MIT Explains What Happens When the Trillion-Dollar AI Bubble Bursts

MIT Technology Review examines what happens when the AI bubble bursts, warning that hyperscalers may need to nearly triple productivity by 2030 just to break even on their trillion-dollar infrastructure bet.

Wharton finance professor Jessica Wachter and a coauthor built the estimate around confirmed hyperscaler outlays, projecting nearly $1.1 trillion in data center spending through 2027 across Alphabet, Microsoft, Amazon, Meta, and Oracle.

The wager matters well beyond Silicon Valley. Inside the AI Bubble’s Productivity Math Wachter previously served as chief economist at the Securities and Exchange Commission (SEC). She found hyperscaler productivity must grow 2.7 times over to break even by 2030.

Hyperscalers have a lot of ground to make up. Image Source. MIT Review That calculation accounts for the cost of capital, a 15% return, and depreciation of the assets. Absent that growth, Wachter and her coauthor reach a stark conclusion.