Economists Warn AI Infrastructure Spending May Outpace Productivity and Revenue Returns

A report highlights that global spending on AI infrastructure has surpassed historical capital booms like the dot-com era and railway expansion, yet broad economic productivity gains have not yet materialized. Cloud hyperscalers and AI companies are estimated to require over $4.2 trillion in incremental revenue over the next five years to sustain their infrastructure buildout. The growing gap between lofty financial valuations and realistic near-term corporate earnings raises the risk of market corrections if new revenue streams like AI robotics or automated material discovery fail to emerge quickly. However, even if financial bubbles contract, the physical computing infrastructure will remain as a foundation for long-term economic transformation. Economists calculate that sustaining Nvidia's valuation demands a 3% to 5% annual US productivity surge, well above the Congressional Budget Office's 1.75% benchmark forecast. Additionally, Stanford research indicates early labor market friction, with entry-level hiring for 22-to-25-year-olds in AI-vulnerable fields like accounting down 19% compared to non-exposed jobs.