The job posting did not ask for a computer science degree or a background in machine learning. It asked for a pipefitter.
Across the United States, the companies building and operating artificial-intelligence data centers have begun competing for workers they once had little reason to court: welders, plumbers, HVAC technicians and electricians. The AI boom, it turns out, is also a blue-collar hiring boom.
Maria Flynn, president and chief executive of the nonprofit Jobs for the Future, described the shift in plain terms. “These are occupations that are becoming increasingly important to the AI economy,” she told CNBC.
The pay reflects the new demand. An apprentice-level technician can expect $40,000 to $60,000 a year, Flynn said, while experienced electricians command more than $100,000. Job postings for welders and pipefitters have climbed 164 percent year over year, according to data CNBC cited. The mean minimum salary for data-center roles jumped 125 percent in a year, to nearly $208,000.
Rachel Lipson, founder of Harvard’s Project on Workforce, pointed to welders who graduate from high school and reach $65,000 to $75,000 a year within a short time, with what she called “almost unlimited overtime.”
“A lot of job opportunities today don’t require advanced degrees,” Lipson said. “We are seeing a blue collar resurgence.”
The scale of the buildout explains the wages. Electrical work alone accounts for an estimated 45 to 70 percent of data-center construction costs, according to the International Brotherhood of Electrical Workers, which has said the country will need roughly 300,000 new electricians over the next decade on top of replacing the 200,000 expected to retire. Nvidia chief executive Jensen Huang has called the AI boom “the largest infrastructure build-out in human history,” and predicted it would create work for plumbers, electricians and steelworkers.
The construction itself is everywhere. Workers have poured concrete, hung pipe and run conduit at sites in Stone Ridge, Virginia, and Eagle Mountain, Utah, raising the shells that will hold the servers powering the AI race. The work is physical, fast and, increasingly, well paid, and it has reversed part of the familiar story about AI and labor, in which the technology is cast mainly as a threat to jobs.
That reversal has limits. A large share of the new positions are tied to construction, which ends when the building does. The demand is real enough to reshape local labor markets for the workers who can reach it, CNBC reported, but many of the jobs themselves are temporary by nature.
The geography matters. An electrician or pipefitter in Birmingham, where data centers are rising fast, now has more bargaining power than a counterpart in Boston. Demand has pulled wages up unevenly across the country, concentrated in the regions willing to host the buildout.
The second catch is political. New York and Texas, two states central to the industry’s next phase of expansion, have both pressed the brakes.
New York Governor Kathy Hochul signed an executive order on July 14 imposing a one-year moratorium on discretionary environmental permits for any data center of 50 megawatts or more, making New York the first state to freeze the projects itself. Hochul said the facilities threatened to raise utility bills and strain natural resources.
Texas took a softer route. In early August, Governor Greg Abbott directed state regulators to audit new data-center proposals, and put new grid connections for large projects on hold until that review is complete. The interconnection queue for the state’s grid holds more than 1,800 projects, roughly 90 percent of them data centers, a measure of how much the state’s ambitions now run ahead of its power supply.
The local resistance runs wider still. By early August, more than 500 towns and counties across the country had passed temporary or permanent restrictions on new data centers, up from roughly 300 in late June, according to reporting by The Information. A Gallup poll cited in coverage of the backlash found that 71 percent of Americans oppose data centers near their homes.
The tension is between two futures. In one, the pipeline of approved projects keeps feeding the trades for years. In the other, the moratoria spread and the stream of new work thins out.
Lipson does not expect the pullback to bite immediately. “In the short term, the moratoria and public backlash are unlikely to have a big impact on demand for these roles, because there are so many AI and data center projects already underway or approved,” she said.
“The longer-term picture could be different,” she added. “If moratoria spread and significantly fewer new projects are approved, eventually that would flow through to labor demand — but with a lag.”
The skills themselves are moving faster than the classrooms. Some of the specialized welding the data centers require outruns what local trade schools teach, forcing employers to train hires on the job for work the students have never seen. The companies doing the hiring are, in effect, paying a premium for people willing to learn quickly.
For now, the trades are enjoying something rare: a technology boom that pays people without a college degree, on jobs the machines they help build cannot yet do for themselves. Whether the work lasts depends less on the demand for AI than on the patience of the towns being asked to host it.
