Everyone obsessing over gigawatts is missing the simpler constraint sitting right next to the substation: nobody can find enough electricians to build the thing.
At Bloomberg's New Economy Forum in New Delhi this week, NTT Data CEO Abhijit Dubey put a number on it that should stop every infrastructure investor mid-sentence: a qualified electrician in the US now commands roughly $200,000 in compensation, and the shortage of that labor — not turbines, not transformers — is the bottleneck "nobody talks about," according to Bloomberg. Power and talent are the two constraints data center executives named as binding right now, full stop.
This matters because the capital side of the equation is moving fast enough to make the labor gap look almost quaint by comparison. Look at what Black Hills Corp just signed to power a proposed Google data center in Cheyenne, Wyoming: $1.8 billion in new generation investment between 2027 and 2029, 564 MW of new capacity, a 2.7-gigawatt resource mix managed partly through a privately-run microgrid, with power deliveries starting in late 2027 and peak load arriving in 2030. Google is covering the full cost of that buildout so it doesn't land on other ratepayers — a structural detail that matters more than the headline number. Separately, Google also locked in a 3.59-GW contract with Constellation Energy, which is putting more than $4.3 billion behind its own fleet to serve it.
That is civilizational-scale capital deployment, announced and contracted — not yet operational. The 2027-2030 delivery timeline on the Cheyenne project is the tell: this is planned capacity, and the gap between signing the contract and energizing the load is exactly where electricians, linemen, and control-room engineers become the critical path.
Money is not the constraint anymore. People who can pull cable and commission a substation are.
I've spent months on this platform tracking interconnection queues, PPA structures, and uranium supply chains as the binding constraints on abundance. The labor market for build-out trades deserves the same attention, because it scales differently than capital. You can raise $4.3 billion in a quarter. You cannot train ten thousand journeyman electricians in a quarter — the pipeline runs through apprenticeship programs measured in years, not financing rounds measured in weeks.
Watch whether Black Hills and Constellation disclose actual commissioning timelines as 2027 approaches, and watch whether any hyperscaler starts funding electrician training programs directly, the way they've started funding reactor restarts. The company that solves the skilled-trades bottleneck first gets its gigawatts online years ahead of competitors stuck waiting on permits for people, not power.