MW The AI boom really isn't - for the broader economy, according to these economists

By Steve Goldstein

Barclays team brushes back a popular belief about the impact of artificial-intelligence capital expenditures

Data-center spending, impressive as it sounds, isn't having a huge impact on the broader economy, one investment bank argues.

For all the spending being done by the so-called hyperscalers on data centers and microchips to enable artificial-intelligence applications, the effect on the broader economy is overstated and pales in comparison with previous investment booms.

That's the argument from economists at Barclays led by Jonathan Miller, who brushed back a popular belief on the impact of artificial-intelligence capital expenditures.

"In our view, popular discussions about the massive scale of hyperscaler capex considerably overstate the momentum this would create for aggregate GDP growth," said Miller. "Although the hundreds of billions being plowed into capex by individual hyperscalers is impressive, it pales in comparison to aggregate U.S. capex, which typically exceeds $4 trillion per year - especially given lackluster outlays elsewhere."

Estimates that the five largest hyperscalers - Meta Platforms (META), Alphabet $(GOOGL)$ $(GOOG)$, Amazon (AMZN), Microsoft $(MSFT)$ and Oracle $(ORCL)$ - will boost capital spending by about 30% by 2027 is actually a deceleration from 2024 and 2025 ramp-ups, they note. "If this is representative of the broader trend, most of the impetus to aggregate demand growth from this spending is already behind us," they say.

The ramp-up compared with past investment booms also is on the tepid side, they say. During the dot-com boom, for example, the rate of business fixed investment more than doubled over a seven-year period, or some 12% annually. During the first half of 2025, business fixed investment rose at an 8.1% annual rate. Moreover there have been a number of other episodes of double-digit growth in fixed investment, during the early 1960s and the mid-2000s.

The Barclays team take a stab at quantifying AI-related investment in GDP terms. Adjusting for the fact that some of the AI expenditure is spent on imports, Barclays estimates the GDP growth contribution during the first half of the year was about eight-tenths of a percentage point.

But what about a multiplier effect? They dismiss that idea. "Our view is that such adjustments would be inappropriate in the current context, with the economy operating near full employment and unconstrained monetary policy working to limit widening of the output gap," say the economists.

Another issue: Won't AI boost productivity growth? Possibly, but not so much by the capital deepening - again, the numbers just aren't enough - as by reducing labor inputs or enabling new ways to add value. "But the historical record suggests that it is difficult to summon such gains, even when innovation seems to be at a high rate," they say. During the 1980s, for instance, the introduction of personal computers showed only limited acceleration in productivity.

Left unanalyzed in the Barclays report is the impact on the economy from stock-market gains fueled by AI. During the first half of the year, the value of corporate equities rose by $3.1 trillion, according to Federal Reserve data, though that's down from $5.3 trillion in the first half of 2024.

The S&P 500 SPX has gained 13% this year and is sitting just below record highs.

-Steve Goldstein

This content was created by MarketWatch, which is operated by Dow Jones & Co. MarketWatch is published independently from Dow Jones Newswires and The Wall Street Journal.

(END) Dow Jones Newswires

October 15, 2025 08:47 ET (12:47 GMT)

Copyright (c) 2025 Dow Jones & Company, Inc.