Tech Firms Push Back Data Center Timelines by Years as AI Buildout Hits 1 Million-Engineer Shortfall
Updated
Updated · PwC · Jun 12
Tech Firms Push Back Data Center Timelines by Years as AI Buildout Hits 1 Million-Engineer Shortfall
3 articles · Updated · PwC · Jun 12
Summary
Major technology firms are already delaying some data center completions by years as labor shortages, not just permits or power, become the binding constraint on AI infrastructure projects.
Up to 1 million engineers are lacking in the US, more than 80,000 electricians a year are needed over the decade, and about 500,000 new construction workers are required to meet current demand.
That gap is colliding with hyperscalers’ timelines: a full apprenticeship takes four years, some regions need six months just to assemble qualified crews, and a 250,000-square-foot data center can require roughly 1,500 skilled tradespeople.
Costs are rising as data centers, utilities, chip plants and reshoring projects compete for the same labor pool, driving wage inflation, thinner contractor benches, schedule slippage, rework and budget overruns.
The report argues investors have not fully priced workforce risk into AI infrastructure models, even though a two-year delay can materially alter returns and the US project backlog already approaches a decade of construction at current capacity.