In recent discussions regarding artificial intelligence (AI), we’ve explored various factors that could hinder its economic and performance aspirations. While data center necessities, political resistance, water shortages for cooling, and power grid limitations are well-known challenges, one issue that deserves further attention is the shortage of skilled workers in the industry.
By Robert Rapier, a chemical engineer with 25 years of international experience in the chemical, oil and gas, and renewable energy sectors, and holder of several patents. Originally published at OilPrice
- Goldman Sachs estimates that U.S. data center power demand will more than double from 31 GW in 2025 to 66 GW by 2027, necessitating a substantial expansion of grid infrastructure.
- The U.S. power sector is projected to require roughly 510,000 additional workers by 2030, coinciding with a wave of experienced construction workers nearing retirement.
- This labor shortage empowers infrastructure contractors like Quanta Services, MYR Group, MasTec, and EMCOR, although it also restricts the speed at which they can complete projects.
While much of the conversation surrounding AI revolves around semiconductors, data centers, and power demands, an emerging bottleneck is the need for skilled labor. This reality poses an overlooked but significant obstacle.
The boom in AI technology requires a workforce of electricians, line workers, substation technicians, grid engineers, mechanical contractors, welders, construction teams, and commissioning specialists. These roles cannot be filled swiftly through software updates or financial investments; they necessitate thorough training, experience, and a consistent pipeline of labor—factors that are currently insufficient in the power sector.
This situation serves as a crucial reminder that the AI boom extends beyond the digital realm; it is equally a tale of physical infrastructure development.
From Chips to Construction
The initial phase of AI expansion has largely focused on securing computing power, prompting investors to concentrate on semiconductors, cloud services, and companies constructing overwhelmingly large data centers for AI operations.
However, all these facilities must connect to the grid, incorporating transformers, substations, backup generation, cooling systems, transmission access, and a skilled workforce qualified to build and maintain that essential infrastructure.
The complexity of the situation escalates from here.
According to a recent report by Reuters, the rush to develop data centers exacerbates the existing shortages of power and grid workers, including electricians, line workers, and other engineering roles. The challenge lies not only in the rising demand but also in the fact that a considerable number of experienced construction workers are nearing retirement.
This presents a unique constraint that investors may not commonly consider. Utilities can secure funding, hyperscalers can sign power purchase agreements, and developers can order equipment. Yet, without access to trained workers, projects may still experience delays.
The Scale of the Need
Goldman Sachs Research estimates that the U.S. data center power demand will escalate from 31 gigawatts in 2025 to 41 gigawatts in 2026 and 66 gigawatts in 2027, more than doubling the estimated data center capacity from late 2025 to late 2027.
To meet this demand, a massive expansion of generation, transmission, interconnection, and backup systems will be required. Goldman also predicts that the U.S. power sector will need about 510,000 additional workers by 2030, with Europe requiring an additional 250,000 workers.
These projections clarify why labor shortages could limit progress. The power sector faces competition not just internally but also from data centers, utilities, renewable developers, manufacturers, industrial projects, and grid modernization programs, all vying for a similar pool of skilled workers.
The Bureau of Labor Statistics anticipates that employment for electricians will grow by 9% from 2024 to 2034, significantly faster than the average for all occupations, with approximately 81,000 job openings annually due to retirements and workers leaving the field. For electrical power-line installers and repairers, a similar trend is projected, with a 7% growth rate and about 10,700 openings each year.
While these roles promise good job security, training skilled electricians or line workers takes time, and often, the most experienced hands are needed for more complex projects.
Costs, Delays, and Utility Bills
A labor shortage does not halt the AI development; instead, it often leads to increased costs and delays. Projects with the strongest sponsors, ideal locations, and well-defined utility partnerships are more likely to proceed as planned, whereas others might face setbacks, financial overruns, or elongated interconnection timelines. Similar pressures may affect transmission upgrades, renewable initiatives, natural gas plants, and grid-resilience projects.
This situation carries significant ramifications for energy policy, utility customers, and investors alike.
If utilities must expand infrastructure to cater to large data centers, someone will ultimately bear the costs. Regulators are currently grappling with whether the financial burden should rest primarily on the major customers driving this demand, or if it should be distributed across all users. Labor shortages further complicate this debate, as rising construction costs will inevitably impact overall project economics.
This is a primary reason the data center boom is evolving into a conversation beyond technology, encompassing issues of utility regulation, construction labor, power markets, and local economic development.
Who Benefits?
For investors, the immediate beneficiaries may not be the AI companies themselves. Instead, electrical contractors, grid builders, equipment suppliers, and utility infrastructure firms stand to gain the most.
Companies such as Quanta Services, MYR Group, MasTec, EMCOR, Eaton, and Vertiv are closer to the physical buildout compared to most software firms. However, it’s essential to consider that labor shortages can have a dual impact: they may enhance pricing power and backlogs, yet they also limit project timelines. Additionally, many of these companies’ stock values have skyrocketed over the past year.
The Big Picture
AI may exist predominantly in the cloud, yet this cloud requires construction, power, connectivity, cooling, and maintenance. An algorithm runs in the cloud, a chatbot responds to inquiries, and search results appear promptly, but these digital experiences are supported by a network of physical assets.
While chips and electricity generation are receiving increased attention, the workforce may emerge as a crucial limiting factor.
This is not a critique of AI or data centers; rather, it highlights the importance of understanding the entire supply chain behind them.
Those best positioned for the next phase of AI advancement may not only be the ones with the superior chips or the largest data centers but could also include utilities, contractors, equipment suppliers, and infrastructure firms that can access skilled labor and execute significant projects.
Ultimately, while the AI boom may present itself as a digital phenomenon, at its core lies a traditional construction challenge. In this context, electricians and line workers are just as vital as the algorithms that drive the technology.
