tan, said the company's "immediate priority is to unlock additional supply to meet extraordinary demand": Intel cannot make enough of the server chips that data centres want, and its finance chief said he expected the slowdown in personal computers to continue at least through the first half of next year.10
"Companies that are performing and meeting the very lofty expectations that have been set are doing well," said Oliver Pursche of Wealthspire Advisors.5 This week the expectations were the problem.
Anthropic goes public
The week began with a filing that showed the same arithmetic from the other side. On Monday 26th October Anthropic, one of the companies whose models the data centres are built for, published its plans to list its shares on the Nasdaq. Its revenue in the first half of 2026 was $16.3bn, up from $1.2bn a year earlier, and it made an operating loss of $693m. Since the end of 2025 it has committed to about $417bn of new spending on computing capacity and data centres.11
Its filing says: "We do not expect our revenue will continue to grow at comparable rates, and we cannot assure you that it will continue to grow at all."11 That is the standard caution of a company preparing to sell shares. It is also the question this week's investors were asking of everyone.
Anthropic makes Claude, the AI model Hindsight Media uses to research and draft each edition.
Where it stands
The companies say demand for computing is greater than they can supply, and their cloud businesses are growing fast enough to support that. Investors who sold this week were not saying AI does not work. They were asking whether revenue will catch up with spending that is growing faster, financed increasingly with debt. Whether this week was a pause or a turn depends on that.