US AI infrastructure investment projected at $10.3 trillion from 2025 to 2032

Data center demand straining labor, power, and crowding out manufacturing

Van Nieuwerburgh projects $10.3T, 3.6% of GDP annually

U.S. investment in data centers and AI infrastructure is on track to surpass the railroad, highway, and internet build-outs as the largest single-industry economic commitment in American history. Stijn van Nieuwerburgh, whose paper was published by the Brookings Institution and reported by The Wall Street Journal, projects total U.S. spending on data centers and related AI infrastructure at $10.3 trillion from 2025 to 2032 — averaging 3.6% of GDP annually — and on track to dwarf the investments that built the railways, the highway system, and the plumbing for the internet. Van Nieuwerburgh writes that “Never before has the U.S. economy been so dependent on the build-out of a single industry.”

The Journal noted that projecting investment is tricky, and total spending might well end up substantially lower than van Nieuwerburgh’s $10.3 trillion forecast. Even so, the money poured into data centers this year already represents an investment unprecedented in recent history. Goldman Sachs projects U.S. AI investment will reach 1.9% of GDP in 2026. The railroad boom of the late 19th century marked the last time the build-out of a single industry accounted for a larger share of the economy, according to the Journal’s analysis.

The data-center spending has reshaped the U.S. construction industry. Through July of this year, a seasonally adjusted $37 billion went to private data-center construction, roughly $9 billion more than the first seven months of 2025, according to the Commerce Department. Private construction spending on everything else — homes, apartment buildings, shopping centers — ran about $46 billion below year-earlier levels over the same period.

Hyperscaler demand is straining labor and electricity supplies. The Richmond Fed recently reported that data-center construction is tightening labor availability in its district. In Mississippi, a planned aluminum smelter that would have brought an estimated 1,000 permanent jobs chose Oklahoma instead after a data center was announced near one of the proposed sites near Vicksburg, tying up electricity the smelter needed, according to a person familiar with the operator’s decision-making. “It’s crowding out manufacturing,” said Didi Caldwell, a site-selection consultant who works with heavy industry.

Analysts at FactSet estimate the five largest hyperscalers — Alphabet, Amazon.com, Meta Platforms, Microsoft, and Oracle — will spend $4.2 trillion on capital projects over the four years ending in 2029, with a growing share financed by debt. Van Nieuwerburgh noted that tech companies often use off-balance-sheet entities to borrow from banks and private-credit firms, with limited public reporting on the deals. “If AI doesn’t generate enough revenue to service the debt raised to build data centers, the fallout could ripple through the financial system,” he said.

The investment is generating well-paid work across the trades and white-collar ranks. LinkedIn estimates that AI was behind more than 750,000 new U.S. jobs from 2023 through 2026, with a median salary of around $180,000 for AI-related job listings compared with $80,000 for all jobs. “It’s one of the robust areas of a very slow labor market,” said Kory Kantenga, LinkedIn’s head of economics for the Americas. Employers added 117,000 data-center jobs since the start of 2024, separate from construction roles. In the Washington, D.C., area, the number of unionized electricians rose from 9,000 to 17,500 in recent years, according to Don Slaiman, political coordinator of IBEW Local 26. Kwaku Afriyie, 23, left an entry-level IT role out of concern AI would take his job and now assembles data-center components as an apprentice electrician, making around $30 an hour. A more experienced colleague, Tyler Beam, 28, said he made $62 an hour plus overtime working on Amazon data-center sites and has used the earnings to buy a new GMC Yukon and begin house-hunting. “They want to put them up as fast as possible,” he said.

The AI-powered rally has driven gains in stock-market wealth. U.S. stock and mutual fund holdings totaled $63 trillion as of the second quarter, according to the Federal Reserve — nearly double the figure at the end of 2022. The gains have buoyed consumer spending even as inflation-adjusted wage growth has faltered, with the largest concentration among higher-wealth households, who hold a disproportionate share of equities. In Silicon Valley, AI money is fueling a surge in luxury-home sales. “This is the best it’s been since 2000,” said Ken DeLeon, a real-estate agent who recently listed a five-bedroom home for $9.9 million and accepted an offer of more than $13 million within two weeks from an AI entrepreneur.

Demand for data-center equipment is feeding through to consumer prices. Import prices for computers, peripherals such as hard drives, and semiconductors were 20% higher in August than a year earlier. Those costs are pushing up prices for consumer electronics such as iPhones and gaming consoles, contributing to broader inflation.

Federal Reserve officials have flagged the spillover effects. Chicago Fed President Austan Goolsbee said data-center investment is pushing up wages in related fields. Fed Chairman Kevin Warsh named borrowing by hyperscalers as one factor pushing long-term interest rates higher, which has made homeownership less affordable for millions of Americans. Electricity bills have surged in some regions with concentrations of data centers.