US Would Be In 'Recession' Without AI, As Spending Goes 'Through The Roof,' Comprising 45% Of S&P 500 CapEx

New data reveals the US would be in a recession without AI. Tech spending hits a record 45% of S&P 500 Capex as data center investment soars.

The United States economy is now dependent on a historic surge in artificial intelligence (AI) investment to stay afloat, according to recent market analysis. New data reveals that without the massive capital infusion into AI infrastructure, the U.S. would currently be in an economic recession.

Recession Is Imminent Without AI Spending?

Analysis by The Kobeissi Letter, utilizing data from BofA Global Research, highlights that AI-related spending was the dominant force behind the nation’s economic performance in the first six months of 2025.

AI spending contributed 62.5%, or a full percentage point, to the total 1.6% U.S. GDP growth during that period. “Without AI, the US would be in a recession,” The Kobeissi Letter noted on X.

The US economy is now BUILT on AI:

US private nonresidential fixed real investment in data centers has soared +300% over the last 3 years.

During the same period, inflation-adjusted nonresidential investment in structures excluding data centers has barely seen any growth.… pic.twitter.com/ryyUSUwI30
— The Kobeissi Letter (@KobeissiLetter) November 28, 2025

Spending On AI Is Bigger Than The Dot-Com Bubble

This economic reliance is driven by what analysts are calling “historic” levels of tech spending. According to data from Topdown Charts and LSEG, technology and related stocks now comprise a record-breaking 45% of all S&P 500 capital expenditure (Capex).

This figure has risen nearly 20% points in the last decade, surpassing the peak of approximately 39% seen during the 2000 Dot-Com Bubble. The charts illustrate a stark divergence between the “new economy” and traditional sectors.

While real private nonresidential fixed investment in data centers has soared nearly 300% over the past three years, inflation-adjusted investment in traditional structures—such as offices, hotels, warehouses, and factories—has remained almost flat.

US tech CapEx spending is historic:

A record 45% of S&P 500 CapEx now comes from technology and related stocks.

The percentage has risen ~20 points over the last 10 years.

By comparison, the peak in the 2000 Dot Com Bubble was ~39%.

Meanwhile, the commodity sectors' CapEx… pic.twitter.com/mSb1UuTeJk
— The Kobeissi Letter (@KobeissiLetter) November 27, 2025

The AI Demand Debate

Meanwhile, as tech spending goes “through the roof,” the capital expenditure weight for commodity sectors in the S&P 500 has halved since 2015 to just 15%, according to Kobeissi’s post, hovering near its lowest recorded level in the past 45 years.

Peter Andersen, CIO of Andersen Capital Management, has warned of massive "overbuilding" amid questionable demand, noting that major tools like ChatGPT are not "locking up" from overuse.

However, Futurum Equities strategist Shay Boloor argues this cycle differs from the dotcom bubble, contrasting the idle fiber optics of 1999 with today’s 80% GPU utilization.

This is why the AI cycle is nothing like the dot-com era.

Early-2000s fiber ran at roughly 7% utilization because the industry built far ahead of demand that never showed up. The physical layer scaled faster than the software.

Today is the inverse. $NVDA clusters inside $MSFT,… pic.twitter.com/4A0bHuifLX
— Shay Boloor (@StockSavvyShay) November 26, 2025

Here is a list of some AI-linked ETFs that investors can consider:

ETF Name	YTD Performance	One Year Performance
iShares US Technology ETF (NYSE:IYW)	23.64%	23.81%
Fidelity MSCI Information Technology Index ETF (NYSE:FTEC)	20.45%	20.25%
First Trust Dow Jones Internet Index Fund (NYSE:FDN)	9.50%	10.26%
iShares Expanded Tech Sector ETF (NYSE:IGM)	26.37%	27.53%
iShares Global Tech ETF (NYSE:IXN)	22.90%	23.94%
Defiance Quantum ETF (NASDAQ:QTUM)	29.69%	51.62%
Roundhill Magnificent Seven ETF (BATS:MAGS)	22.43%	28.14%

Disclaimer: This content was partially produced with the help of AI tools and was reviewed and published by Benzinga editors.

Photo courtesy: Shutterstock