The AI infrastructure boom is shaping up to be the largest economic bet in American history, with data center investment already exceeding the combined spending on canals, railways, and power grids, creating jobs and wealth while simultaneously driving up inflation.
The artificial intelligence infrastructure expansion is poised to become the biggest economic wager in US history, far surpassing the investment scale of major American infrastructure projects such as railways, the highway system, and the underlying infrastructure of the internet.
According to the latest estimates from Brookings Institution economist Stern Van Nieuwerburgh, total investment in data centers and supporting AI infrastructure between 2025 and 2032 is projected to reach $10.3 trillion, with average annual spending accounting for 3.6% of GDP. The US economy has never been so dependent on a single industry's infrastructure expansion.
This investment is reshaping every sector of the economy, creating hundreds of thousands of jobs and spawning new billionaires. At the same time, it carries enormous risks: much of the investment is supported by debt. If growth suddenly stalls, shockwaves would ripple across the entire US economy. Investment forecasts carry uncertainty, and the final total expenditure could very well fall far short of expectations. Even so, the scale of capital flowing into data centers this year is unprecedented in modern times.
Goldman Sachs' latest estimate puts US artificial intelligence investment as a share of GDP at 1.9% in 2026. The last time a single emerging industry's expansion claimed such a high proportion of the economy was during the railway boom of the late 19th century.
The massive influx of capital is having significant impacts across multiple dimensions of the economy: a flood of construction funds is pouring into data center projects, providing a bright spot amid an otherwise sluggish construction sector. US Commerce Department data shows that as of July this year, seasonally adjusted private data center construction spending reached $37 billion, roughly $9 billion higher than the first seven months of last year. Meanwhile, spending on all other private construction projects — residential, apartment, and shopping center — fell by approximately $46 billion year-over-year in the first seven months of this year.
Major cloud providers are continuously absorbing scarce labor and electricity resources, driving up operating costs for other businesses. A recent report from the Federal Reserve Bank of Richmond noted that data center construction is exacerbating labor shortages in its district. According to people familiar with the matter, Mississippi was originally expected to land an aluminum smelter last year, a project projected to bring 1,000 permanent jobs. But a new data center was announced near the proposed site in Vicksburg, consuming the electricity the smelter needed, and the smelter ultimately relocated to Oklahoma.
The pressure isn't just from electricity: data center projects in many locations are driving up land costs. Industry site selection consultant Didi Caldwell said: "It is squeezing out the development space for manufacturing."
Financing and risks
FactSet data shows that analysts estimate the five major cloud giants — Alphabet, Amazon, Meta Platforms, Microsoft, and Oracle — will collectively spend $4.2 trillion in capital expenditures over the four years through 2029, with an increasing share of that funding relying on debt. If this boom turns out to be a bubble that bursts, spending of this magnitude would pose risks to the financial industry. Van Nieuwerburgh noted that tech companies often borrow from banks and private credit institutions through off-balance-sheet entities, transactions that involve little public disclosure. He added that this model makes it difficult for outsiders to assess the true financial risk. If revenue generated by AI businesses proves insufficient to repay the debt incurred in building data centers, the risk would spread through the financial system.
Labor economists remain divided on whether AI has already significantly impacted white-collar employment. But one thing is clear: this infrastructure expansion has created extreme shortages in certain roles. LinkedIn estimates that from 2023 through 2026 so far, AI-related new positions in the US have exceeded 750,000, with generous salaries: the median advertised salary for AI roles on LinkedIn is approximately $180,000, compared to a median of just $80,000 across all industries. Cory Kantenga, LinkedIn's economic head for the Americas, said: "In an overall sluggish job market, this is one of the few sectors that remains strong."
White-collar roles such as data annotators and AI engineers make up the bulk of new employment; since early 2024, employers have added another 117,000 data center positions. This doesn't even count construction jobs, many of which are non-permanent. Don Sleeman, political coordinator for the International Brotherhood of Electrical Workers Local 26, said the union's electrician membership in the Washington DC area has grown from 9,000 to 17,500 in recent years. "Many people come here to work and pay off their college student loans."
Kwaku Afriyie, 23, holds a bachelor's degree in cybersecurity and previously worked in an entry-level IT position, at one point worrying that AI would take his job. Last year he switched to becoming an electrician, and now assembles data center equipment, earning about $30 per hour as an apprentice. Senior electricians can earn twice that. Taylor Beam, 28, has spent most of the past two years building data centers, working for a company that ships prefabricated components to Amazon data center sites across the country. This electrician recently earned $62 per hour, plus overtime pay. He typically works 58 hours a week, with double pay during certain periods. "They want to build these server rooms as fast as possible."
Rising incomes are driving consumption: Beam recently bought a GMC Yukon SUV and is currently looking at houses.
Wealth effect
The AI-driven stock market rally has generated enormous paper wealth. Federal Reserve data shows that as of the second quarter, US household stock and mutual fund assets totaled $63 trillion, nearly double the level at the end of 2022. Even with weak inflation-adjusted wage growth, this trend continues to support household consumption. Wealth gains are concentrated among high-income groups, as wealthy individuals allocate more of their net worth to stocks compared to the middle class.
Nationwide home sales have remained sluggish for four consecutive years, but in Silicon Valley, AI-generated wealth is fueling a luxury housing boom. Real estate agent Ken DeLeon said: "This is the best market since 2000." DeLeon recently listed a five-bedroom home priced at $9.9 million, received seven offers, and signed a contract within two weeks for over $13 million, with the buyer being an AI entrepreneur.
Inflation
Strong demand for data center equipment — particularly memory chips — has created supply shortages, driving up the cost of tech products. In August, prices paid by importers for computers, hard drives, peripherals, and semiconductors rose 20% year-over-year. Higher import prices, in turn, put upward pressure on the costs of certain consumer goods such as Apple phones and gaming consoles, contributing to inflation. Federal Reserve Bank of Chicago President Austan Goolsbee recently warned that data center investment is driving up wages in related industries. Federal Reserve Governor Kevin Warsh suggested that borrowing by major cloud providers is one of the reasons for elevated long-term interest rates, making homeownership less affordable for millions of Americans. Electricity bills have also risen sharply in many data center-dense areas.
Comments