Goldman Sachs Bombshell Report Finds Trillions in AI Investment Generated Basically Zero Economic Growth in 2025
A bombshell Goldman Sachs analysis published February 24 2026 found that despite an estimated 600 billion dollars in AI-related capital expenditure across U.S. companies in 2025 the technology contributed negligibly to GDP growth reigniting fierce debate about whether the AI investment boom is a financial bubble.
Goldman Sachs Drops a Bombshell: All That AI Spending Bought Almost Nothing
For three years, Wall Street told a very specific story about artificial intelligence: the investment is enormous, but the payoff will be even bigger. Patience, they said. Wait for the productivity wave.
Goldman Sachs, one of the firms that told that story most loudly, just published a report that quietly dismantles it. The finding: despite an estimated $600 billion in AI-related capital expenditure across U.S. companies in 2025, the technology measurable contribution to gross domestic product was, in the bank own words, basically zero.
The Numbers Behind the Conclusion
Goldman economists analyzed productivity data, corporate earnings calls, and sector-by-sector output figures across the U.S. economy. What they found was a stark mismatch: AI adoption in enterprise settings has been slower than projected, the use cases generating real efficiency gains remain narrow, and the industries spending the most on AI have not shown statistically significant productivity improvements that can be directly attributed to the technology.
The report notes that previous general-purpose technologies, such as electricity and the internet, also showed long lag periods between adoption and measurable economic impact. But it cautions that those comparisons may not apply cleanly to AI, given the concentration of investment among a small number of companies.
According to Stanford economist Dr. Erik Brynjolfsson, one of the leading researchers on technology and productivity, we have seen this pattern before. The gains from new technologies often do not show up in GDP for a decade or more. But the Goldman report raises a legitimate question: are we building the right things, for the right users, in the right ways?
Silicon Valley Pushes Back and the Bubble Question Grows
The response from the tech industry was swift and defensive. Executives at Microsoft, Google, and OpenAI all disputed the report methodology, arguing that GDP metrics are poorly suited to capturing the value of AI-driven automation. One senior Google executive called the analysis a category error, like measuring the value of the internet in 1997.
Nvidia, whose stock has been among the biggest beneficiaries of AI infrastructure spending, saw its share price dip 4 percent on Tuesday following the report release before partially recovering by afternoon trading.
That word, bubble, is now circulating openly in financial circles in a way it was not six months ago. Several prominent venture capital investors have begun quietly reducing their AI-adjacent positions. Short interest in major AI-exposed ETFs has risen sharply over the past 60 days.
As tech companies prepare to report Q1 2026 earnings next month, investors will be watching AI return-on-investment figures more closely than ever before.
