Goldman Sachs says AI boom fueling U.S. economy will soon peak

[Financial News]  Warnings are mounting that the momentum behind the artificial intelligence (AI) boom, which has driven economic growth in the United States for at least a year, will soon peak. 
On the 26th (local time), business magazine Fortune reported that major Wall Street financial institutions say that, despite massive investment and high expectations, tangible financial returns have yet to reach companies outside the major IT firms.
Goldman Sachs Chief Economist Jan Hatzius said, "Even before concerns about the sustainability of recent AI investment began to grow, the boost the AI boom gives to economic growth was expected to peak in late 2026 or early 2027."
In a recent client note, Apollo Global Management Chief Economist Torsten Slok said, "For U.S. companies outside the tech sector, the financial benefits brought by AI technology are virtually 'zero (0).'"
According to Apollo's analysis, recent gains have been concentrated entirely among large tech companies. Inflation-adjusted sales per employee and operating margins rose at the seven Big Tech companies known as the 'Magnificent Seven' (M7), while the remaining 493 companies in the S&P 500 made little progress. The Federal Reserve Bank of San Francisco's total factor productivity measure also showed no signs of AI-driven productivity gains.
Slok said, "Because AI has not helped profitability, non-tech companies have begun to view spending on AI tokens as a 'cost' that eats into margins, rather than as a driver of growth." He added, "As long as agentic (autonomous) AI and token consumption remain confined to only part of the economy, AI's macroeconomic impact will be very limited."
Goldman Sachs expects the growth in AI investment as a share of gross domestic product (GDP) to slow over the next year. Also weighing on the outlook are rising energy prices amid renewed tensions in the Middle East and an average increase of around 100 basis points (1 percentage point) in government bond yields in advanced economies since early 2026. Hatzius cited "increased AI-related borrowing" as one cause of the rise in yields and estimated that, if it continues, it could reduce U.S. GDP by a total of 0.6%. The boost to consumer spending from rising stock prices (the wealth effect) is also expected to ease in the coming period.
The two experts had differing outlooks for the future.
Goldman Sachs maintained its base-case scenario that AI investment will ultimately lead to productivity gains and higher profits across companies. Hatzius cautioned, however, that "in the short term, we will have to accept the side effect of disruption in the labor market" before those benefits materialize. He also noted a clearer trend toward lower employment growth at non-manufacturing companies with higher rates of AI adoption.
Apollo, by contrast, offered a more skeptical view, saying that even as AI models become widespread, companies will find it difficult to turn them into increased sales. Fortune reported that Slok expressed doubts about major companies' ability to translate AI into business results.


[email protected] Yoon Jae-joon Reporter