Risk of an AI crunch for stock markets is rising

Skip to Content News Archives Economy Defence Energy Oil & Gas Renewables Electric Vehicles Mining Commodities Agriculture Real Estate Mortgages Mortgage Rates Finance Banking Insurance Fintech Cryptocurrency Defence Work Wealth Smart Money Wealth Management Investor Personal Finance Family Finance Retirement Taxes High Net Worth FP Comment Executive Women Puzzmo Newsletters Financial Times Business Essentials More Innovation Information Technology FP500 Podcasts Small Business Lives Told Tails Told Shopping Financial Post Store Obituaries Place a Notice Advertising Advertising With Us Advertising Solutions Postmedia Ad Manager Sponsorship Requests Classifieds Place a Classifieds ad Working Profile Settings My Subscriptions My Offers Newsletters Customer Service FAQ News Economy Energy Mining Real Estate Finance Defence Work Wealth Investor FP Comment Executive Women Puzzmo Newsletters Financial Times Business Essentials This advertisement has not loaded yet, but your article continues below.HomeFinancial TimesRisk of an AI crunch for stock markets is risingThe tremor seen this week could be a foretaste of things to comeAuthor of the article:Last updated 30 minutes ago Nearly four years after the launch of ChatGPT, the promised AI productivity miracle is still more promise than reality. Photo by Michael M. Santiago/Getty ImagesWhen the tech companies behind the boom in data centre spending start to report quarterly earnings next week, talk is likely to focus once again on capacity shortages and surging AI revenue. All the signs are that demand for raw AI power is still running well ahead of supply, adding to the pressure to lift capital spending even higher.THIS CONTENT IS RESERVED FOR SUBSCRIBERS ONLYSubscribe now to read the latest news in your city and across Canada.Exclusive articles from Barbara Shecter, Joe O'Connor, Gabriel Friedman, and others.Daily content from Financial Times, the world's leading global business publication.Unlimited online access to read articles from Financial Post, National Post and 15 news sites across Canada with one account.National Post ePaper, an electronic replica of the print edition to view on any device, share and comment on.Daily puzzles, including the New York Times Crossword.SUBSCRIBE TO UNLOCK MORE ARTICLESSubscribe now to read the latest news in your city and across Canada.Exclusive articles from Barbara Shecter, Joe O'Connor, Gabriel Friedman and others.Daily content from Financial Times, the world's leading global business publication.Unlimited online access to read articles from Financial Post, National Post and 15 news sites across Canada with one account.National Post ePaper, an electronic replica of the print edition to view on any device, share and comment on.Daily puzzles, including the New York Times Crossword.REGISTER / SIGN IN TO UNLOCK MORE ARTICLESCreate an account or sign in to continue with your reading experience.Access articles from across Canada with one account.Share your thoughts and join the conversation in the comments.Enjoy additional articles per month.Get email updates from your favourite authors.THIS ARTICLE IS FREE TO READ REGISTER TO UNLOCK.Create an account or sign in to continue with your reading experience.Access articles from across Canada with one accountShare your thoughts and join the conversation in the commentsEnjoy additional articles per monthGet email updates from your favourite authorsSign In or Create an AccountLook further out towards the horizon, though, and the effects of the projected surge in construction look more daunting. To take one recent estimate, from the Brookings Institution: Spending on the AI services produced by all those data centres would need to account for an astounding 9.2 per cent of United States GDP by 2032 for their operators to make a decent economic return.This advertisement has not loaded yet, but your article continues below.Between those two poles — acute shortages in the short term and the looming shadow of severe overcapacity — there could be a world of hurt for the stock market. The tremor that has gone through some parts of the AI sector this week, particularly AI cloud computing start-ups like CoreWeave and Nebius that are considered the most exposed to any slowdown, could be only a foretaste of what’s to come.Get the latest headlines, breaking news and columns.By signing up you consent to receive the above newsletter from Postmedia Network Inc.A welcome email is on its way. If you don't see it, please check your junk folder.The next issue of Top Stories will soon be in your inbox.We encountered an issue signing you up. Please try againStart with the forces that have been pushing the near-term capital spending plans ever higher. The five biggest hyperscalers — Alphabet, Amazon, Meta, Microsoft and Oracle — have raised their capex budgets throughout the year and currently expect to spend US$800 billion, nearly twice as much as last year. That is set to eclipse their combined forecast cash flow of around US$700 billion.One very visible measure of the red-hot demand driving this is the cost of renting GPUs in the spot market. A B200 chip from Nvidia currently costs more than US$6 an hour, around 50 per cent more than six months ago, according to data from SemiAnalysis.Another way to look at this is the payback period on new data centre investments. On its first earnings call as a public company, SpaceX told investors that the returns from its latest contracts are high enough to cover the entire cost of building the necessary computing capacity in less than a year. With data centre equipment depreciated over five to six years, that points to sky-high returns over the life of today’s servers and networking equipment.This advertisement has not loaded yet, but your article continues below.Left unsaid is what will happen when the extreme supply/demand imbalance eventually ends. Now consider the medium-term view. Even if many of the more speculative data centres on the drawing board are never built, according to the Brookings report, it will take an extraordinary jump in AI demand to justify the construction boom. By 2032, customers would need to spend US$3.7 trillion a year (the nine per cent of U.S. GDP).Tech research firm Gartner came up with a similar calculation this week, estimating that US$8.5 trillion of AI revenue will be needed between now and 2030 to produce decent investment returns on all the new data centres. The actual AI revenue Gartner forecasts over that period: US$3.6 trillion.It is possible to come up with more bullish forecasts to support the investment boom. If AI makes labour much more productive — or starts to replace workers en masse — then nine per cent of GDP might be more plausible. Nearly four years after the launch of ChatGPT, though, the promised AI productivity miracle is still more promise than reality.It’s also possible that the data centre boom could end in a softer landing than the current building plans imply. The political backlash against data centres or shortages of things like power and chips could slow construction and reduce the risk of overbuilding. Alternatively, if the capital markets start to choke on the debt needed to maintain the build-out, particularly as interest rates rise, some planned facilities would never see the light of day.This advertisement has not loaded yet.This advertisement has not loaded yet, but your article continues below.Yet even if forces like this kept capacity more constrained and AI prices higher, it would probably still have serious consequences for equity markets given the high growth expectations baked into AI stock prices.Not that market manias sparked by tech revolutions usually end in soft landings. There is a tendency to overshoot, as investors expect the upward momentum to continue and companies fight to keep a place for themselves in the new tech world that is emerging. That will be something to remember in the coming days as tech executives discuss the latest state of AI demand — and, quite possibly, ratchet up their near-term capital spending plans yet again.© 2026 The Financial Times LtdWe apologize, but this video has failed to load.Notice for the Postmedia NetworkThis website uses cookies to personalize your content (including ads), and allows us to analyze our traffic. Read more about cookies here. By continuing to use our site, you agree to our Terms of Use and Privacy Policy.