AI Is Supposed to Replace Workers. In Construction, It Is Creating a Labor Shortage.

Artificial intelligence is usually discussed as a technology that will reduce the amount of human labor companies need. In construction, the AI boom is producing almost the opposite problem.

Building the infrastructure required for artificial intelligence is creating enormous demand for electricians, engineers, project managers, construction crews, technicians, and other skilled workers at a time when many of those occupations are already experiencing shortages. Reuters Breakingviews estimates that the U.S. construction industry is short roughly 439,000 workers as data-center investment accelerates.

That creates an important counterpoint to the usual debate over whether AI will destroy jobs. Technologies do not simply remove labor demand. They move it, often into parts of the economy that appear far removed from the technology itself.

AI may reduce demand for some forms of office work while simultaneously increasing the value of occupations responsible for building the physical systems on which AI depends.

Artificial intelligence has a very physical economy

The experience of using AI can make the technology seem almost weightless. A person opens an application, types a question, and receives a response within seconds.

Behind that interaction is an enormous industrial system.

AI requires data centers filled with specialized computing equipment. Those facilities require land, cooling systems, electrical infrastructure, backup power, fiber connections, security systems, and continuous maintenance. The electricity supplying them requires generation capacity, transmission networks, substations, and increasingly significant upgrades to regional power systems.

Every layer involves human work.

The scale of investment makes that demand increasingly difficult to ignore. Microsoft, Amazon, Alphabet, Meta, and other companies are spending extraordinary amounts on computing infrastructure as they compete to expand AI capacity. Reuters has cited estimates suggesting U.S. hyperscaler investment could reach roughly 3% of gross domestic product annually between 2027 and 2029.

That is not simply a software-development cycle. It is a construction and infrastructure boom.

And construction was already struggling to find enough workers before AI arrived.

The workers AI creates demand for are not necessarily the workers it displaces

This distinction is essential when evaluating claims that AI will ultimately create enough jobs to compensate for those it eliminates.

At an economy-wide level, both statements can be true. AI can reduce the number of people required for routine analysis, administrative work, customer service, software development, or other knowledge tasks while creating demand for electricians, power engineers, equipment technicians, and construction managers.

For individual workers, though, those gains and losses do not line up so easily.

A worker displaced from one part of the economy cannot simply step into a newly created role somewhere else. An office employee may need years of training to enter a skilled trade, while a junior analyst may have little direct path into power engineering or infrastructure work. The opportunities may also be concentrated in different industries, regions, or cities.

That is why the debate over whether technology creates more jobs than it destroys can miss the more practical question. What matters is not only how many jobs exist, but what those jobs require, where they are located, what they pay, and whether the people losing work can realistically move into them.

The AI infrastructure boom is demonstrating that technological disruption can create labor shortages and worker displacement at the same time.

Construction cannot manufacture experienced workers overnight

The shortage is particularly difficult because many of the occupations involved require meaningful training and experience.

An electrician is not created through a brief corporate reskilling program. Neither is an experienced project manager, power engineer, or technician responsible for critical infrastructure. Apprenticeships, licensing, technical education, and years of practical experience remain important.

That makes supply slow to respond when demand rises suddenly.

Reuters notes that construction employment across 15 OECD countries remained approximately 2 million workers below its 2008 level in 2025, even as overall employment expanded substantially. The industry lost significant capacity following the global financial crisis and has struggled in many markets to rebuild its workforce.

AI investment is now competing for a limited pool of workers who are also needed elsewhere.

That scarcity can produce attractive wage gains. Skilled workers who move into data-center construction may command premiums because employers urgently need their expertise.

For workers, that is an opportunity. For the broader economy, it creates another set of tradeoffs.

AI infrastructure competes with housing and everything else we need to build

The electrician installing systems in a data center could also be wiring an apartment building. The construction manager overseeing an AI facility could be managing a hospital, school, factory, or transportation project. Engineers working on power infrastructure are also needed to modernize electrical grids serving households and traditional industries.

When data-center developers can offer more attractive compensation, scarce workers follow the money.

That is how one industry's investment boom can increase costs elsewhere.

Housing is particularly important because many regions already struggle to build enough homes. Adding massive data-center projects to constrained construction markets does not simply create jobs. It changes the competition for the people, materials, land, and energy required to build everything else.

The economic benefits of AI infrastructure therefore cannot be evaluated only by counting investment dollars or construction jobs. Policymakers and companies also need to consider what other activity the boom makes more expensive or difficult.

That does not make the investment undesirable. It means workforce capacity has become part of infrastructure strategy.

Approving a data center without asking who will build it addresses only half the problem.

The opportunity is to expand the talent pool rather than simply bid up the existing one

The construction boom could ultimately produce a positive long-term effect if sustained demand encourages the industry to rebuild its workforce.

When employers believe strong demand will persist, apprenticeship programs become easier to justify. Community colleges have stronger incentives to expand technical programs. Workers become more willing to invest several years learning a trade when the earnings and career prospects appear durable. Employers can form partnerships with schools and training organizations because they expect to continue hiring.

That is where the AI infrastructure boom could create something more lasting than a temporary wage premium.

Simply competing for the existing supply of electricians, technicians, and engineers will raise labor costs without solving the shortage. Expanding the workforce requires investment in apprenticeship, technical education, career awareness, training capacity, and clearer pathways into skilled occupations.

It may also require reconsidering assumptions about which careers offer security in a technology-driven economy.

For years, young people were often encouraged to think of technology as the future while traditional trades were treated as a less sophisticated alternative. The AI economy is exposing how artificial that division can be.

Advanced computing, cloud infrastructure, and autonomous systems all depend on physical work. Electricians, construction crews, power engineers, and technicians are essential to building and maintaining the systems that make AI possible.

That is why the future of work will not divide neatly between technology jobs and everything else. AI will reduce demand in some occupations, reshape others, and create new pressure in areas far removed from software. Some of the workers benefiting most from AI investment may never write a line of code.

The more useful question, then, is not simply which jobs AI will automate, but where AI investment is creating new shortages, new forms of scarcity, and new opportunities. Right now, one of the clearest examples is construction.

FAQ's

How is AI affecting construction jobs?
AI infrastructure investment is increasing demand for electricians, engineers, construction managers, technicians, and other skilled workers needed to build and operate data centers and energy systems.

Why are data centers creating labor shortages?
Data-center construction requires many of the same electricians, engineers, technicians, and construction professionals already in short supply across housing, manufacturing, utilities, and infrastructure.

Will AI create jobs as well as eliminate them?
Yes, but the jobs created may require different skills, training, and locations than the jobs displaced. Job creation and worker displacement can therefore occur simultaneously.

Why can’t displaced office workers simply move into AI infrastructure jobs?
Many infrastructure occupations require apprenticeships, licensing, technical education, or years of experience, making transitions difficult for workers from unrelated occupations.

What careers may benefit from the AI infrastructure boom?
Electricians, power engineers, construction managers, technicians, utility workers, equipment specialists, and other skilled trades may see stronger demand as AI infrastructure expands.