What the AI Boom Is Doing for Americans

Author: Noah Smith URL: https://www.noahpinion.blog/p/what-the-ai-boom-is-doing-for-americans One-line: The AI investment boom is already benefiting ordinary Americans through construction jobs, data center tax revenues, and estimated $172B annual consumer surplus — far more than the “AI does nothing for regular people” narrative suggests.

Key claims

• Krugman’s thesis — that AI delivers no tangible benefits to ordinary Americans — is wrong on multiple empirical grounds. Smith marshals construction job data, regional economic impacts, GDP growth attribution, and consumer surplus estimates to build the counter-case.
• Construction employment rose sharply after the 2022 AI boom began. The ~900K surge in construction jobs since 2022 doesn’t prove AI causation, but Goldman Sachs estimates ~500K new construction and trades jobs will be needed by 2030 specifically to build data center infrastructure — and that process has already started.
• Loudoun County, Virginia is the premier case study of data center prosperity. The county — nicknamed “Data Center Alley” — has seen its data centers generate $240M/year in taxes without many students, building trophy schools, libraries, and roads. Data center employees earn average wages exceeding $200K. The tax base is so anomalous that the rest of Virginia has had to pass laws to prevent other counties from attracting data centers away.
• Jason Furman estimates AI investment accounts for 92% of US GDP growth in H1 2025. This is a striking single-variable attribution: the investment boom in chips, data centers, and AI infrastructure has directly driven the preponderance of economic growth during this period.
• Erik Brynjolfsson et al. estimate $172 billion in annual US consumer surplus from AI tools — comparable to the surplus from smartphones. The methodology is based on willingness-to-pay surveys asking what people would accept to give up access to AI tools.
• Incumbent industries that benefit from human inattention — banks collecting spread income from deposit inertia, credit card reward programs funded by interchange fees — will face pressure as AI agents help people optimize (cross-citing Matt Levine’s Money Stuff thesis on agents and attention).
• Supply-side effects may be more important than demand-side effects over the 10-year horizon. Even if AI doesn’t immediately raise everyone’s income, the long-run channel runs through productivity growth enabling higher real wages — the standard channel for all technology booms historically.

Notable quotes

“Krugman is wrong. Wrong, wrong, wrong. The AI boom is already helping regular Americans, and is likely to help them far more in the future.”

“Loudoun County, Virginia, which is home to the highest density of data centers in the world, has been using the tax revenues from these facilities to build trophy schools and libraries. Data center employees make an average of over $200,000 a year.”

“Jason Furman, former chair of Obama’s Council of Economic Advisers, estimated that AI investment accounted for 92% of US GDP growth in H1 2025.”

“Erik Brynjolfsson and a team of economists estimated $172 billion a year in US consumer surplus from AI tools, similar to the estimated consumer surplus from smartphones.”

“Goldman Sachs estimates that about 500,000 new construction and trades jobs will be needed by 2030 just to build the data center infrastructure to power AI.”

“The rest of Virginia’s counties have had to pass laws to prevent data centers from fleeing to Loudoun County, because the local governments there have gotten so rich from data center tax revenues.”

How it connects

• Artificial Intelligence — the core technology investment driving all the economic effects described
• Paul Krugman — the direct interlocutor whose “AI does nothing for regular Americans” thesis is the foil
• Data Centers — the infrastructure investment channel linking AI spending to construction jobs and local tax revenues
• Jason Furman — economist who provided the 92% GDP growth attribution figure
• Erik Brynjolfsson — provided the $172B consumer surplus estimate
• Consumer Surplus — the welfare measure used to quantify AI’s benefit to regular users
• Goldman Sachs — source of the 500K construction jobs estimate