If there’s one thing AI excels at, it’s defying every attempt to build a coherent narrative around it.
Is AI destroying jobs, or just masking the same old garbage labor market? Are data centers unlocking prosperity for generations to come, or hemorrhaging value faster than a new car driven off the lot?
Whatever can be said of AI’s consequences for the future, one of the more widely agreed-upon views among economists seemed to be that tech spending is propping up an otherwise dismal economy.
Throughout the last few months, a wide range of experts have concluded that tech industry spending on AI — which includes everything from data center infrastructure and energy bills to massive salaries and lobbying — was responsible for a sizeable chunk of GDP growth in the US across 2025. Though there were disagreements about how much exactly, the consensus seemed clear: AI investments are critically important to the US.
Now, though, experts at a leading bank are throwing their weight behind a different theory: that AI spending has had little impact on the economy whatsoever.
In a recent discussion on AI and the global economy with the Washington Post, Goldman Sachs’ joint-lead of global economics investment research, Joseph Briggs, argued that investment spending on AI had “basically zero” impact on US GDP growth last year.
“It was an intuitive story,” Briggs told WaPo, referring to the narrative that the nation’s economy is dependent on AI spending. “That maybe prevented or limited the need to actually dig deeper into what was happening.”
That appears to be the consensus view among the analysts at Goldman, one of the largest investment banks in the world.
“We don’t actually view AI investment as strongly growth-positive,” Goldman chief economist Jan Hatzius said in a recent interview. “I think there’s a lot of misreporting, actually, on the impact that AI investment had in US GDP growth in 2025, and it’s much smaller than is often perceived because most AI equipment is imported. That means there’s a positive entry in the investment line, but that’s offset by a negative entry in the net-exports line.”
In other words, their argument goes that previous estimates ignored where money spent on AI investments actually ends up. Nobody can dispute that the US is spending a boatload of cash on AI — but most of it flows overseas to the countries making AI chips and hardware. The money’s ultimate destination matters when measuring AI’s real impact on the US economy.
“A lot of AI investment that we see in the US adds to Taiwanese GDP, and it adds to [South] Korean GDP, but not really that much to US GDP,” Hatzius continued.
The commentary comes as Goldman has launched SPXXAI, an inelegantly-named S&P 500 index that specifically excludes stocks related to AI — a new tool for investors looking to avoid exposure to the AI bubble.
How true that is remains to be seen, as much of the data on AI’s economic impact is murky at best. One thing seems certain: the longer we spend in this AI-no-man’s-land, the less cohesive any one narrative seems to become.
More on AI: AI Could Cause Workers to Rise Up Against the Corporations Driving Them Into Poverty