Why AI Spending Hasn't Delivered a Productivity Boom Yet

Companies are pouring billions into AI. Faster workers and higher AI use have not consistently translated into profit or company-wide productivity. That gap between investment and return is the central tension in the current AI cycle.

Corporate spending on AI infrastructure and software has surged. Major cloud providers and enterprise software vendors report double-digit growth in AI-related revenue. The productivity data tells a different story.

U.S. nonfarm business sector productivity rose 2.3% in the fourth quarter, according to the Bureau of Labor Statistics. That is a solid number by historical standards. It is not the dramatic acceleration that AI bulls predicted. Labor productivity growth has averaged roughly 1.5% annually over the past two decades. The current run is above trend, not dramatically so.

The disconnect is partly a measurement problem. Productivity statistics capture output per hour worked across the whole economy. AI tools are still concentrated in specific functions: customer service chatbots and code generation. Those represent a small slice of total labor input. Even if those tools double the output of the workers using them, the aggregate effect is modest until adoption broadens.

There is also an integration lag. Companies that deploy AI tools often find that the gains show up in task completion times, not in the revenue or output metrics that feed productivity calculations. A customer service agent who resolves tickets 30% faster may handle the same volume if call volume is flat. The time saved goes to other tasks, or to slack, not to measurable output.

JPMorgan Chase CEO Jamie Dimon said in his annual shareholder letter that the bank sees "serious" potential for AI to improve productivity and customer service. The benefits will take years to materialize, Dimon wrote. "We are in the early innings of the AI revolution. The full impact will not be felt for some time."

That timeline mismatch creates a risk for companies spending aggressively on AI now, expecting a near-term earnings payoff. If productivity gains remain invisible in the macro data for another two to three years, CFOs may face pressure to justify the spending to boards and investors who want to see the return on the P&L.

Goldman Sachs economists estimated in a February note that AI could boost U.S. GDP by roughly 0.4% annually over the next decade. That is a meaningful contribution, not a revolution. The bank's analysts said the productivity gains would be concentrated in sectors like technology and finance, where AI tools directly augment knowledge work.

For investors, the question is not whether AI will eventually boost productivity. It almost certainly will. The question is whether the current wave of spending is priced for a payoff that is still years away. If the productivity data stays in its current range through 2025 and 2026, the stocks of companies that have sold the AI productivity story hardest could face a reckoning.

The next concrete data point comes May 7, when the Bureau of Labor Statistics releases first-quarter productivity figures. A print below 1.5% would reinforce the view that the AI spending cycle is running ahead of the productivity cycle. The bulls need a print above 2.5% to regain cover.