Goldman Says AI Spending Has Yet to Lift Corporate Earnings

US companies are committing vast sums to generative AI on the expectation that automation will cut labor costs, lift productivity and widen profit margins. So far, investors have found clearer returns among infrastructure suppliers such as Nvidia and major cloud providers than among businesses deploying AI applications. The gap matters because lofty valuations increasingly depend on companies converting technology spending into measurable earnings gains rather than promising benefits years ahead.

Goldman Sachs’ findings, reported on March 3, 2026, tracked S&P 500 management commentary during fourth-quarter 2025 earnings calls. About 70% of management teams discussed AI and 54% linked it to productivity or efficiency, but only 10% quantified gains in specific use cases and just 1% measured an earnings impact. With 2026 technology capital expenditure estimated at $667 billion, the analysis found little evidence that AI adoption had yet produced materially above-market profit growth.