A Wharton study finds that Big Tech's roughly one trillion dollar AI investment only pays off if productivity nearly triples.

New Wharton research finds hyperscalers need to raise AI productivity by a factor of 2.7 to justify close to one trillion dollars in spending by 2027. The authors warn that a shortfall could become the largest capital misallocation in history.

Wharton finance professor Jessica Wachter and co-author Jonathan Wachter of Point72 compared investment data from the five major hyperscalers Amazon, Alphabet, Microsoft, Meta and Oracle against a model built for rare productivity booms. Spending by these firms grew from 155 billion dollars in 2022 to an estimated 755 billion dollars in 2026 and is projected to cross one trillion dollars in 2027. The model implies that each assumed productivity boom would need to lift AI-sector output by a factor of 2.7 over current levels to justify present stock valuations.

The paper compares this to past boom periods: the US IT boom from 1995 to 2005 delivered only 1.5 times per capita growth, the American railroad era produced 2.8 times over 60 years, and the East Asian growth miracles achieved 8 to 13 times over 25 to 30 years. Wachter still considers 2.7 plausible, arguing it can be derived directly from hyperscaler earnings and current market pricing rather than abstract guesswork.

What remains unresolved is whether the assumed productivity boom has actually occurred or merely reflects what managers believe will happen. The authors themselves note their approach rests on firms' investment choices, which shows conviction rather than proof. For decision makers the open question is how long capital markets will keep financing this bet before hard productivity numbers arrive.

This story was produced automatically from the source named above and checked by software before publication. The image is symbolic and shows neither the event nor a real person. How this paper is made