AI Productivity Boom Remains Elusive Despite Spending Surge

Generative AI has been cast as a general-purpose technology capable of doing for knowledge work what electricity and computers did for earlier eras: lift output per hour, expand economic capacity and ease inflation without suppressing demand. The stakes extend beyond corporate returns. A durable productivity acceleration would influence wages, profit margins and Federal Reserve policy, while failure to deliver would deepen concern that vast spending on chips and data centres is producing only narrow task-level efficiencies.

That promise remains hard to find in aggregate data. A July 22 Financial Times analysis, citing Barclays, said Amazon, Alphabet, Microsoft and Meta are on course for about $725 billion in 2026 capital expenditure, yet industry adoption rates show no statistically significant relationship with productivity gains. A St. Louis Fed survey found only 10% of workers use AI daily; adopters average about 30 minutes a day and report roughly 10 minutes saved. The Census Bureau’s latest survey put active business use at 21%, underscoring slow, uneven diffusion.