Hundreds of millions of dollars are flowing into apprenticeships, community colleges and other training programs as AI data center construction strains the supply of electricians and other skilled workers.
High wages, bonuses and heavy overtime at large data center sites are pulling workers from other building projects, worsening labor shortages across the broader construction industry.
Meta's four-week training model has sharpened a fight with unions, which say traditional four-year apprenticeships produce safer, more versatile workers; OpenAI has pledged to work with unions on staffing.
Tech companies are also courting unions and community groups to ease opposition to data centers over land use, local impacts and resource consumption.
The boom's durability remains uncertain because data centers need far fewer workers to run than to build, raising fears of an eventual labor glut and wage pressure.
When the AI data center construction boom inevitably slows, what happens to the millions of newly trained electricians left without jobs?
Can a four-week corporate crash course truly replace years of apprenticeship without risking catastrophic failures in multi-billion-dollar AI facilities?