A prevailing narrative that massive spending on artificial intelligence is single-handedly propping up the U.S. economy is now facing significant scrutiny. A growing number of economists are challenging initial high-impact estimates, arguing that the technology's actual contribution to economic growth last year was minimal, perhaps even close to zero.
This debate highlights the complexities of measuring a new technology's economic footprint, especially when global supply chains are involved. As companies invest billions in AI infrastructure, the true effect on U.S. Gross Domestic Product (GDP) remains a subject of intense discussion among financial experts.
Key Takeaways
- Economists are divided on how much AI spending contributes to U.S. economic growth, with some estimates ranging from over 50% to effectively zero.
- The core of the debate centers on how to account for imported components, like computer chips from Asia, which are subtracted from U.S. GDP calculations.
- Major tech companies are expected to spend a combined $700 billion on AI infrastructure this year, fueling a construction and technology boom.
- The difficulty in measuring AI's current impact suggests future debates over its long-term effects on jobs and prosperity will be challenging.
The Heart of the Disagreement
The discussion about AI's economic role intensified last year when some analyses suggested it was responsible for a substantial portion of U.S. growth. This idea was quickly adopted in both political and business circles, framing AI as a critical engine for an otherwise sluggish economy.
However, a counter-narrative has emerged, led by economists from prominent institutions including Morgan Stanley, JPMorgan Chase, and Goldman Sachs. Their analyses present a starkly different picture.
A Tale of Two Calculations
Initial reports suggested AI investment accounted for anywhere from 39% to over 90% of U.S. economic gains in 2025. In contrast, researchers at Goldman Sachs recently concluded that this spending made "basically zero" difference to U.S. economic growth last year.
The discrepancy hinges on a fundamental aspect of economic accounting: Gross Domestic Product. GDP measures the final value of goods and services produced domestically. When a U.S. company builds a data center, the money spent on imported hardware is subtracted from the GDP calculation because it benefits foreign economies.
"It was a very intuitive story," said Joseph Briggs, who co-leads global economics investment research at Goldman Sachs. According to Briggs, this intuitive appeal may have limited the incentive to "actually dig deeper into what was happening."
Imports and the GDP Puzzle
Understanding the debate requires a closer look at how GDP is calculated. If an American company sells a U.S.-made truck for $50,000 that contains $10,000 worth of foreign parts, only the remaining $40,000 contributes to U.S. GDP.
The AI Supply Chain
Analysts estimate that computer hardware, such as advanced processing chips, accounts for roughly three-quarters of the cost of a new AI data center. Many of these essential components, even for American tech giants like Nvidia, are manufactured in Asia. This heavy reliance on imports is central to the argument that AI's contribution to U.S. growth is overstated.
Economists who argue for a lower impact point out that a significant portion of the billions being spent on AI data centers flows overseas to pay for this equipment. Therefore, while the investment is massive, its net effect on the U.S. economy is significantly diluted.
"This is a big deal, but not the be-all and end-all," noted Joseph Politano, an economic analyst and writer of the Apricitas Economics newsletter. He calculates that AI-related spending contributed a more modest 0.2 percentage points to the overall 2.2 percent U.S. economic growth last year. Politano added that while stripping out import costs is not standard for every sector, AI "is a special case, because so much is imported."
An Undeniable Spending Boom
Regardless of the precise GDP calculation, the scale of investment in AI is undeniable. Across the country, massive data centers are being constructed at a rapid pace, creating a surge in demand for land, construction labor, cement, and power infrastructure.
"The two engines of today’s economy are the AI ecosystem and wealthy consumers."
– Tom Barkin, Richmond Fed President, in a January speech.
Five of the largest U.S. technology companies are projected to spend a combined $700 billion on AI infrastructure and other major projects this year alone. This figure is roughly equivalent to the entire economy of Sweden. This spending has tangible effects, from creating construction jobs to boosting the stock portfolios of Americans invested in tech companies.
Some experts believe the narrow focus on GDP calculations misses the bigger picture.
Joe Brusuelas, chief economist at the consulting firm RSM US, suggests that while the critics may be correct on the technical details, they are "missing something much larger happening in the economy." He argues that the debate overshadows the more profound questions about how this massive capital investment is reshaping industries and labor markets.
Navigating the Fog of Data
The difficulty in measuring AI's current impact foreshadows future challenges. As the technology becomes more integrated into business and daily life, determining its precise role in productivity, job creation, or displacement will become even more complex.
Hannah Rubinton, an economist at the Federal Reserve Bank of St. Louis, co-authored an analysis that found AI-related business spending contributed about 39% of U.S. economic growth in the first nine months of 2025. While she stands by her calculation, she acknowledges it likely represents the maximum possible impact, as it included broader tech spending that was not exclusively for AI.
Rubinton agrees that the narrative of AI propping up the entire economy was likely overstated. She explained that this idea gained traction early last year when the economy seemed fragile and data was skewed by other factors. Once that story took hold, it was difficult to displace.
"It’s not like AI is propping up the economy," Rubinton concluded.
As economists continue to refine their models and more data becomes available, the true measure of AI's economic contribution will become clearer. For now, leaders in business and government are left to navigate conflicting numbers while making high-stakes decisions about the future. As Brusuelas put it, "We’re all trying to peer through the fog here on what is driving growth."