Goldman says AI drove S&P 500 profits 51% higher but the tailwind is fading

The AI boom drove S&P 500 profits 51% higher, and its biggest tailwind is fading¹. Goldman Sachs chief US equity strategist Ben Snider estimates that almost half of 2026 S&P 500 earnings growth can be attributed to AI-related investment¹. Goldman Sachs Research has asked directly whether AI investment can drive S&P 500 earnings even higher from here².

Recent earnings growth has been “extraordinary”, Snider wrote in a Goldman Sachs Research report². AI appeared on 67% of S&P 500 earnings calls in the second quarter⁶. Nearly half of the index’s earnings-per-share growth came from AI investment⁶. Profits at the index level jumped 51% on the boom⁴.

The depreciation problem

Depreciation from $800 billion in 2026 hyperscaler capital spending will nearly offset all AI earnings gains by 2028, Goldman estimates⁵. That is the mechanism behind the fading tailwind, and it is an accounting one rather than a demand one⁵. Capital spending at that scale converts into a depreciation charge that runs through reported earnings for years after the cash is committed⁵.

Goldman projects continued but slower earnings growth and a modest stock market rise if AI delivers on its productivity promise⁴. The bank holds an 8,700 S&P 500 target implying 14% upside, contingent on AI generating broad economic productivity gains rather than infrastructure spending alone³. Goldman Sachs Research has framed the open question as whether AI investment can drive index earnings even higher from here².

What to watch

The forecast is conditional rather than directional. The target rests on productivity gains appearing across the economy, not on the build-out continuing at its present pace, and the AI-driven component of growth is described as slowing^3,4. Goldman’s projection of slower growth with a modest market rise is explicitly conditioned on the productivity promise being met⁴.

Two indicators follow directly. Hyperscaler capital expenditure guidance is the variable the depreciation arithmetic turns on, and hyperscaler capex is already a standing subject on earnings calls^5,6. And the share of index earnings growth attributable to AI is the number to track, since almost half of this year’s growth came from that source and the headline profits figure is a 51% rise^1,6. Goldman’s own framing treats broad productivity, not infrastructure spending, as the test the market has to pass³. Hyperscaler capex is already a standing debate item for the largest US operators⁶.