Demand for new AI data centers is surging, but they can’t build themselves.
Big Tech is funneling billions into building out these specialized facilities, with the four hyperscalers, Alphabet, Microsoft, Meta, and Amazon, committing nearly $700 billion in combined capex spending this year to fund these developments.
Amazon said last month that it’s committing $12 billion to build a new AI data center in Louisiana, which will create 540 full-time jobs on site as well as 1,700 other roles for electricians, technicians, and security specialists.
Meta also invested $27 billion last year in a joint venture with Blue Owl Capital to construct its mammoth Hyperion data center in Louisiana, which is expected to consume more electricity than the city of New Orleans.
While anxiety around AI replacing white-collar jobs has reached a fever pitch, the data center boom is creating lucrative opportunities for skilled traders.
“The digital revolution requires a massive physical foundation,” Sander van’t Noordende, CEO of the world’s largest recruitment firm Randstad, told CNBC. “Ultimately, the real constraint on global tech growth isn’t solely related to a shortage of microchips, energy, or capital; it is the severe scarcity of the specialized talent required to build it.”
Between 2022 and 2026, demand for robotic technicians increased by 107%, according to a global analysis of 50 million job postings released by Randstad on Wednesday. For cooling — or HVAC — system engineers, the growth rate was 67%, and vacancies for industrial automation technicians grew by 51%.
Meanwhile, job listings for traditional skilled trade jobs such as construction workers and electricians increased by 27%, according to Randstad’s analysis.
“The debate around AI’s impact on the labor market often focuses entirely on the software side, specifically, whether generative models will displace white-collar jobs. But a critical reality is being completely overlooked: AI cannot build its own data centers,” Noordende said via emailed comments.
With roughly 12,000 data centers existing globally today, and thousands more to be built to house high-performance AI computing capacity, it’s essential to update outdated mechanical, electrical, and plumbing systems every four to six years, according to Mike Mathews, digital infrastructure leader at reinsurance firm Marsh.
Mathews noted “massive growth areas specific to labor” as a result of these retrofitting requirements, with network engineers, electricians, mechanical engineers, and plumbing and heating contractors deployed to install new liquid cooling systems.
A fourth-generation plumber himself, Mathews described these roles as “new-collar” jobs, which will see traditional white-collar and blue-collar employees working alongside each other and being valued the same.
“The data center space will be the first time when we’ve had highly compensated, high-skilled trades workers physically working next to network engineers who have college degrees. And I think it’s a great social blend that we’re going to have throughout data centers,” Mathews added.
Skills shortage driving wage premium
Randstad’s Noordende said that a “scarcity premium is taking effect” and that advertised wages for HVAC engineers have increased by roughly 10% to 15% in the past four years.
“As AI infrastructure demand outpaces a shrinking labor supply, wage growth is rising significantly for these specialized roles,” Noordende said, adding that six-figure salaries are achievable in the sector.
Specialized and technical professionals moving into high-level data center roles often see a 25% to 30% pay increase, according to staffing firm Kelly Services, which shared estimates with CNBC based on internal and third-party data. It said the premium can vary by role and full-spectrum salary data is still emerging for data center positions.
Nvidia‘s CEO Jensen Huang – a central figure behind the AI data center boom – predicted in January that “six-figure salaries” are on the horizon for the workers building AI factories.
A key factor driving these salaries higher is a shortage of trade workers, with the U.S. facing a potential shortfall of 1.9 million manufacturing workers by 2033, according to 2025 data from the National Association of Manufacturers.
Meanwhile, the U.S.-based Associated Builders and Contractors trade group estimates that nearly half a million new workers will be needed in 2027, up from the 349,000 needed in 2026.
“[The skills shortage is] a huge issue now, and it’s only going to get worse,” Gary Wojtaszek, CEO of Pure Data Centres, told CNBC.
“The good thing is AI won’t replace any of those jobs — someone needs to man those machines, so those are really important jobs, and that is a huge challenge for the industry overall.”
In order to overcome the skills shortage, businesses and governments will need to invest in training programs, Randstad’s Noordende said.
The skills shortage is also leading to cross-industry poaching as there’s a lot of overlap in the operational technical skills needed for energy, defense, and tech, according to William Self, chief workforce strategist at global professional services firm Mercer.
Earlier in March, BlackRock launched a $100 million initiative to empower the next generation of trades workers, as CEO Larry Fink stressed that capital alone is not enough to realize the $10 trillion of investment needed for infrastructure.
“[The] skills profile is evolving faster than traditional job descriptions can track,” Mercer’s Self said in a virtual press briefing on the topic.
“Given the supply shortage, the companies that win this talent race will be the ones investing in both traditional recruiting and non-traditional workforce development. I’m seeing apprenticeship programs, community college partnerships, military veteran pipelines, and even internal talent academies to grow companies’ own talent.”
‘Hazard pay rate’
But the boom in demand for key AI infrastructure jobs also faces several hurdles, including an aging demographic and geographical constraints.
Roughly 1 in 4 workers globally is nearing retirement age, and the talent pool is not being replenished fast enough, according to Noordende.
“Unlike software developers who can often work remotely, skilled trades possess very low geographic mobility,” he said.
“An equipment technician or construction worker must be physically on-site. When a company builds a new AI data center or manufacturing plant, it is a region-changing event that can instantly exhaust local talent pools.
What will be challenging to predict going forward is how much of a “hazard pay rate” will be built into salaries in the future, Mercer’s Self said.
Earlier this month, Amazon Web Services’ data centers in the United Arab Emirates were targeted by Iranian drone strikes. As the conflict in the Middle East unfolds, “you could also imagine a certain psychic burden for people who are working in facilities that they know very well might be hot targets or bad actors, which could additionally increase the types of compensation packages that we would see to lure this population to these centers,” Self added.
