AI Data Center Boom Faces a Critical Hurdle: A Shrinking Workforce
Table of Contents
The rapid expansion of AI data centers is hitting a snag-a dwindling pool of qualified workers.
- demand for power from AI data centers in the U.S. could reach 106 gigawatts by 2035, a 36 percent increase from previous estimates.
- A shortage of engineers, technicians, and skilled construction workers is emerging as a major constraint.
- companies are broadening recruitment efforts, seeking talent from industries like nuclear energy, the military, and aerospace.
- Technical colleges and industry partnerships are ramping up training programs to address the skills gap.
The artificial intelligence revolution is demanding an unprecedented amount of computing power, and that means building a lot of data centers. But a December report from BloombergNEF reveals a looming problem: the U.S. could need 106 gigawatts of power for these facilities by 2035-a significant 36 percent jump from earlier projections.The challenge isn’t just securing that power; it’s finding enough people to actually build and maintain these massive facilities.
The growing skills Gap
The power grid engineering workforce is currently shrinking, and data center operators are struggling to find trained electrical engineers. Laura Laltrello, the chief operating officer for Applied Digital, notes that demand has recently accelerated for civil, mechanical, and electrical engineers, as well as construction management and oversight positions. Applied Digital is currently building two data center campuses near Harwood, North Dakota, which will require 1.4 GW of power when completed.
“As we anticipate a shortage of conventional engineering talent, we are sourcing from diverse industries,” says Laltrello. “We are finding experts who understand power and cooling from sectors like nuclear energy, the military, and aerospace.Expertise doesn’t have to come from a data center background.”
The need extends far beyond engineers. According to the Association for Computer Operations and Management’s (AFCOM) state of the Data Center survey, 78 percent of respondents reported difficulty finding qualified data center professionals. “Demand for each category is rising significantly faster than supply,” says Hawkins.
Training the Next Generation
Technical colleges and applied education programs are emerging as key to workforce progress. These programs focus on hands-on skills, facilities operations, power and cooling systems, and job readiness.Texas, with its surge in data center construction, is seeing a proliferation of these programs, including the SMU Lyle School of Engineering’s Master of Science in Datacenter Systems Engineering (MS DSE) in Dallas, which blends electrical engineering, IT, facilities management, business continuity, and cybersecurity. Dallas College and Texas State Technical College near Waco also offer 12-week AI data center technician programs.
“Technical colleges are driving the charge in bringing new talent to an industry undergoing exponential growth with an almost infinite appetite for skilled workers,” says Wendy Schuchart, an association manager at AFCOM.
Industry leaders are also stepping up. microsoft’s Datacenter Academy is a public-private partnership with community colleges in regions where Microsoft operates data centers. Google supports local nonprofits and colleges offering IT and data center operations training,and Amazon offers data center apprenticeships. The Siemens Educates America program has surpassed 32,000 apprenticeships across 32 states, with a commitment to train 200,000 electricians and electrical manufacturing workers by 2030. The National Electrical Contractors Association (NECA) operates the Electrical Training Alliance, while the Society of Manufacturing Engineers (SME) offers ToolingU-SME, and Uptime Institute Education programs are also working to accelerate technician and operator readiness.
“Every university we speak with is thinking about this challenge and shifting its curriculum to prepare students for the future of digital infrastructure,” said Laltrello.”The best way to predict the future is to build it.”
