Goldman Sachs Analysis Finds AI Investment Added Zero to 2025 US Growth

2026-02-25 economy

New York, Tuesday, 24 February 2026.
Goldman Sachs reveals AI added “basically zero” to 2025 US GDP, as billions in capital expenditure on imported hardware effectively boosted the economies of Taiwan and South Korea instead.

A Reality Check on the AI Economic Boom

While Wall Street has been buzzing with the narrative that artificial intelligence is the primary engine driving the U.S. economy, a sobering analysis from Goldman Sachs suggests otherwise. On Monday, February 23, the bank’s Chief Economist, Jan Hatzius, stated that AI investment contributed “basically zero” to U.S. GDP growth in 2025 [2][4]. Despite a massive spending frenzy by tech giants, Hatzius argues that there has been significant misreporting regarding the actual macroeconomic impact of these investments, which is much smaller than often perceived [1][8]. This assessment challenges the prevailing optimism that capital expenditures alone are sufficient to prop up broader economic metrics.

The Import Offset Mechanism

The discrepancy lies in the mechanics of how Gross Domestic Product is calculated, specifically regarding the hardware required to power AI models. While companies like Meta, Amazon, and Google spent billions in 2025, a vast majority of the critical equipment—specifically microchips and data center components—is imported [1]. Hatzius explained that this spending effectively boosts the GDP of manufacturing hubs like Taiwan and South Korea rather than the United States [1][4]. Technically, because imports are subtracted from GDP calculations, the influx of foreign hardware offsets the domestic investment figures, neutralizing the sector’s contribution to the headline growth number [1].

Conflicting Narratives and Economic Data

This findings stand in stark contrast to previous economic indicators and political rhetoric that championed AI as a dominant growth driver. In November 2025, President Donald Trump cited AI investment as the force making the U.S. economy the “hottest” in the world to argue against state-level regulations [1]. Similarly, Harvard economist Jason Furman had previously noted that investments in information processing equipment accounted for 92% of GDP growth in the first half of the year [1][8]. However, newer data offers a more granular view; economic analyst Joseph Politano estimates that AI investment was responsible for only 0.2% of the U.S. economy’s 2.2% growth in 2025—a contribution of roughly 9.091 percent [4][6].

The Productivity Waiting Game

Beyond the immediate GDP calculations, the anticipated productivity boom from AI adoption appears to be lagging behind the capital expenditures. A recent survey of nearly 6,000 executives found that while 70% of firms are actively using AI, approximately 80% reported no impact on employment or productivity [1]. Goldman Sachs Research had previously forecasted that measurable impacts on labor productivity would not begin until 2027 [6][8]. With top tech companies expected to spend roughly $700 billion on AI infrastructure in 2026, the pressure is mounting for these investments to yield tangible economic returns [1][4].

Sources

Artificial Intelligence GDP Growth