The AI spending boom is very real. The productivity boom? Not so much.
Goldman Sachs' economists have coined a term for the gap: "AI-nxiety." Corporate AI talk has become completely untethered from measurable economic results, and Goldman's chief economist has calculated that AI's actual contribution to US GDP growth last year was essentially zero.
The numbers are stark. A survey of nearly 6,000 executives across four countries, coordinated by four central banks and published through the National Bureau of Economic Research, found that over 80% of firms report no meaningful impact from AI on productivity or employment over the past three years. Seventy percent of those firms are actively using the technology.
Companies, however, can't walk it back. Moody's chief economist recently described the dynamic as a "Cortes moment" — a reference to the conquistador who burned his ships so his crew had no choice but to press forward. AI budgets have ballooned, teams have been restructured, and headcount decisions have been made around productivity gains that haven't materialized. Goldman Sachs Research projects AI companies will invest more than $500 billion in infrastructure this year alone.
The executives leading the charge barely use the technology themselves. The NBER study found that while two-thirds of top executives say they regularly use AI, they average just 1.5 hours a week with it. One in four don't touch it at all.
That ambition-execution gap is showing up everywhere. In the January 2026 Business Conditions Survey from the National Association for Business Economics, survey chair Martha Moore noted that "all firms report using artificial intelligence in some form, ranging from limited or experimental use to AI being core to their operations," but that "employment impacts are currently uncertain." Jon Groves, CEO of Logicalis US, struck a similar chord in his company's 2026 CIO Report: "AI ambition is high, but the real challenge is turning that ambition into something scalable and secure."
Some companies that bet on AI to trim headcount are already reversing course. Gartner predicts that 50% of organizations will abandon plans to reduce their customer service workforces through AI. "AI is not a panacea," said Kathy Ross, Senior Director Analyst in Gartner's Customer Service & Support practice. "The human touch remains irreplaceable in many interactions."
The more patient read is that none of this is unusual. Stanford economist Erik Brynjolfsson has long argued that transformative technologies follow a "productivity J-Curve" — heavy upfront investment, a frustrating stretch where the gains feel invisible, then a sharp payoff. He's pointed to US productivity growth tracking at roughly 2.7% in 2025, nearly double the prior decade's average, as an early signal the curve may be starting to bend.
OUR VALLEY VIEW
The tech industry has been here before. Cloud computing and mobile enterprise apps each went through their own hype-to-reality corrections before delivering real value. But no previous technology wave swallowed this much capital this fast with this little to show for it. Companies are probably right that the payoff is coming. They just can't tell you when, and at $500 billion a year, the meter is running.
