AI data center boom puts skilled trades in the policy spotlight
Former Sen. Saxby Chambliss says a new Meta-backed training effort shows AI growth depends on electricians, welders and other trades workers.
By Sofia Marchetti · World Affairs Correspondent
3 min read
A new private-sector training program for skilled trades is drawing attention to a bottleneck in the buildout of artificial intelligence: the workers needed to construct data centers and energy infrastructure. Former Sen. Saxby Chambliss, a Georgia Republican, argued in a Fortune commentary that the United States cannot treat AI leadership as a software contest alone.
Meta, the National Urban League, Associated Builders and Contractors and CBRE announced America’s Workforce Academy last week, according to Chambliss. He described it as a $115 million effort to train Americans for skilled-trade jobs at no cost, pay participants during training and connect graduates to jobs building AI infrastructure, largely data centers.
The first training sites are set to open this year in Louisiana, Ohio, Indiana and Texas, Chambliss wrote. Graduates are expected to receive an industry-recognized credential that they can use beyond the initial job placement, he said.
Data centers need more than chips
Chambliss, who served in the Senate from 2003 to 2015 and was vice chairman of the Senate Select Committee on Intelligence, framed the issue as a national security concern. He said years of intelligence briefings focused on threats such as Chinese intellectual-property theft, weak supply chains and espionage, but not on the trades workforce needed to sustain U.S. technology capacity.
His central argument is that AI systems depend on a chain of physical assets: chips, data centers, electricity and the grid. Electricians, welders, pipefitters, line workers and other tradespeople build and maintain much of that chain, Chambliss wrote.
Chambliss also pointed to China’s energy buildout as a competitive risk. He said the Chinese Communist Party is adding power and transmission capacity at a pace the United States has not matched in decades, and noted that before two new reactors came online at Plant Vogtle in Georgia, the U.S. had gone about 30 years without building a nuclear reactor from the ground up.
Labor shortage becomes an AI constraint
The construction labor gap is already large, according to industry groups cited by Chambliss. Associated Builders and Contractors has said the construction industry needs nearly 350,000 additional workers in 2026 to meet demand.
Chambliss also cited NovaArc Technologies for the figure that the average American welder is now 55 years old. JLL has warned that more than 2 million skilled-trade jobs could go unfilled by 2030.
Chambliss called America’s Workforce Academy the largest private-sector skilled-trades commitment with a job guarantee in U.S. history. He said the program’s structure matters because participants receive a conditional job offer before training, with employment tied to completing the course.
That model, Chambliss argued, links training to actual demand rather than broad government workforce goals. He said private companies have a direct financial stake in getting the program to produce workers for projects they need built.
Chambliss urged government to support similar efforts by speeding permits for energy projects, allowing trade credentials to transfer across state lines and expanding Pell grants to short-term credential programs. He also said Washington should encourage more companies to create training programs tied to specific jobs rather than launch a broad federal initiative.
Chambliss serves on the national security advisory board of the American Edge Project. His commentary said AI infrastructure should be seen as a modern form of American manufacturing, with data centers and power facilities creating stable work for people without college degrees.
This story draws on original reporting from Fortune.