What's at stake in AI’s trillion-dollar gamble

The article highlights a critical financial analysis of the massive investments being poured into AI infrastructure by a few "hyperscalers." Jessica Wachter, a finance professor at the University of Pennsylvania’s Wharton School and former SEC chief economist, has taken a pragmatic approach to assess the economic implications. Instead of predicting AI's widespread utility, she focused on the earnings growth required to justify the staggering expenditures. Wachter and her collaborator estimate that by 2027, these investments will reach nearly $1.1 trillion. To break even by 2030, accounting for capital costs, a 15% return, and asset depreciation, the AI companies will need to boost their own productivity by a factor of 2.7. While challenging, Wachter notes this isn't impossible, comparing the potential economic growth to the US IT boom of the mid-1990s. However, she cautions that achieving such growth by 2030 means "a lot of growth compressed into a few years." The stakes are incredibly high. Hyperscalers are projected to spend around $750 billion this year alone on building vast data centers. If these companies fail to meet the aggressive profit goals, Wachter warns of potential bankruptcies due to inability to cover interest payments. She and her coauthor conclude that if a productivity boom "fails to materialize," the current AI buildout could become "the largest misallocation of capital in history." This underscores the immense financial risks associated with today's unprecedented investments in artificial intelligence infrastructure.