The AI conversation used to center on model launches and benchmark scores. Now it centers on concrete, copper, and kilowatts.
A new Washington Post analysis by Shira Ovide shows how the AI investment boom is quietly reshaping parts of the US economy. It brings new momentum to sectors like manufacturing and real estate, but it also creates new vulnerabilities. Data from the EIA, the Federal Reserve, and academic research backs up that picture.
One Bet Carries a Lot of Weight
Start with how much the economy leans on AI. Earlier this year, the Post reported that some economists put AI spending at half or more of US growth, while Goldman Sachs calculated the contribution at basically zero.
Both numbers can’t be right. That gap shows how hard it is to measure a boom still in progress. The exposure stays real either way.
The money also isn’t reaching workers the way past booms did. The AI boom is minting billions but not jobs in Silicon Valley, which makes the growth story look lopsided.
Factories Gain Orders but Lose the Bidding War
The industrial heartland is getting a second act, though not the one politicians promised. Bloomberg Businessweek reported that Foxconn, OpenAI, and SoftBank plan to turn a former GM plant in Lordstown, Ohio, into an AI equipment hub. The same report noted that AI is absorbing capital, power, and people, the three things every factory project needs.
So AI creates demand for cooling systems, chips, and electrical gear. It also competes with every other manufacturer for the same inputs. A plant that serves the boom thrives. A plant that doesn’t faces bidders with far bigger budgets.
Electricity Is the Real Bottleneck
The boom turns physical at the power meter. The EIA now expects US electricity demand to climb from a record 4,195 billion kWh in 2025 to 4,288 billion kWh in 2026 and 4,356 billion kWh in 2027. It says AI and crypto data centers drive much of that growth.
The global picture looks similar. The IEA projects that data center electricity use could more than double by 2030, to roughly 945 terawatt-hours.
That growth puts pressure on the grid itself. Gary Marcus has warned that AI could trigger power grid failures, and he argues that weak oversight repeats an old mistake from aviation. Utilities that planned for flat demand now have to plan for growth, and someone pays for it: ratepayers, factories on the same grid, or taxpayers.
Local effects are already measurable. An NBER working paper on data centers and local economies found that data center growth raises employment, house prices, and electricity prices in nearby areas.
Why AI Could Keep Prices High
The Fed has noticed. San Francisco Fed president Mary Daly told Axios that companies are preparing for an AI-driven chip squeeze that could lift prices well beyond the data center buildout. She called the pressure less of a one-off.
That creates an awkward policy problem. Higher interest rates slow the broader economy, but they do little to cool the firms at the center of the boom.
Companies feel the same squeeze at a smaller scale. Teams that want to keep compute bills in check can start with practical steps to cut AI cloud costs and improve performance.
What Happens If the Boom Continues or Stalls
The usual question is whether AI is a bubble. A better question is what happens to everything around the boom in each case.
- If the boom continues, the strain shows up in power bills, chip prices, and competition for industrial capacity.
- If the boom stalls, the factories, grid upgrades, and real estate built for it become stranded assets.
Either way, AI is no longer just a sector. It is a load-bearing part of the economy, and that changes the kind of risk the country carries.
Is AI Really Driving US Economic Growth?
Economists disagree. Some estimates credit AI investment with half or more of recent growth, while Goldman Sachs puts its contribution close to zero. What is not in dispute is that AI spending now shapes electricity demand, factory activity, and price pressure.
