The most needed talent in the US AI industry: plumbers and electricians

The artificial intelligence boom is triggering an unexpected talent war. While tech giants offer millions of dollars in salaries for top researchers and engineers, the construction of AI data centers is facing a more fundamental bottleneck: the United States is severely lacking enough electricians, plumbers, and HVAC technicians to build the physical infrastructure needed to support AI operations.

According to the U.S. Bureau of Labor Statistics, the country will face an average annual shortage of about 81,000 electricians between 2024 and 2034. Consulting firm McKinsey draws an even starker conclusion: between 2023 and 2030, the U.S. will need to train an additional 130,000 electricians, 240,000 construction workers, and 150,000 construction supervisors.

This shortage is placing a real constraint on the expansion of the AI industry. According to a blog post from the International Brotherhood of Electrical Workers in May this year, some local branches "face data center projects that require two, three, or sometimes even four times their current membership numbers." Chris Madello, international representative of the United Association, which represents pipefitters and plumbers, said that data center projects currently demand more workers than any other single industry.

Tech companies have begun taking steps to address this crisis. Last spring, Google announced donations to the Electrical Training Alliance, planning to help 100,000 existing electricians upskill and train 30,000 new apprentices by 2030.

The Skilled Worker Shortage Is Longstanding

Anirban Basu, chief economist of the Associated Builders and Contractors, said the shortage of skilled construction workers in the U.S. has lasted for years. In earlier times, skilled workers would pass on their expertise to their children, but in recent years they are more likely to encourage the next generation to pursue a four-year college degree. As a result, the most highly skilled construction workers are now reaching retirement age.

"For years, the industry has been warning of a coming silver tsunami, meaning the mass retirement of these highly skilled baby boomers," Basu said. "That moment has arrived, so the industry has been dealing with an increasingly widespread shortage of skilled workers."

Data center construction must compete with other projects—residential, hospitals, factories, and energy facilities—for skilled workers. In all these sectors, the supply of workers is stretched thin. The Bureau of Labor Statistics predicts that the number of electricians will grow by 9% over the next decade, 'much faster than the average for all occupations.'

Training Speed Can't Keep Up With Demand

Dan Quinonez, an executive at the Plumbing-Heating-Cooling Contractors Association (PHCC), said, the plumbing industry is 'doing everything it can' to increase the number of available workers, but the issue needs a long-term solution and can't be solved overnight.

Worker demand also varies by industry and region. Madello said that in Northern Virginia, despite a boom in data center construction, there is no shortage of people applying to become plumbers or pipefitters. "The number of applicants always far exceeds the number we actually accept into the apprenticeship program," he explained. The number admitted is mainly determined by how many union members are expected to retire in the coming years.

However, one issue is that workers can't be trained quickly enough for data center projects. Quinonez said the work itself isn't much different from other plumbing jobs, but data centers are built on a tight schedule, with little room for delays or mistakes. Apprentices usually learn on active construction sites under experienced plumbers, but for data center projects, companies are less willing to take risks since even small errors can slow the schedule.

"Mistakes can be costly for contractors," Quinonez explained. Apprentices and trainees may need 'more rigorous training' before being assigned to help build data centers.

High Salaries Intensify Talent Flow

Charles White, responsible for regulatory affairs at PHCC, said that data center projects offer plenty of incentives for plumbers, pipefitters, and HVAC technicians. These projects typically pay more than other types of construction, partly because their tight schedules make overtime more likely for workers.

High demand encourages workers to jump jobs, as they're often drawn by longer hours and higher pay offered by data center developers. "I hear these stories all the time," Quinonez said. "You get paid quickly, since you're working for Amazon, Google, or another big tech company. So competition is everywhere."

"The competition really is fierce," Quinonez added. "Part of the problem is there just aren't enough plumbers and HVAC technicians."

In the construction industry, some of the pressure is relieved by a reserve of mobile skilled workers, who are ready to "show up whenever something needs building in a remote location." David Long, CEO of the National Electrical Contractors Association (NECA), said the organization has done well training new electricians to keep pace with retirements, but the scale and technical requirements of data center projects make it a "challenge" to get all workers onboard quickly and safely.

At present, it's unclear how long the demand for skilled workers will last after the AI boom eventually wanes. Once construction is complete, data centers usually retain a small group of full-time staff on-site, along with a network of outside contractors to handle maintenance on multiple facilities.

Risk Warning and DisclaimerThe market carries risks, and investments require caution. This article does not constitute personal investment advice, nor does it take into account the special investment objectives, financial situation, or needs of any individual user. Users should consider whether any opinions, views or conclusions herein are suitable for their particular situation. If you invest based on this information, you do so at your own risk.