MIT Study Warns AI Bubble Burst Could Trigger Historic Capital Misallocation

A new analysis from MIT Technology Review looks at what might happen if the artificial intelligence investment bubble collapses, noting that hyperscale cloud operators may need to nearly triple their productivity by 2030 simply to recover their trillion-dollar infrastructure outlays.

Wharton finance professor Jessica Wachter, together with a coauthor, developed the projection by tallying confirmed spending by hyperscalers. They estimate nearly $1.1 trillion will go into data center construction through 2027 at Alphabet, Microsoft, Amazon, Meta and Oracle. The consequences of that wager extend far beyond the technology sector.

Productivity Requirements Under the AI Investment Boom

Wachter, who previously served as the Securities and Exchange Commission's chief economist, determined that hyperscaler productivity would need to expand by a factor of 2.7 to break even by 2030.

The figure incorporates the cost of capital, a 15% return and asset depreciation. Without that level of growth, Wachter and her coauthor arrive at a grim assessment.

“The current buildout will be the largest misallocation of capital in history.”

— Jessica Wachter, Wharton finance professor

Alphabet posted a $5.9 billion free cash flow deficit last quarter, its first since its 2004 initial public offering. As spending becomes more concentrated, investors are increasingly seeing AI outlays as a risk to broader markets.

Debt Financing Extends the Risk Beyond Big Tech

Morgan Stanley calculates that hyperscalers will fund more than half of their planned $2.9 trillion in data center investment through 2028 by borrowing rather than using cash. That money is increasingly sourced from outside capital rather than internal reserves.

In Louisiana, Meta transferred an 80% stake in its Hyperion data center to private-credit firm Blue Owl Capital. The deal illustrates how complex hyperscaler financing structures have become.

Columbia Business School's Stijn Van Nieuwerburgh warns that debt of this kind increasingly moves through pension funds and private credit vehicles. He notes that few people recognise how deeply that exposure has penetrated their own retirement and insurance savings.

Crypto strategist Arthur Hayes has raised a parallel possibility. He contends that an AI credit bust could compel the Federal Reserve to print money, driving bitcoin (BTC) toward $1 million.

“A parlay bet by the capital markets and the economy.”

— Gary Gensler, former SEC chair and MIT Sloan School professor

Gensler anticipates a pullback eventually, though the timing is unclear. Whether it happens slowly or suddenly may decide how much of that trillion-dollar wager turns into a permanent loss.