Yves here. In posts and links, we have identified many reasons why AI will backfire against its economic and operational promises. On this list are major data center needs, from political opposition to lack of sufficient cooling water to grid capacity limitations. This post explains one more: too few skilled workers.
By Robert Rapier, a chemical engineer with 25 years of international experience in the chemical, oil and gas, and renewable energy industries who holds several patents related to his work. Originally published at OilPrice
- Goldman Sachs estimates that US data center energy demand will more than double from 31 GW in 2025 to 66 GW in 2027, requiring a massive build-out of grid infrastructure.
- The US energy sector needs about 510,000 additional workers by 2030, according to Goldman, as a wave of experienced construction workers approaches retirement.
- The labor shortage gives infrastructure contractors such as Quanta Services, MYR Group, MasTec, and EMCOR pricing power, although it also limits how quickly they can execute projects.
Most discussions about artificial intelligence focus on chips, data centers, power plants, and the need for electricity. All of these are important. But another dilemma is beginning to emerge, and it may be an underappreciated challenge.
The AI boom needs electricians.
It also requires line workers, substation technicians, grid engineers, mechanical contractors, welders, construction workers, and dispatch technicians. These are not jobs that can be filled quickly with a software update or a new funding round. They need training, experience, and a stable labor pipeline that the energy sector does not currently have in abundance.
That’s an important reminder that the AI boom isn’t just a digital story. It is also a matter of physical infrastructure.
From Chips to Construction
The first phase of AI buildout was dominated by the race for computing power. Investors are focused on semiconductors, cloud providers, and companies building large data centers to support artificial intelligence workloads.
But all those facilities must be connected to the grid. It must have transformers, substations, backup generation, cooling systems, transmission access, and the right personnel to build and maintain that infrastructure.
This is where the problem becomes more complicated.
Reuters recently reported that the rush to build data centers is exacerbating shortages of energy and grid workers, including electricians, line workers, and other engineering, procurement, and construction roles. The problem is not just that demand is increasing. It is growing while a large portion of the experienced construction workforce is about to retire.
This creates a different type of problem than most investors are used to thinking about. The device can raise money. A hyperscaler can sign a power purchase agreement. The engineer can order the equipment. But if qualified workers are not available, projects may be delayed.
Scale of Need
Goldman Sachs Research estimated that US data center power demand could increase from 31 gigawatts in 2025 to 41 gigawatts in 2026 and 66 gigawatts in 2027. That would more than double the estimated data center capacity from the end of 2025 to the end of 2027.
Meeting that demand will require massive construction of generation, transmission, communication, and support systems. Goldman also estimated that the US energy sector will need about 510,000 additional workers by 2030 to meet growing demand, while Europe needs another 250,000.
Those numbers help explain why the labor crisis can be so limiting. The energy sector does not simply compete with itself. Data centers, utilities, renewable engineers, manufacturers, industrial projects, and grid modernization programs all run many of the same workforces.
The Bureau of Labor Statistics projects that employment of electricians will grow by 9% from 2024 to 2034, much faster than the average for all occupations. It also generates about 81,000 job openings each year, many of which are tied to workers leaving or retiring.
For installers and electricians, the BLS projects employment growth of 7% over the same period, and faster than normal, with about 10,700 openings per year.
Those are good works. But it takes time to train a qualified electrician or electrician, and experienced workers are often needed for more complex jobs.
Costs, Delays, and Usage Bills
A shortage of skilled workers does not mean that AI buildout is stopping. It means that construction can be expensive and uneven.
Projects with the strongest sponsors, best locations, and clear resource coordination are more likely to move forward. Others may experience delays, cost overruns, or long communication timelines. The same pressure may also affect transmission development, renewable projects, natural gas plants, and grid strengthening work.
This has direct implications for energy policy, utility customers, and investors.
If utilities have to build additional infrastructure to service large data centers, someone has to pay for it. Regulators are already at loggerheads over whether the costs should be borne primarily by large customers who drive demand or spread more widely across all prices. Labor shortages add another layer to that argument because the high cost of construction ends up being reflected in the economics of the project.
This is one of the reasons that the data center is becoming more than just a technical issue. It is involved in resource management, infrastructure, energy markets, and local economic development.
Who Are the Beneficiaries?
For investors, the most obvious beneficiaries aren’t necessarily the AI companies themselves. The bottleneck is likely to be electrical contractors, grid builders, equipment suppliers, and utility infrastructure companies.
Firms like Quanta Services, MYR Group, MasTec, EMCOR, Eaton, and Vertiv are closer to virtual architecture than most software companies. The caveat is that labor shortages cut both ways. They can increase pricing power and backlog, but they can also limit how quickly projects can be completed. The second caveat is that the shares of most of these companies have already risen significantly over the past year.
Big Picture
AI may live in the cloud, but the cloud must be built, powered, wired, cooled, and maintained. The model runs in the clouds. The chatbot answers the question. The search result appears instantly. But behind that experience there is an array of tangible goods.
Chips may get a lot of attention. Power plants and natural gas turbines are also getting more attention as electricity demand forecasts rise. But labor force may be one of the most important obstacles.
This is not an argument for or against AI or data centers. It is an argument to understand the full supply chain behind them.
The companies best positioned for the next phase of AI buildout may not be the only ones with the best chips or the biggest data centers. It may also be utilities, contractors, equipment suppliers, and infrastructure companies that have access to skilled workers and the ability to undertake large-scale projects.
The AI boom may be digital on the surface, but underneath it lies a classic architectural challenge. And in that world, electricians and line workers may be as important as algorithms.
