Leading Wall Street bank Goldman Sachs has revised a prominent narrative about artificial intelligence (AI) driving U.S. economic growth, concluding that AI investment contributed “basically zero” to the United States’ gross domestic product (GDP growth) in 2025. While earlier stories had circulated that AI spending — especially on data centers and infrastructure — was a major engine of growth, Goldman economists now say that when measured properly, direct AI-related investment didn’t noticeably boost official growth figures last year. This reassessment challenges the idea that AI was the dominant force in recent U.S. economic expansion.
Goldman’s chief economist Jan Hatzius and colleague Joseph Briggs noted that even though U.S. companies poured billions into AI hardware, software, and data centers, much of that spending benefited foreign economies because key components like semiconductor chips are imported rather than manufactured domestically. Because GDP calculations count only value created within the U.S., this import-heavy investment meant that a large proportion of AI spending did not contribute to U.S. GDP growth — offsetting the impact from domestic expenditure.
The bank’s reassessment follows a broader debate about how to quantify AI’s economic impact. Early estimates from economists and central bank analysts suggested that AI-related investment might explain a large share of recent economic growth — with some figures claiming up to 92 % of growth in parts of 2025 was tied to AI infrastructure spending. However, these estimates have since been challenged as overstatements once investment content and statistical measurement issues are taken into account.
Although Goldman’s analysis implies that AI hasn’t yet shown up strongly in official GDP data, many economists still argue that measuring AI’s real impact is difficult, and that the technology could drive productivity gains later this decade as adoption broadens. The debate highlights both the limits of current economic statistics to capture fast-evolving tech trends and the uncertainty about when AI’s long-term benefits will become visible in aggregate economic performance.
