Goldman warns AI’s $115B blind spot is distorting growth data

Goldman Sachs said the current artificial intelligence boom is not showing up in government growth statistics. The bank’s analysts argued that roughly $115 billion is uncounted in official growth statistics.

The analysts calculated that AI technology has contributed about $160 billion in real U.S. economic activity since 2022, representing 0.7% of GDP. Only around $45 billion (0.2% of GDP) of AI-led growth has been recorded.

Goldman Sachs believes the impact of AI on GDP was smaller

Goldman Sachs analysts revealed that the uncounted $115B indicates the difference between what companies report and what the government measures. They also said it’s the Commerce Department’s Bureau of Economic Analysis method of calculating growth.

The analysts argued that the measured impact of AI on GDP was smaller because BEA’s methodology for estimating growth considers semiconductors as intermediate inputs. They revealed that semiconductors are only counted towards final demand when the products, such as consumer laptops, that they enable are sold.

The observers stated that the AI chips developed in recent years are being used for training and supporting AI models. They believe that the semiconductors build an intangible asset whose final output value has not been fully capitalized or measured in growth statistics.

The analysts also said that new import policies disrupted the impact of AI technology on GDP. They noted that business investment in AI equipment surged in the first half of 2025 because companies rushed to import servers and networking gear ahead of President Donald Trump’s import levies.

The analysts acknowledged that the trend reflects one-time frontloading ahead of tariffs, thus exaggerating normal AI investment demand. They believe the AI investment boost was partly offset because imports are subtracted from GDP.

AI propels the S&P 500 to all-time highs

According to a separate note from Goldman Sachs earlier this month, AI’s impact on economic growth is hard to pin down since companies are also struggling to show it in their bottom lines. They said that the shares of S&P 500 companies quantifying the impact of AI on earnings today remain limited despite most firms mentioning AI on earnings calls in the second quarter of the year.

Artificial intelligence has propelled the S&P 500 to all-time highs, while data from Slickcharts shows that Nvidia makes up approximately 7% of the index. The S&P 500’s top eight publicly traded companies are also heavily invested in artificial intelligence. Those firms make up more than 36% of the S&P 500 and are ramping up their AI spending to release products and services that use the technology.

Other companies in the S&P 500 are also heavily invested in AI, including Oracle, Palantir, and Cisco. The trio makes up more than 2% of the S&P 500 combined, and their continued increase in AI investments could cause the stock market to continue to grow.

The financial institution recently warned that the stock market could drop by 20% once AI spending slows down. Goldman Sachs Analyst Ryan Hammond said there’s the danger of hyperscalers inevitably cutting back on AI expenditures.

“A revision of long-term growth estimates back to early 2023 levels would imply 15% to 20% downside to the current valuation multiple of the S&P 500.”

-Ryan Hammond, Analyst at Goldman Sachs.

AI spending has surged in the first half of 2025, but Hammond wrote that a few analysts expect a sharp deceleration in Q4 2025 through to 2026. However, tech giant Meta Platforms revealed last week that the company will spend $600 billion on AI over the next three years. 

The firm’s CEO, Mark Zuckerberg, also highlighted that there’s a possibility the company will invest more than $600 billion during the specified period. He added that a significantly high number was possible through the end of the decade.

Want your project in front of crypto’s top minds? Feature it in our next industry report, where data meets impact.