Ben Horowitz, co-founder of Andreessen Horowitz, highlighted a critical labor shortage that could impede the expansion of AI infrastructure. He noted that only 2% of electricians in the United States possess certification for DC power, a requirement for modern high-density AI data centers.
Current AI facilities are transitioning to high-voltage DC systems, including 800V setups, to support power densities of 120 to 140 kW per rack. In contrast, traditional data centers operate at 10 to 14 kW per rack using standard AC configurations.
The US Bureau of Labor Statistics anticipates approximately 81,000 new electrician openings annually through 2034, with a projected 9% employment growth. McKinsey estimates that the US will require 130,000 additional trained electricians between 2023 and 2030 to meet infrastructure demands.
To address these gaps, Google and Meta have collectively committed more than $165 million toward training programs for electricians and skilled trades. However, experts suggest that the labor shortage may remain a significant bottleneck for project timelines due to the time required for specialized certification.
