AI is reshaping electricity demand faster than almost anyone predicted a few years ago, and the infrastructure race to keep up is running into a workforce problem that's easy to overlook amid the gigawatt headlines. Data centers and the power companies that supply them are now drawing from the same narrow pool of skilled technical workers — and new research shows that competition is intensifying fast.
Deloitte estimates that data center power demand in the US will jump from 47 gigawatts in 2025 to more than 176 gigawatts by 2035. Separately, US data center electricity demand is projected to reach 75.8 GW in 2026 for IT equipment, cooling, and other uses, expanding to 108 GW by 2028 and over 134 GW by 2030. Behind those numbers sits a simple physical reality: AI server racks now draw 50 to 100 kW of power, compared to just 5 to 10 kW for traditional racks, meaning each new AI data center facility places a dramatically larger load on the grid than the facilities it's replacing.
Deloitte's research center for energy and industrials analysed US job postings from 2023 through 2025 across power companies and data center developers, focusing on 39 core occupations — computer specialists, engineers, technicians, power plant operators, and line workers — that together make up more than 40% of the existing workforce in both sectors. The finding: more than one-third of new job postings in both industries targeted the exact same set of workers.
The competition is accelerating. Between 2023 and 2025, power sector job postings for these core roles rose 20%, while data center postings for the same roles surged 64% — far outpacing the 4% growth rate for these occupations across the broader economy. Some of the sharpest increases came from nuclear-related roles, with postings for nuclear power plant operators rising nearly tenfold and nuclear engineer postings up almost 60%, as utilities lean on nuclear and other firm power sources to meet AI-driven demand.
On the data center side, postings for electrical technicians climbed more than 180% — the single largest increase of any occupation in Deloitte's analysis — while postings for power plant operators rose just over 56%, reflecting the growing number of facilities managing onsite power assets and backup systems where downtime carries real financial penalties.
Interconnection delays of five to seven years are pushing some data center operators to build on-site generation — gas turbines, microgrids, and increasingly nuclear or solar-plus-storage — rather than wait on utility upgrades. That shift creates entirely new hiring categories, including microgrid engineers and on-site power strategists, roles that barely existed in most job markets three years ago.
This isn't a forecasted problem — it's a current one. In Deloitte's 2025 AI Infrastructure Survey, power sector leaders ranked competition for skilled employees as their top workforce challenge, while 63% of data center executives cited a shortage of data center-related skilled labor as their number one obstacle to securing talent. Both sectors are scaling simultaneously, both increasingly need workers with digital and AI-adjacent skills layered onto traditional technical training, and the training pipelines for these roles take years to build out — not months.
For power companies and data center developers alike, the practical implication is that traditional recruiting playbooks aren't built for this level of cross-sector competition. A lineworker or electrical technician with in-demand certifications now has genuine leverage between two rapidly growing industries bidding for the same skill set, and compensation benchmarks that felt current eighteen months ago are already out of date in many markets.
The organisations navigating this well are treating workforce planning as core infrastructure strategy — forecasting hiring needs against multi-year capacity buildout plans, investing in training partnerships that widen the talent pipeline rather than just competing harder for the existing one, and working with recruiters who understand both the utility and data center sides of the market well enough to source across the overlap rather than being boxed into one sector.
InfraRec recruits across the infrastructure roles powering the AI economy — from line workers and electrical technicians to power plant operators and grid engineers. If workforce competition is slowing down your buildout, connect with our team to build a hiring pipeline ahead of your next capacity milestone.
3 Jobs for Every 1 Qualified Engineer: What the US Infrastructure Talent Shortage Means for Hiring Managers in 2026
If your engineering roles have been sitting open longer than expected, you are not doing anything wrong. The market is simply that tight.
Data Centers to Consume 300% More Power in the Next Decade
What It Means for the US Grid and the People Who Build It
Data Centers vs. Power Companies: The 2026 Fight for the Same Workforce
AI data center growth and grid buildout now compete for the same skilled workers. Here's what Deloitte's new research reveals about the 2026 hiring squeeze.
