We’re already there when looking at the eye-watering tallies for artificial intelligence investment from the four biggest hyperscalers. Microsoft, Amazon, Facebook parent Meta Platforms, and Google parent Alphabet are likely to spend a staggering $340 billion this year alone in building out their AI data centers and developing their individual products.
Looking further ahead, a recent McKinsey report pegged the collective outlay needed to meet AI demand over the next five years at around $7 trillion.
But that doesn’t mean experts share a consensus view on the impact that AI spending is likely to have on the world’s biggest economy.
Some statistics suggest second-quarter AI spending powered nearly half of GDP growth. Others say the money being spent will ultimately find its way overseas.
AI deployments will nudge growth figures higher, as users get more efficient with using the technology, according to some economists. One Nobel laureate, however, became famous for proving that information technology spending has little immediate impact on productivity statistics.
David Laidlaw, portfolio manager at Carnegie Investment Counsel, is in the former camp.
“The productivity gains associated with AI could easily be multiples of the capital expenditures themselves,” he wrote in a spring report on the impact of surging tech investment.
“The economic stimulus from the AI data center build is a secular trend that will power the U.S. economy for the next few years,” he said.
Digging into the data from the Bureau of Economic Analysis’ second-quarter GDP report underscores his prescience.
Amazon, Microsoft, Google parent Alphabet, and Facebook parent Meta Platforms spent around $69 billion over the three months ending in June, according to economist Paul Kedroksy, equating to an annual pace of $276 billion.
That’s around half of the annualized amount of domestic IT equipment spending. As a result, Kedrosky estimates the contribution of AI capex to second-quarter GDP was around 1.3 percentage points of the 3% advance estimate.
“This would be an even larger share than that contributed to GDP by AI capex in the first quarter,” he wrote in a blog post. “For practical purposes, it ate Q2 GDP growth.”
However, Samuel Tombs, chief U.S. economist at Pantheon Macroeconomics, notes that President Donald Trump’s trade policies make drawing conclusions from the AI spending wave a challenge.
“It’s difficult to judge how much of this leap (in IT investment) was due to the ongoing flood of money being poured into AI infrastructure or a one-time bump linked to companies stockpiling tech-related goods, rather than risk paying tariffs,” he said in a recent client note.
Still, AI capex has been a powerful force in 2025. Callie Cox, chief market strategist at Ritholtz Wealth Management, notes that AI spending over the first half of the year has grown by $152 billion, more than double the $77 billion increase in consumer spending over the same span.
But she adds an important caveat.
“You could say, okay, ‘AI is propping up the economy and is adding more than this component of the economy that is so large and so dominant,’” she told Ritholtz’s Josh Brown on The Compound and Friends podcast earlier this week. “Or you could say ‘Consumer spending is really setting a low bar.’ It’s stalling out,” she added. “I think it’s the latter.”
“AI capex is going gangbusters, but can it prop up the economy? I’m not so sure,” Cox said. “I don’t think you have a thriving U.S. economy without the consumer.”
Peter Berezin, chief global strategist at BCA Research, is also skeptical of the idea that AI investments are a pure driver of U.S. growth.
The combined spending of Amazon, Microsoft, Google, and Meta over the past year is equal to around 1% of domestic GDP, he argued in a report earlier this week, but noted that “most of this capex consists of spending on Nvidia chips and other tech equipment, much of which isn’t manufactured in the U.S.”
“Although it is possible that some of this spending will be re-shored back to the U.S., it will take a while for that to happen,” he added. He noted that construction spending in tech-related manufacturing and data center construction has been trending down and “employment in computer manufacturing and related sectors is near record lows.”
His final citation was from Noble Prize-winning economist Robert Solow, creator of the Solow paradox, which argues that tech investment doesn’t immediately make workers more efficient.
“You can see the computer age everywhere but in the productivity statistics,” Solow wrote in a 1987 New York Times review of “The Myth of the Post-Industrial Economy.”
Is it different this time? Maybe ChatGPT will know the answer.
Write to Martin Baccardax at martin.baccardax@barrons.com