AI Investment Surpasses Railroad and Internet Booms, but Returns Questioned

PwC estimates that global data center investment could exceed $30 trillion by 2050, far surpassing the railroad and internet boom eras. As a major player in the AI race, Anthropic plans to invest $518 billion in the coming years, more than 100 times its 2025 revenue.

JPMorgan notes that in the US, which leads the AI race, broad productivity gains have yet to appear, raising doubts about whether current AI valuations can be sustained. A Bain study suggests that productivity gains from existing markets alone are not enough to support such massive investment, and entirely new markets must emerge to fill the funding gap.

A Columbia Business School economist estimates that US AI investment from 2025 to 2032 could reach about $9 trillion. For these investments to yield a 10% return, the US AI industry would need to generate about $3.55 trillion in annual revenue by 2032, while current revenue is only a small fraction of that.

Stanford researchers found that in occupations heavily affected by AI, such as accounting and paralegal work, employment levels for workers aged 22 to 25 are 19% lower than in jobs AI struggles to replace, like cleaners.