How big is AI's contribution to the US economy, really? Institutions give different answers.

Artificial intelligence (AI) is powering the US economic expansion, and in recent months that expansion has shown signs of accelerating. Economists expect data due Oct. 29 to show that US gross domestic product grew in the third quarter at the fastest pace in a year. AI doesn't appear as a separate line item in the GDP report, but its effects cut across multiple categories. AI contributes to GDP through several channels, including the billions of dollars in direct investment that tech companies are pouring into the technology, the data-center construction boom that has followed, and the boost to consumer spending from the stock-market gains AI has driven. So how big is AI's contribution to GDP? The question is disputed. Some economists see a modest effect, while others calculate a significant contribution. Goldman Sachs' estimate is relatively conservative. Its economists say AI's impact is constrained by a range of factors, including AI-related imports and inflation, as well as the crowding-out effect of AI on investment in other parts of the economy. Goldman estimates that AI contributed 0.1 percentage point to GDP over the past year. From the second quarter of 2025 to the second quarter of 2026, the US economy grew 2.2%. A Federal Reserve working paper published in July examined business investment in software, data centers, power facilities, and computers and peripheral equipment, concluding that AI-related gains were uneven across quarters but amounted to about one-fifth of overall GDP growth on a one-year basis. The economics team at Nationwide Mutual Insurance estimates that over the past year, direct AI investment plus spillover effectssuch as job growth from data-center construction and wealth created by the AI-led stock-market boomcontributed 0.6 percentage point to GDP, or about a quarter of overall growth. "That's quite substantial," said Chief Economist Kathy Bostjancic. ING expects the US economy to grow about 2.4% in 2026, with AI's direct and indirect effects contributing about 0.8 percentage point, or one-third. Bloomberg Economics, meanwhile, argues that official statistics understate AI's contribution to GDP. Its estimates show that if government statisticians had better captured the technology's impact, US economic growth in 2025 would have been a stronger 3.1%, rather than the recorded 2.3%. How large is US investment in AI? AI is poised to become the largest investment in US history. Stijn Van Nieuwerburgh, a professor of finance and real estate at Columbia Business School, predicts that total AI-related investment in data centers, power infrastructure, chips and computer equipment from 2025 to 2032 will reach $10.3 trillion, or about 3.6% of US GDP annually. That share would exceed the proportion historically invested in canals, railroads, electrification, highways or internet infrastructure. The largest of these, railroad construction from 1870 to 1890, accounted for about 2.2% of GDP. Where does AI's quantifiable impact on GDP mainly show up? Spending on software, data centers, chips, computers and power infrastructure shows up most clearly in the equipment and intellectual-property line items of the GDP report. However, a substantial portion of the spending underpinning the AI boom goes to components manufactured outside the US, such as chips, computers and computer equipment. US government statisticians record the billions of dollars spent on AI goods and services, then strip out the portion produced overseas from the GDP calculation, because GDP measures only domestic production. As a result, AI-related imports drag down the headline measure of economic growth. How does AI affect consumer spending? Through the so-called "wealth effect," Americans spend more when their wealth rises, even if much of that is only paper wealth. Since 2024, the stock market has repeatedly hit record highs, partly because of investors' bets on AI. The result: US household wealth rose to a record high in the second quarter of this year. To be sure, not all stock-market gains come from AI, and not everyone benefits from rising share prices. The wealthiest 10% of Americans hold about two-thirds of household wealth, while the bottom half holds just over 2% of total wealth. But wealthier Americans tend to own stocks and are driving consumer spending, a key driver of economic growth. Some estimates show that about half of AI's contribution to economic growth comes from the wealth effect. "We're seeing high-income households continue to spend strongly. They have income, and they hold wealth. Lower- and middle-income households are feeling greater financial strain from weak wage growth and rising living costs," said James Knightley, chief international economist at ING. What would happen to the economy if AI investment pulled back? "It's flattering GDP growth, but it's not everything," said Thomas Ryan, senior North America economist at Capital Economics. Other pillars of the economy include the healthcare sector, which accounted for about three-quarters of new jobs over the past year. Surveys show that US factories have been adding workers in recent months as output rises. Companies across US industries have also reported strong profits. Even so, an AI slowdown could still have a huge impact on the economy. In that scenario, business investment would slow. If stock values stopped growing or shrank, the wealth effect would reverse. The labor market could also be affected. AI, for example, is propping up the construction industry. Since the fall of 2024, residential builders have cut 97,500 jobs. But data-center construction has pushed nonresidential building to a record high, enough to offset the job losses at residential builders. What is AI's long-term impact on the economy? It is a huge unknown. AI leader Anthropic PBC said in its listing filing that AI could pose an "existential threat" to humanity, or could "accelerate global growth and significantly improve the quality of life for people around the world." So far, AI's impact on productivity has not been obvious. But the technology appears to have begun affecting parts of the labor market, with some research showing it is reducing hiring of entry-level job seekers in fields such as software development and customer support. A study by consulting firm McKinsey predicts that over the next decade, AI could force about 11 million US workers to find other work. "Even in the best-case scenario, the transition to the new AI era will be one of the most turbulent periods in human history," Microsoft co-founder Bill Gates wrote in an August article. He said "there will be new jobs, but without the right policies, there will be far fewer of them than there are today." Despite possible disruptive changes in the medium term, many economists and policymakers predict that AI will deliver substantial returns over the long run. "In the longer run, I am optimistic that AI-driven productivity growth can raise living standards for all Americans," Federal Reserve Governor Lisa Cook said in a September speech.