NVIDIA and AI-exposed equities keep climbing. Daron Acemoglu, Nobel laureate in economics, says the productivity gains justifying those prices do not yet exist.1
"There's a huge amount of uncertainty," Acemoglu told MIT Technology Review, pointing to anecdotes about worsening job markets for college graduates alongside no measurable effect on aggregate productivity.1 His view: AI agents are tools that augment specific tasks, not systems capable of replacing whole roles.1
That assessment sits uncomfortably against current valuations. Tech equity multiples are priced for a productivity revolution. The macro backdrop is not cooperating.
CPI is running at 3.8%. Services inflation remains stubbornly above 3% annually.2 The Iran conflict has pushed average U.S. household gasoline costs up $857 in 2026, according to the Stanford Institute of Economic Policy Research.3 Consumer sentiment is deteriorating across middle- and higher-income cohorts — the demographic that drives discretionary tech spending.
A Federal Reserve leadership transition adds another variable. Markets are repricing rate-cut timelines, and higher-for-longer rates compress the present value of future earnings — exactly the earnings AI companies are promising but not yet delivering.
Google DeepMind is reportedly hiring economists specifically to shape the AI growth narrative, a signal that the productivity credibility gap is being taken seriously inside the industry.1 The effort itself acknowledges the problem: the numbers are not yet making the case on their own.
Commodities markets are flagging the same tension in real time. Silver swung between tariff-driven optimism and inflation-driven disappointment within a 48-hour window — a sign that macro uncertainty is acute, not background noise.
The core trade-off for equity investors is straightforward. AI infrastructure spending is real and accelerating. The productivity payoff that justifies growth-stock premiums is, by Acemoglu's account, still small and slow to materialize.1 In a 3.8% inflation environment with a Fed in transition, that gap is expensive to carry.
Tech equity bulls need either falling rates or credible productivity data. Right now, neither is clearly in view.
Sources:
1 Daron Acemoglu, MIT Technology Review, May 11, 2026
2 Bureau of Economic Analysis, finance.yahoo.com
3 Stanford Institute of Economic Policy Research, finance.yahoo.com, May 16, 2026
