AI needs hands: the U.S. faces a 1.7 million-job shortfall in the skilled trades

Behind the stock market's expectations for the profits of artificial intelligence lies a problem that rarely makes the headlines about chips and models: someone has to physically build the data centers and sustain the supply chain around them, and there are fewer and fewer qualified hands to do it. A report published by the Alliance for America's Skilled Trades warns that the United States will need to fill 1.7 million skilled-trade vacancies a year through 2035, while current training programs produce only 55 workers for every 100 needed.

The figure matters because the alliance behind it is no marginal outfit: it was founded by BlackRock, Ford, Google and Carhartt, and has recently added Meta, Microsoft, Nvidia and Waymo. The fact that the very companies leading the AI race are joining a coalition focused on electricians, welders and HVAC technicians sums up the problem neatly: the AI infrastructure boom depends on trades that have spent years failing to attract a new generation.

The report details that employment of electricians, HVAC technicians and industrial machinery mechanics will grow two, three and as much as five times faster, respectively, than U.S. occupations as a whole. At state level, the projections for 2034 are specific: Arizona will need 29,000 electricians, Texas 86,000 auto mechanics and North Carolina 19,000 carpenters. The source does not explicitly link those state projections to the location of data centers or semiconductor plants.

The concept framing all of this is what the report calls the 'essential economy': construction, utilities, agriculture, transportation, oil and gas, and equipment manufacturing. According to the figures cited, it accounts for $12 trillion of U.S. GDP, 95 million jobs and 3 million businesses. It is, in other words, the physical base underpinning any AI-driven growth narrative, and it is currently suffering labor shortages because of its ageing workforce and Generation Z's lack of interest in work seen as too physically demanding.

On top of this comes a further paradox noted in the article: even workers already in these jobs do not always have the skills needed to adapt as AI is integrated into the industrial supply chain itself. A PRT Staffing survey cited in the article indicates that 17.4% of manufacturing companies already report staff shortages, and projects that 3.8 million manufacturing jobs will need to be filled over the next decade.

Ford CEO Jim Farley stressed in a statement that 'no company can close the skilled trades gap alone' and that training only 55 of every 100 workers needed 'puts economic growth at risk' and leaves many people without a clear path to a decent career. Dirty Jobs host Mike Rowe, for his part, pointed to a structural cause: the disappearance of shop class from American high schools over decades, which, combined with stigmas and stereotypes about manual work, has left younger generations estranged from these trades. Through his MikeroweWORKS foundation, Rowe funds vocational scholarships to train the next generation of plumbers, welders and electricians, in what he describes as an 'AI-proof portfolio of careers.'

For those tracking investment in AI infrastructure, the underlying message is significant: the bottlenecks in AI deployment are no longer just chips, power or permits, but also the availability of electricians and technicians able to physically build and maintain data centers. That tech giants such as Meta, Microsoft, Nvidia and Waymo are now joining carmakers and asset managers to tackle this problem suggests the sector is beginning to treat the shortage of skilled labor as a strategic risk every bit as real as the shortage of computing capacity.