Washington Post: AI & Tech Brief: The skilled trade shortage
The Lead Brief
The AI industry is encountering a key bottleneck in its race to supply the demands of computing power: finding and training enough electricians, lineworkers and welders to build out data centers and support the new critical infrastructure.
As our lead Energy Analyst Kathryn Clay lays out in a new report, the supply of skilled tradespersons is simply not keeping up with demand as Big Tech forges ahead with billions of dollars of data center investments around the country.
The data points: OpenAI Chief Global Affairs Officer Chris Lehane says that OpenAI’s data center expansion alone will employ 20 percent of the current workforce within the skilled trade. That doesn’t leave much of the workforce left for other Big Tech companies and investors that have committed to similarly massive data center projects.
The Bureau of Labor and Statistics estimates that there will be 10,700 new openings for electric power line workers every year through 2034.
The agency also found that there will be 46,000 annual openings for welders, cutters, solderers and brazers over the next decade.
“We have a need for digital specialists, scientists and engineers of all kinds — but what’s really lagging is our ability to hire the blue-collar workforce,” Sandra Guerra, head of human resources for Siemens Energy North America, said at a recent WP Intelligence briefing.
Addressing the shortage is especially challenging because many skilled workers, including electricians, learn their trade over four- to five-year apprenticeships. Meanwhile, the Department of Energy estimates that 40 percent of the electric utility workforce will be retirement-eligible by 2030.
Katherine recommends that C-suite executives start looking at labor availability as a core strategic variable and invest upstream in apprenticeship programs for long-term projects.
Cracking Down on AI Hype
Companies across a range of industries are attaching AI to its products and services. The technology has become both the hot new thing and a buzzword.
As interest in the marketing of AI explodes, Nikki Bhargava, a partner at international law firm Reed Smith, says the Federal Trade Commission is on the lookout for companies that are making deceptive claims about business growth and earnings potential from their AI products.
“The FTC is certainly focused on AI-washing cases,” Bhargava said in an interview. “They have already sent a number of letters targeting AI-washing companies …. We’ve seen attention from both state and federal regulators.”
These kinds of investigations from the FTC are not new, according to Bhargava. The agency has long charged so-called “greenwashing” firms and other companies that misrepresent their products under the FTC Act’s Section 5 : “unfair or deceptive acts or practices.” But AI companies have increasingly come under scrutiny in the past year.
For example, in an order filed against information services company Air AI in March, the District Court of Arizona proposed a monetary judgment of $18 million against the company for making false and unsubstantiated claims about its goods and services and its unsubstantiated earnings claims.
The takeaway: Bhargava says that companies could put themselves in legal jeopardy if they make absolute promises on rates of returns for their clients in quantitative terms.
“There’s a difference in talking about AI innovation and the ability to achieve efficiency and guaranteeing results, particularly in a monetary amount,” said Bhargava. “Because when you start to put monetary amounts on what could be achieved, then you’re driving consumers to believe that, if I do this, I’m guaranteed to make a return on my investment.”
ICYMI
Our Monday edition explored how agentic technology increases the risk of bank runs. In light of a compressed window to respond to financial crises, Rep. Bill Foster (D-Illinois) says that the Federal Reserve will need more insight into banks — both big and small.
Our Tuesday edition analyzed whether OpenAI’s new policy blueprint should be viewed as a public relations document rather than a serious policy road map. Still, AI experts expect that step-changes in models will soon redefine what is possible in the policy space.
Our Wednesday edition explored how Anthropic’s new model has changed the power dynamics around use of frontier models. Only a handful of private and public entities will have access to them and their capabilities to infiltrate the world’s most critical digital infrastructure.
Our Thursday edition broke the news that the House Foreign Affairs Committee is planning to vote on a new slate of export control bills April 22. One bill, the MATCH Act, would force foreign ally countries to align with the U.S.’s export control regime.
This newsletter is published by WP Intelligence, The Washington Post’s subscription service for professionals that provides business, policy and thought leaders with actionable insights. WP Intelligence operates independently from The Washington Post newsroom. Learn more about WP Intelligence.