The AI has added $ 160 billion to “real GDP” since 2022, said Goldman Sachs. There is just one problem: th

Elsie Peng, Joseph Briggs and Sarah Dong, of the team of chief economist Jan Hatzius, explain in a research note of September 13 that “the measured impact” on the GDP of the AI is an “intermediate impact” of the Bureau of Economic Analysis (BEA), so this only matters to the final demand when a final product is sold. For example, a semiconductor only appears in GDP when, let’s say, a laptop is purchased. This means that billions of dollars in economic activity compatible AI are not measured, Peng, Briggs and Dong compete.
Since the launch of Chatgpt in 2022, they calculate, the dramatic wave of income awarded to the infrastructure of IA has stimulated the “real GDP” by 160 billion drug dollars. This figure highlights the transforming role of AI as a growth engine, but underlines a gap in confusion in official government statistics. According to Goldman Sachs, most of the real economic contributions of AI have remained largely invisible in American GDP numbers to date.
Research is based on business reports and government data, revealing expenditure on AI infrastructure by US companies, an increase of $ 400 billion since 2022. A significant part of these expenses has been focused on information processing equipment, which increased at an annualized rate of 39% in the first half of 2025.

Dig into the “real GDP”
To analyze the real national economic impact, the Goldman Sachs team has adjusted data on the company’s income by subtracting the effects of swollen prices, foreign sales of the equipment produced abroad and importing imports. This has resulted in the figure of $ 160 billion, around 0.7% of American GDP since 2022, which results in approximately 0.3 annualized growth percentage points.
The same analysts calculate the amount officially counted in the measured GDP is much lower – just 45 billion dollars, or 0.2%, since 2022. This only reflects 0.1 point of percentage of annualized growth.
The team also analyzed the four main channels by which AI affects the American economy. The first is investment in equipment, such as semiconductors and servers; The second is structures such as data centers and electrical installations. The third comes the intellectual property, including software and R&D expenses, and finally net exports of goods and services related to AI.
However, analysts warn that a large part of the overvoltage of investments has been motivated by imports, which means that it did not directly contribute to the growth of net GDP, and it has developed particularly quickly this year, which suggests that it was loaded before tariff hikes.

An image that appears
The analysis of Goldman Sachs intervenes while political decision -makers attack the way of accurately monitoring the economic impact of rapid evolution technology. Analysts warn that, although AI knows indeed significant growth, a large part remains hidden from standard indicators.
In the middle of this summer, The Wall Street JournalChristopher Mims noted a striking increase in IA infrastructure spending, arguing that it added more to GDP growth than consumer expenditure, the backbone of the American economy, responsible for some third parties of GDP as measured by the BEA. Retail sales run, more recently at 0.6% in August, but the American consumer is widely understood as being besieged. An image emerges from a significant amount of GDP from the booming AI sector and a tired economy everywhere else, with slow growth in employment. If there is an AI bubble, and if it appears soon, many could drop with it.
For investors and business leaders, the point to remember is that the United States wave could be more powerful and potentially more sustainable, than the title numbers suggest it. While the government adapts its statistical methodologies, a more complete image of the economic imprint of the AI could emerge, but for the moment, it remains largely in the shadow of the official calculation.
For this story, Fortune Used a generative AI to help an initial project. An editor checked the accuracy of the information before the publication.
https://fortune.com/img-assets/wp-content/uploads/2025/09/GettyImages-2233531854-e1758058136948.jpg?resize=1200,600