The U.S. build-out of data centers and related artificial-intelligence infrastructure could absorb $10.3 trillion between 2025 and 2032, a scale that would make it one of the largest investment waves in the country’s history.
The estimate comes from economist Stijn Van Nieuwerburgh and was cited in Wall Street Journal reporting republished by Mint. Spread across the period, the projected investment averages about 3.6% of U.S. gross domestic product each year. The report cautions that forecasting a build-out of this size is difficult and that the final total could be substantially lower.
The spending is already reshaping construction
Through July 2026, seasonally adjusted private data-center construction spending reached $37 billion, about $9 billion more than during the same period a year earlier, according to Commerce Department figures cited in the report. The demand is creating work for electricians and other skilled trades while also competing with housing and manufacturing projects for labor, land and electricity.
Goldman Sachs estimates AI investment will equal 1.9% of U.S. GDP in 2026. The Journal compares the concentration to the 19th-century railroad boom, the last time one new industry accounted for a larger share of the economy.
Debt makes the boom a financial risk
FactSet estimates that Alphabet, Amazon, Meta, Microsoft and Oracle will spend $4.2 trillion in the four years ending in 2029. A growing portion of the wider build-out is financed with debt, including off-balance-sheet structures that provide limited public visibility.
If AI services do not generate enough revenue to cover that financing, losses could spread through banks and private-credit firms. The same investment is also pushing up demand for chips, power and specialized workers. The $10.3 trillion figure is therefore not a committed budget or guaranteed outcome; it is a measure of how large the current trajectory could become and how much of the economy now depends on it continuing.
