- Digital Realty is widening hiring to bring in master electricians, HVAC techs and people with no tech background, then training them for engineer roles
- The U.S. faces roughly 81,000 openings for electrician positions alone
- The aging construction workforce will only make the shortage worse
The data center boom is driving a skilled worker shortage and operators like Digital Realty are responding by widening their hiring criteria and then training them from scratch.
Data center engineers — the technicians who keep facilities running — are the roles Digital Realty most needs to fill, said Audrey Escalante, head of global operations learning and development, in an interview with Fierce. There is no pipeline producing them ready-made. "We don't have data center engineers coming out the door," Escalante said. "We're always going to have to train people."
Construction is outpacing recruitment. North America has more than 66 GW of data center capacity on the ground, drawing more electricity than all of Germany consumes, according to JLL's midyear 2026 report. Vacancy has held at 1% for three straight years even as builders race to add supply. Constructing and operating that capacity takes electricians, mechanical and cooling specialists and technicians the industry cannot hire fast enough, and operators are pursuing three responses at once: train new workers, poaching experienced ones from one another, and leaning on automation to reduce labor requirements.
How is Digital Realty widening the hiring pool?
Digital Realty's answer is to lower the barrier at the entry point and bring recruits up to speed with training. A data center engineer maintains and inspects equipment, works with vendors on repairs and replacements, sets up new customers and troubleshoots problems on site — work that mixes keyboard tasks with hands-on labor. Rather than wait for candidates who arrive with all of that, the company assesses what a hire already knows and builds a program around the gaps.
A master electrician or a former HVAC technician brings transferable knowledge and can sometimes enter at a mid-level engineer role after some bridge training, Escalante said. Someone with no relevant experience — such as a ride-share driver with a high school education — can start as an entry-level technician and be trained up, though it takes more time and more supervised practice before the company will put that person on shift. Trainees hired as full-time employees are paid throughout onboarding.
"Instead of what used to be a very narrow door, we just keep opening the door," Escalante said. But that requires greater investment in training. "You have to make sure you're building a good strong program that sets them up for success."
The talent pool "keeps getting smaller" as employers draw down the limited supply, Escalante said. Beyond data center engineers, Digital Realty runs a graduate program for architecture engineers and an onboarding track for remote hands — technicians who perform physical tasks inside a colocation facility on customer-owned equipment. To recruit, the company uses job boards, early-career programs and partnerships with schools, and Escalante said it is exploring high school outreach to raise awareness that data center jobs exist for people who are not engineers.
Digital Realty, a data center real estate investment trust founded in 2004 and based in Austin, Texas, operates more than 300 facilities across six continents and reported $6.11 billion in revenue for 2025.
How big is the skilled-labor gap?
The numbers around adjacent trades suggest the scale of the problem. The U.S. Bureau of Labor Statistics projects employment of electricians to grow 9% from 2024 to 2034, more than double the pace of the average occupation, with about 81,000 openings a year over the decade. A McKinsey analysis put the need higher, estimating that the U.S. would require an additional 130,000 electricians, 240,000 construction laborers and 150,000 construction supervisors between 2023 and 2030.
Demographics makes closing the gap harder. About 41% of the current construction workforce is projected to retire by 2031, according to the National Center for Construction Education and Research, and nearly one in five electricians is already 55 or older, Associated Builders and Contractors chief economist Anirban Basu told Reuters Events. As for broadband, a 50-state survey for the Fiber Broadband Association and the Power & Communication Contractors Association found the U.S. short about 58,000 tradespeople to build government-funded networks.
To fill the skills gap, Meta, Google and BlackRock have together pledged more than $265 million to train electricians, carpenters and other tradespeople to build AI data centers, The New York Times reported in July. Meta's own LevelUp program offers free four-week fiber technician training, then routes graduates to jobs with contractors that staff its construction sites. Google Cloud and AWS run their own fiber training tracks.
Data center technicians earn between $45,000 and $145,000 a year depending on experience, location and certifications, and experienced electricians managing multi-site crews in Northern Virginia now command more than $200,000.
Can operators keep the workers they train?
Training a worker and keeping that worker are different problems. Digital Realty's pitch to recruits, she said, is a career rather than a single role: clear onboarding, a mentor, recognition programs and a "technical guides" system that pairs experts with new hires to transfer the "tribal knowledge" a veteran would otherwise take out the door at retirement.
Not every operator builds. Many simply buy, hiring experienced staff away from competitors — a fix that raises pay without adding a single new worker to the industry. Iron Mountain, which operates roughly 30 data centers globally, has called the practice unsustainable. "We end up hiring from each other and driving up wages, which isn't a bad thing for those that have [skills]," said Mark Kidd, EVP and GM of Iron Mountain's data center and asset lifecycle management businesses. "But it's not sustainable, because ultimately this cost gets passed back onto the user." Workers moving into data center construction can see their paychecks get 25% to 30% fatter, the Wall Street Journal reported. Kidd has said the deeper issue is structural: "We don't have a talent development system that matches the increase in needed skills."
Automation provides an alternative
If the industry cannot train or buy its way out fast enough, some suppliers are betting on needing fewer people per facility. Calix is pitching agentic AI as a force multiplier that lets operators grow without adding proportional headcount. The constraint is demographic, not technological, said Shane Eleniak, chief product officer at Calix. "It doesn't matter whether you're a small operator in Nebraska or you're Verizon. There's not enough people," he said. Schneider Electric, a major supplier of data center power and cooling, is pushing prefabricated builds that shift work off the construction site, even as it warns the shortage is global. "The number of trained people — that's a big constraint right now globally," said Rob Bunger, the company's global director of data center solution architecture, in a Fierce Network interview.
Read more about the data center labor shortage
- AI data center boom exposes labor crunch as Schneider Electric, Iron Mountain push workforce development
- Calix bets agentic AI can offset broadband's aging workforce as engineers head for the exits
- Data center labor shortage pushes costs higher as operators poach each other's workers, Iron Mountain says
- The U.S. needs 58k more workers to deploy broadband