Paul Krugman is arguably the world's most influential economics columnist, and his Nobel Prize credentials are beyond dispute. But his track record on the economic value of new technologies is spotty. In 1998, he famously predicted the internet's impact would be no greater than the fax machine's. He was spectacularly wrong. The internet reshaped supply chains, slashed communication costs, and created trillions in value—not to mention the consumer surplus from e-commerce and high-speed access that researchers have since quantified.
Now Krugman has turned his skeptical eye to artificial intelligence, arguing that AI is a technology "of, by and for oligarchs" that does little for ordinary Americans. He points to soaring corporate investment with surprisingly few jobs created. But a closer look at the data suggests he is repeating his old mistake.
Construction jobs are booming
Krugman claims that the data center buildout has barely moved the needle on nonresidential construction spending. But that's a misleading metric. Look at employment instead. Since late 2022, when the AI investment wave began, construction employment has risen steadily, both in absolute numbers and as a share of the workforce. This is despite the Trump administration's mass deportations, which removed many undocumented workers—about 13% of the construction workforce—and had knock-on effects that cost native-born workers their jobs too. Even so, the sector kept adding positions.
The gains are concentrated precisely in the trades needed for data centers: electricians, plumbers, and sheet metal workers. Real wages for construction workers have climbed since mid-2022, and the increase relative to the national average has been especially sharp this year. Goldman Sachs estimates that sustaining the buildout will require 500,000 new construction and trades jobs by 2030. Virginia, the epicenter of data center activity, has documented thousands of direct jobs from the boom.
These are not just numbers on a spreadsheet. They represent steady paychecks for electricians in Northern Virginia, plumbers in Texas, and sheet metal workers in Ohio. That's a tangible benefit that flows to working-class Americans, not just tech executives.
Tax windfalls for local communities
Beyond jobs, data centers are becoming a fiscal bonanza for state and local governments. They pay property taxes, sales taxes, and corporate taxes, often in jurisdictions that have struggled to fund schools, roads, and public safety. In some rural counties, a single data center can double the local tax base, allowing for new schools and better infrastructure without raising taxes on residents.
This is not "trickle-down" economics in the Reagan sense. It's direct revenue that communities can see and spend. Krugman's dismissal of these benefits ignores the fiscal reality that many local governments face.
To be sure, the AI boom has its downsides—energy consumption, water use, and the risk of job displacement in other sectors. And the benefits are not evenly distributed; some regions are left out. But to claim that AI "does nothing" for most Americans is to ignore the construction workers building these facilities and the communities that are reaping the tax rewards.
Krugman's skepticism is valuable, but it should be grounded in evidence. The evidence suggests that, at least for now, the AI boom is creating real economic value for ordinary Americans—not just for the oligarchs who own the algorithms. As the buildout accelerates, those benefits are likely to grow. Whether they will outweigh the long-term costs remains an open question, but the current picture is far from the dystopian portrait Krugman paints.
