Google is funding 300,000 electricians and welders, because the AI boom is running out of them

The AI boom has a problem no amount of capital can solve on its own: there are not enough electricians, welders, and pipefitters to build it. Google’s answer is to pay to train them.

Its philanthropic arm, Google.org, said it is committing $50M to prepare more than 300,000 skilled-trade workers across over 20 US states, channelling the money through 14 labour unions and four trade associations.

Google is candid about why. The work it is funding, in its own words, is “the kind of work that goes into building and maintaining data centres”: the electricians and fibre technicians wiring “advanced network grids”, and the welders and pipefitters fitting the “complex cooling systems” that keep AI servers from overheating.

Hundreds of thousands of such roles sit open across the country, and industry projections cited by Google estimate that 2.1 million skilled-trade jobs could go unfilled by 2030.

The bottleneck behind the build-out

The 💜 of EU tech

The latest rumblings from the EU tech scene, a story from our wise ol' founder Boris, and some questionable AI art. It's free, every week, in your inbox. Sign up now!

Google is not alone, and that is the real story. Big Tech is pouring hundreds of billions into data centres, from Meta’s $200bn Louisiana campus to the $1.4tn utilities plan to power them, only to run into a shortage of the people who physically build them.

In the past week alone, Meta announced a $115M skilled-trades training programme and Anthropic a $150M fellowship; OpenAI has been working with building-trades unions on its data centres. Labour, not chips or capital, is emerging as the constraint that could slow the whole expansion.

Google’s package, drawn from its AI Opportunity Fund, backs programmes run by the electrical workers’ etA, the building trades’ TradesFutures, the plumbers’ and pipefitters’ training fund, and the sheet-metal workers, each modernising apprenticeships and folding AI tools into the training.

It arrives alongside eight workforce-development policy proposals Google says it is endorsing. Since 2022, the company says it has put more than $1bn into skilling worldwide.

There is obvious self-interest here: the companies straining the labour market are the ones now funding it, and a $50M grant is a rounding error against the hundreds of billions going into the data centres themselves. The shortage also has a political dimension Google’s announcement skirts, the trades gap has widened under the Trump administration’s immigration crackdown, which has hit construction harder than any other sector.

Still, the shift is telling. After two years of treating AI as a story about chips, models, and capital, the industry is conceding that the binding constraint may be something far more old-fashioned: whether enough people know how to wire a building.