AI’s boost to the U.S. economy is nearing its peak, Goldman says, as Apollo finds ‘zero financial payoff’ outside Big Tech

The AI boom has been giving U.S. growth a lift for more than a year, but Goldman Sachs thinks that lift is close to peaking. Chief economist Jan Hatzius said in a note Monday that the growth impulse should crest in late 2026 or early 2027, even before the recent worries about whether the investment wave can keep going. That’s a softer landing than a sudden stop, but it still means the easy gains are fading.

Hatzius is not calling AI useless. He still expects investment in the technology to rise sharply, just at a slower pace as a share of GDP over the next year. But he also sees other drag building around it: higher fuel prices tied to fears of renewed Middle East escalation, and long-term rates that are up about 100 basis points on average across developed-market yield curves since early 2026. He partly blames large deficits and AI-linked borrowing for that, and says the rate move would subtract a cumulative 0.6% from GDP if it sticks.

The consumer side has already gotten a boost from the stock market. Goldman estimates the AI wealth effect added 0.4 percentage points to U.S. consumer spending growth over the past year, equal to 0.25% of GDP. Hatzius expects that to fade over the next year. If AI stocks keep holding up, the decline should be gradual. If they slide, the drop in support gets sharper.

Apollo’s Torsten Slok is much harsher. In a note Saturday, he said AI has delivered “zero financial payoff” to corporate America outside tech, despite the hype and the money. His argument is that the gains are piling up mainly at the Magnificent 7, where revenue per worker and profit margins are rising, while the rest of the S&P 500 — the S&P 493, in his shorthand — is basically flat. He also pointed to a San Francisco Fed measure showing no AI lift in productivity.

That matters because mainstream companies, in Slok’s view, are starting to treat token spending like a cost, not a growth engine. Venture-backed firms are spending more on tokens, but the broader corporate world “does not need the latest model and is unwilling to pay for it.” Goldman still thinks AI will eventually boost productivity and profits across the economy. The split between the two camps is simple enough: one sees a delayed payoff, the other sees a very expensive waiting room.