They have spent a lot of it on the Nvidia NVDA -1.37%decrease; red down pointing triangle chips that serve as the computational workhorses of the boom. Before the past few days, this had lifted Nvidia’s shares to unprecedented heights; it was briefly the largest listed company in the world.

Sustaining the boom hinges on the tech companies’ willingness to stick with current and future spending against a backdrop where recession appears increasingly likely.

That is a lot to ask. Probably too much.

Some of the businesses that give Big Tech its big pockets aren’t exactly recession-proof. Meta is almost entirely funded by ad sales, an area that could come under threat with higher tariffs that send prices up and make consumers wary. Google also is ad-dependent: About three-quarters of its revenue came from ads last year.

Tech companies were already pretty far over their skis with spending on AI before tariffs came into the picture. They were set to lay out some $325 billion next year in capital expenditures on data centers, according to a Citi estimate in February.

That spending wouldn’t necessarily be unsustainable if it weren’t for a troublesome fact—AI is still trying to find its own legs as a business. So far, AI hasn’t been enormously profitable for anyone—at least not commensurate with the investment it requires. 

John Blackledge, a tech analyst at TD Cowen, said Amazon’s cloud-computing arm historically has generated $4 of incremental revenue for every $1 of capital spending. With investments in generative AI, the ratio is currently something like 20 cents for every dollar—although Blackledge estimates it will get closer to the usual $4 return in the next several years.

Given the promise AI holds, there is an argument that tech companies will take tariffs as a cue to spend more on AI while dialing back other parts of their business. Some analysts believe they will stick to their big spending plans.

The more probable scenario is that at least some of them dial back. They will likely use tariffs as an excuse to moderate spending. In many ways, the spending boom had been driven by not wanting to fall behind competitors who are also spending massively on the technology.

In what may be a foretaste of things to come, Microsoft is already slowing data-center construction across the globe, including a $1 billion project in Ohio. The company is sticking with plans to spend more than $80 billion on infrastructure this fiscal year, but the moves signal a more cautious stance in the longer term. Analysts at TD Cowen said in a note last month that Microsoft canceled leases in the U.S. and Europe, which the analysts attributed partly to an oversupply of data-center space relative to the company’s demand forecasts.

Tech companies have shown they can adjust quickly to changed circumstances. They did so a few years ago when Covid hit: Google delayed ad launches and slowed hiring, and Meta pledged to moderate spending growth.

That crisis was also lined with opportunity in a way this one isn’t. With more people learning and working from home, demand back then rose for a lot of the services tech companies supplied. There is no such opportunity when prices are on the upswing and people rein in spending.

Meta is perhaps most sensitive to economic weakness of any of the tech giants, given its relative lack of revenue streams beyond ads.

Microsoft, Amazon and to some extent Google may be more resilient. Those three companies have big cloud-computing divisions that largely serve corporate customers. There, the return on AI investments is more straightforward, even if it is no slam dunk. That said, Google is still largely ad-driven and Amazon is exposed to consumer weakness in its e-commerce and ads businesses.

The other big casualty in a tariff-induced AI slump would be Nvidia, whose chief executive, Jensen Huang, was touting near-limitless spending on AI data centers just last month. “We’ve got to go and work with the supply chain upstream and downstream, to prepare the world for hundreds of billions of dollars, working towards trillions of dollars of AI infrastructure build-out,” he told analysts.

Such pronouncements were taken at face value by investors eager to cash in on the next big thing. Now, in the blink of Trump’s eye, they may prove to be a pipe dream.