AI Spending Faces A Trillion-Dollar Reckoning If Productivity Falls Short, MIT Says

Research by Wharton finance professor Jessica Wachter and coauthor Jonathan Wachter, examined by MIT Technology Review, analyzed spending by Alphabet, Microsoft, Amazon, Meta, and Oracle. The study concludes the AI sector must achieve a 2.7-fold productivity increase by 2030 after accounting for capital costs, depreciation, and a 15% return, warning that failure to meet this target could make the buildout "the largest misallocation of capital in history." Morgan Stanley estimates approximately $2.9 trillion in global data-center spending through 2028, with roughly $1.5 trillion requiring external capital, pushing AI infrastructure risk into corporate debt, securitized credit, and private-credit markets. Alphabet reported negative free cash flow of $5.9 billion in Q2 2026—its first such quarter since Google's 2004 IPO—as capital spending surged. Meta's $27 billion Hyperion project in Louisiana exemplifies the financing model, with Blue Owl Capital owning 80% of the joint venture while Meta retains 20%. Former SEC chair Gary Gensler described the AI investment cycle as "a parlay bet by the capital markets and the economy." The unresolved question is whether productivity and revenue can rise fast enough to match increasingly externally financed infrastructure commitments.

(Source：yellow.com)