Let’s go back to the beginning of the year here for a moment. (I hear you and your 401(k) eagerly nodding yes.) Back then, all the big tech companies, including Meta, Amazon, Alphabet and Microsoft, were still planning on spending unfathomable gobs of money on artificial intelligence, and specifically unfathomable gobs of money on AI data centers. Just how big of an unfathomable gob? $320 billion, in 2025 alone.
Fast forward to now, and all the uncertainty we’re seeing touch just about every aspect of this economy is touching the future of this economy too.
In 2023, Jeff Brown, CEO of the real estate investment firm T2 Capital Management, invested in the construction of four Midwest data centers. Three in Iowa, one in Illinois.
These weren’t data centers built specifically for AI, but they use a lot of the same materials — materials that often come from overseas.
“I mean on the outside, they’re just concrete shells, right? And so even concrete, we import so much from Canada, even China,” he said.
Not to mention all the imported steel structures and the cooling systems and computer chips you may find on the inside.
Brown said because of tariffs and other rising construction costs it would be significantly more expensive to build those data centers now.
“I think it’s very safe to say 20% or more,” he said.
Microsoft has said it’s pausing some of its data center construction, including a $1 billion project in Ohio. Amazon has reportedly slowed some of its data center leasing as well.
Neither company has publicly blamed higher costs or tariffs.
Angelo Zeno, senior vice president of equity research at CFRA Research, said he expects big tech companies to still spend big on AI data centers this year. But:
“If we were to see significant deterioration within the macro economy we think there could be a reevaluation or a scale back,” he said.
If a recession does come, tech companies may have less revenue to steer toward AI. And billion-dollar data centers may be a tougher sell if adopting AI technology comes slower than what investors expect.
“It doesn’t turn on like a light switch overnight. It’s going to be gradual, and companies are trying to figure out what AI projects they use,” said Brent Thill, tech sector leader at Jefferies.
Thill said right now most non-tech companies are still in the experimental phase with AI. And they’re dealing with their own uncertainties, too.
