How much AI data-center spending is projected through 2027?

Jessica Wachter and her collaborator estimate expenditures will reach nearly $1.1 trillion through 2027.

What productivity gain would hyperscalers need to break even?

They would need to increase their own productivity by a factor of 2.7 to break even by 2030, accounting for the cost of capital and a 15% return, and depreciation of the assets.

What happens if the hyperscalers cannot meet those profit goals?

Wachter says they will fall behind on their interest payments, which risks bankruptcy.

AI Stocks & MarketsAI Business & IndustryMIT Technology Review AISep 15, 2026Sep 15, 2026, 22:00 JST

Wachter: AI buildout needs 2.7× productivity to break even

Send on LINE

3 Key Points

What happened
Wharton's Jessica Wachter estimates hyperscaler AI data-center spending will reach nearly $1.1 trillion through 2027, requiring a 2.7× productivity gain to break even by 2030; she warns of bankruptcy risk and "the largest misallocation of capital in history."
Why it matters
Wachter frames the spending as an accounting problem, not a technology forecast, so the bet hinges on whether hyperscalers can generate returns large enough to match costs of capital and depreciation. She calls the required growth "a lot of growth compressed into a few years."
What to watch
The test is whether that productivity boom materializes; if it does not, Wachter says the hyperscalers will fall behind on interest payments. She was previously the SEC's chief economist and director of its division of economic and risk analysis.

WHO IT HITSThis lands hardest on hyperscaler finance chiefs and their creditors, who must justify data-center spending against borrowed money, and on pension-fund and insurance managers whose portfolios carry that debt indirectly.

Not sure about something? Ask the AI

Questions and answers are published on this page.

Summaries like this, every morning.

Get these free every morning

Context & Analysis

Wachter, a University of Pennsylvania Wharton School finance professor and former SEC chief economist, starts from what she calls a "remarkable fact" that is not in question: a handful of hyperscalers are investing huge amounts to build AI data centers. Rather than forecast how widely AI models will be used, she and her coauthor asked how fast hyperscaler earnings must grow to justify the spending.

The result would resemble the kind of economic growth seen during the US IT boom over a period of about 10 years starting in the mid-1990s. But Wachter says compressing that into a few years is "a lot of growth." If a productivity boom "fails to materialize," she and her coauthor conclude, "the current buildout will be the largest misallocation of capital in history."

That caution connects to the wider financing question the article raises: hyperscalers are increasingly borrowing to fund these data centers, so the productivity question is likely to shape whether the buildout proves sustainable for creditors and the broader economy.

FAQ

How much AI data-center spending is projected through 2027?: Jessica Wachter and her collaborator estimate expenditures will reach nearly $1.1 trillion through 2027.
What productivity gain would hyperscalers need to break even?: They would need to increase their own productivity by a factor of 2.7 to break even by 2030, accounting for the cost of capital and a 15% return, and depreciation of the assets.
What happens if the hyperscalers cannot meet those profit goals?: Wachter says they will fall behind on their interest payments, which risks bankruptcy.

MIT Technology Review AIRead Original Article

Send on LINE

AI news that matters for your work, delivered every morning.

Pick your industry and the AI tools you use, and get news related to your work every day.

Get it free

Free · 30 seconds with Google · unsubscribe anytimeWhat is AIToday? →

Ask AI

Ask AI anything about this article. The AI reads this article, earlier AIToday articles, and Wikipedia, and cites its sources. Q&As are published on this page for other readers too.