Goldman Says AI Will Drive 40% of S&P 500 Profit Growth in 2026 — Then Fade

Goldman Sachs warned that AI spending propping up S&P 500 earnings growth will wane after this year. The assessment was reported on Sept. 19 and positions 2026 as the high-water mark for the current capex contribution to index earnings. Business Insider

Chief U.S. equity strategist Ben Snider said the firm's base case is for S&P 500 earnings growth to decelerate, not collapse, in coming years. Yahoo Finance Snider holds an 8000 year-end price target for the S&P 500. He is chief U.S. equity strategist at Goldman Sachs. MarketWatch

Goldman Sachs expects S&P 500 earnings per share to increase 12% in 2026 and 10% the following year. Goldman Sachs Earnings per share means profit per share. Snider estimates that AI-related investment will drive approximately 40% of S&P 500 earnings-per-share growth in 2026. In the same outlook, Goldman Sachs forecast US stocks to rise 6% in 2026. Goldman Sachs Capex means spending on things like chips and data centers.

Goldman's research momentum signal stood in the 98th percentile and its market breadth signal stood in the 94th percentile, indicating extreme elevation. MarketWatch Momentum tracks how persistent price trends are. Breadth tracks how many stocks take part.

The broader context here is quality of growth, not just the headline rate. When close to half of growth comes from one investment cycle, it persists less well than growth from broad customer demand, pricing power or lasting productivity gains. Capex pays suppliers first and builds capacity that must later earn money. That works while spending accelerates. It squeezes when spending only levels off, even if total profits still compound.

In my view the tension is between that earnings path and crowded positioning. A path of 12% then 10% already steps down. If the AI share normalizes, the rest of the index must carry more weight to reach the same 8000 target. When momentum and breadth sit near extremes, estimate cuts do not stay isolated. They spread across sectors pulled along by the same theme. Crowding does not set the business outcome, but it shapes the adjustment if estimates move.

Looking at what this means for risk management, the split to watch is capex-linked earnings versus the rest. That remainder decides whether the 8000 target can rest on stable valuations or needs higher prices for slower growth. Orderly slowing lets revenue from AI use replace profit from AI building. A sharper pause forces the market to re-price future earnings without the pass-through that had flattered margins. That argues for watching capex guidance, order books and estimate revisions outside the direct AI supply chain.

For trading desks and allocators, the message is sequencing. Goldman is not describing an earnings collapse. It is describing peak contribution. Headlines can still look strong while early inputs soften. That phase puts weight on revision breadth, earnings-call language on project returns, and whether heavy spending still converts to cash. If those hold, slower growth fits a higher index. If they loosen, the same 8000 implies a richer price for lower-quality earnings, with less room for disappointment.