NVIDIA shares under criticism: Goldman Sachs sees no GDP effect from AI

• Goldman Sachs chief economist Hatzius described AI’s GDP contribution in 2025 as “basically zero”
• This contradicts the narrative that AI spending significantly supports US growth
• US tech companies are planning joint AI investments of around $700 billion for 2026

Why Goldman Sachs is challenging the AI ​​narrative

The major US technology companies have invested hundreds of billions of dollars in AI infrastructure in 2025. Meta, Amazon, Alphabet, Microsoft and Oracle plan combined capital spending of around $700 billion in 2026, according to the companies’ respective quarterly reports. These numbers have fueled a narrative on Wall Street that massive AI spending is significantly supporting US economic growth. But Goldman Sachs clearly contradicts this representation.

In an interview with the Atlantic Council on January 8, 2026, Goldman Sachs chief economist Jan Hatzius stated that the contribution of AI investments to U.S. GDP growth in 2025 was “basically zero.” Hatzius literally said that Goldman Sachs does not view AI investments as “strongly growth-oriented.” Much of what is reported about the economic contribution of AI is simply misrepresented, says Hatzius. His colleague Joseph Briggs, co-head of economic investment research at Goldman Sachs, described the previous interpretation in a February 23, 2026 Washington Post article as a “very intuitive story” that may have discouraged people from taking a closer look.

The import problem: who really benefits

The central point in the Goldman Sachs analysis concerns the structure of the value chain. Gross domestic product measures domestic production rather than domestic spending. When US corporations spend billions on AI hardware, much of that money goes to manufacturers abroad. NVIDIA, by far the most important supplier of AI chips, is based in the USA, but has its processors manufactured by TSMC in Taiwan. The HBM memory chips also come from South Korea, where SK hynix and Samsung produce.

Hatzius summed it up in the Atlantic Council interview: A large part of the AI ​​investments made in the USA actually increase the Taiwanese and South Korean GDP, but not the American one. In the GDP calculation, the positive entry on the investment side is offset by a negative entry on net exports because the imported equipment largely neutralizes the effect. Even the record sales that NVIDIA recently achieved quarter after quarter with its data center business do little to change this mechanics, as the physical value creation predominantly takes place in Asia.

Lack of productivity gains and the Solow paradox

In addition to the import issue, there is a second finding: Even at the company level, it has so far been difficult to demonstrate any measurable productivity gains through AI. A study published in February 2026 by the National Bureau of Economic Research (NBER Working Paper No. 34836) surveyed almost 6,000 executives in the USA, Great Britain, Germany and Australia. The result: Although around 70 percent of companies are actively using AI, over 80 percent reported no impact on productivity or employment in the past three years. Even top managers who use AI personally only spend an average of 1.5 hours per week on the technology.

However, as the NBER study also shows, the managers surveyed expect a productivity increase of 1.4 percent and a decline in employment of 0.7 percent due to AI in the next three years. This is reminiscent of the so-called Solow paradox from the 1980s, when Nobel Prize winner Robert Solow stated that computers could be seen everywhere but in productivity statistics. Goldman Sachs itself predicted in a research report in 2023 that a measurable impact of AI on US GDP and labor productivity could only be expected from 2027. For investors who bet on AI stocks like NVIDIA, the question remains as to when the billions in spending by technology companies will actually translate into overall economic growth, or whether a relevant part of the value creation will remain permanently in Asia.

D. Maier / editorial team finanzen.net

Selected leverage products on Alphabet C (ex Google)

With knock-outs, speculative investors can participate disproportionately in price movements. Simply select the lever you want and we will show you suitable open-end products on Alphabet C (ex Google)

The leverage must be between 2 and 20

Advertising