The Unexpected Key Players in the AI Boom: "Electricians" — A Shortage of Skilled Workers Halting Data Center Construction

The future of generative AI is not determined solely by the code written by researchers or the latest semiconductors.

To train AI models and deliver responses to users worldwide, buildings that house a vast number of servers are necessary. These buildings require facilities to supply large amounts of power stably, systems to cool heated equipment, emergency power supplies to continue operations during power outages, communication cables, fire protection equipment, and monitoring devices.

The people who install these on-site are electricians, plumbers, HVAC technicians, carpenters, steelworkers, welders, and communication equipment technicians.

One of the most lacking resources in the glamorous AI industry is "skilled human hands."

The Great Migration of Craftsmen Beginning in Michigan

Around Detroit, Michigan, the construction of data centers is beginning to significantly change the local labor market.

A 29-year-old apprentice electrician featured by The New York Times usually enters underground manholes to repair cables and electrical equipment. However, many of his colleagues are now being sent to a massive data center construction site about an hour's drive away.

The reason is clear. Jobs at data centers often offer higher wages and allowances than usual sites, with the potential to significantly increase income through overtime.

The OpenAI-related project in Saline Township, Michigan, is considered one of the largest investment projects in the state's history. The construction site requires a large number of electricians, with long hours and consecutive days of work expected.

For young skilled workers, this could be an opportunity to save enough for a down payment on a house in a short period. On the other hand, continuing a lifestyle of long working hours every day could sacrifice time with family, health, and community life.

High income or life balance?

Such choices are becoming a reality at AI infrastructure construction sites.

Why Do Data Centers Need So Many Electricians?

Electrical work is indispensable even in general buildings, but in AI-oriented data centers, the importance of electrical equipment is extraordinarily high.

Servers lined with high-performance AI semiconductors consume an extremely large amount of power. Facilities are equipped with substations, distribution boards, transformers, uninterruptible power supplies, emergency generators, storage batteries, cooling devices, and more. It is not uncommon to design power routes to be duplicated or triplicated to prevent the entire facility from stopping in case of a failure.

It's not a job that ends with just passing electricity through.

It is necessary to read blueprints, wire according to strict standards, test equipment, confirm safety, and continue inspections and maintenance even after completion. Mistakes in the work could lead to multi-billion-dollar equipment shutdowns, fires, or electric shock accidents.

Gathering unqualified newcomers won't solve the problem immediately.

According to the U.S. Bureau of Labor Statistics, many electricians acquire skills through a 4- to 5-year apprenticeship. In addition to approximately 2,000 hours of on-the-job training annually, they study electrical theory, blueprint reading, regulations, and safety management.

In other words, even if companies realize a labor shortage today, they cannot deploy a large number of sufficiently skilled personnel to the field starting tomorrow.

The supply constraint AI companies face cannot be resolved in a short period with just funding.

AI Demand Adds to Already Shortage Occupations

The shortage of electricians did not suddenly begin with the advent of generative AI.

In the U.S., the aging of skilled workers has progressed, and the number of retirees has outpaced the entry of young talent. The U.S. Bureau of Labor Statistics predicts that electrician employment will increase by 9% from 2024 to 2034, with about 81,000 job openings annually, including replacements due to job changes and retirements.

Furthermore, not only data centers but also power grid enhancements, renewable energy, semiconductor factories, electric vehicle-related facilities, and the renewal of aging infrastructure require personnel with the same skills.

According to a Reuters report, a significant portion of U.S. construction workers is expected to retire by 2031. Competition for electricians and power line workers is already fierce, with urgent data center projects increasingly offering higher-than-usual compensation.

AI investment has added massive demand to occupations that were already in short supply, rather than creating new jobs from scratch.

The "Craftsman Magnetism" Created by Huge Compensation

High wages are a welcome change in terms of improving the treatment of skilled workers.

For years, career perspectives centered on college advancement have been emphasized, while professions responsible for construction and electrical equipment have not received evaluations commensurate with their expertise and social importance.

The construction rush of AI data centers has the potential to reverse that evaluation.

On social media, reactions include "It's more rational to earn a salary as an apprentice while obtaining qualifications than to go to college and incur large student loans" and "Young people should be encouraged to pursue skilled trades that are less likely to be replaced by AI."

On the other hand, the extremely high annual income introduced in reports and on social media is not necessarily attainable with just the basic salary. It is often the result of accumulating business trips to remote areas, night shifts, hazard pay, weekend work, and a large amount of overtime.

If only the high annual income walks alone, the nature of the work, training period, physical burden, and safety risks become less visible.

Electrical work is not a job where anyone can earn a high income just by taking a short course. Qualification systems and regional regulations differ, and it takes several years to become fully qualified.

Why Tech Companies Are Funding Vocational Training

If the labor shortage is left unaddressed, the completion of data centers will be delayed. If the facilities are not completed, AI services cannot be expanded even if expensive semiconductors are purchased.

Therefore, tech companies and major investment firms have started to become suppliers of vocational training themselves.

Google is supporting electrician training organizations and has announced plans to assist in training about 100,000 existing electrical workers and 30,000 new apprentices. They aim to significantly increase training capacity over several years and expand the base of personnel supporting the power grid and data center construction.

Meta has launched "America’s Workforce Academy," investing $115 million in its first year. The program eliminates participant costs and guarantees jobs at partner companies for those who complete the training. The targets include not only electricians but also welders, plumbers, and communication and fiber optic technicians.

Microsoft is also collaborating with North American construction labor unions to support AI education, vocational training, and career development for skilled workers. Google, BlackRock, Ford, and workwear manufacturer Carhartt have also launched a corporate alliance to expand the training of skilled trades.

This is both a charitable activity and a clear business strategy.

For AI companies, the shortage of electricians is a business risk that hinders their growth before it is a social issue.

"Weeks of Training" and "Skilled Expertise" Are Not the Same

While there are high expectations for company-led training, there is also cautiousness from existing labor unions and skilled professionals.

Short-term intensive training is effective in broadening the entry into the construction industry. If companies cover tuition and living expenses and provide employment after completion, it creates opportunities for those who previously found it difficult to choose skilled trades.

However, a certificate that can be obtained in a few weeks is not the same as a skilled qualification acquired after several years of apprenticeship.

In large data centers, the same design equipment may be installed repeatedly. Apprentices placed in specific processes for extended periods may become proficient in those tasks but may not gain the broad experience needed for homes, factories, hospitals, and public infrastructure.

In the electrician community on social media, questions arise such as "Can personnel trained only in data centers move to other sites after the construction boom ends?"

Increasing the number of trainees alone is not enough.

It is crucial that those who receive training can acquire skills that are applicable in other industries without being dependent on specific companies or types of facilities.

The "Blue-Collar Revival" Welcomed on Social Media

A prominent reaction on social media to this topic is the welcome of the re-evaluation of the value of skilled workers.

There are voices saying, "The country's infrastructure has always been driven by such people," and "It's not that craftsmen have become important only in the AI era; society has finally recognized their importance."

While there is discussion about how white-collar jobs will change due to AI, the structure where demand for jobs that install, repair, and ensure the safety of equipment on-site increases is being received with surprise by many.

For the younger generation, skilled trades are being reconsidered as an active career with rewards and expertise, rather than a path chosen by those who couldn't pursue higher education.

Instead of paying tuition, they earn income as apprentices while learning. After obtaining qualifications, they can work not only in data centers but also in homes, manufacturing, energy facilities, and public infrastructure.

From these points, it is natural that opinions evaluating the AI boom as a turning point in vocational education are emerging.

The Frustration of "Not Being Able to Hire for Home Repairs"

Conversely, there is a strong reaction that the absorption of local craftsmen by data centers delays necessary everyday work.

On Hacker News, there are posts stating that multiple data center constructions have begun in the area, and electricians and HVAC technicians have moved to remote sites, making it impossible to hire for home repairs. There are also testimonials about significantly increased estimates and having to learn even simple tasks themselves.

This indicates that the profits from AI infrastructure investments are not evenly distributed.

Skilled workers earn high incomes. Construction companies secure large contracts. Municipalities expect tax revenue and investment.

However, local residents may face long wait times and high fees for home wiring repairs, HVAC equipment replacements, store renovations, and new home constructions.

Hospitals, schools, factories, homes, and data centers will compete for the same limited personnel.

Costs not reflected in AI service usage fees may manifest as increased living expenses and delays in housing supply in the region.

Concerns About Electricity, Water, and Environmental Impact

While employment is welcomed on social media, there are also voices raising concerns about the power consumption and water usage of data centers.

The project in Saline Township is said to involve enormous power demand. In Michigan, discussions continue about the impact of data centers on electricity rates, power grids, water resources, noise, and the environment, with an increasing number of municipalities temporarily halting development.

The word "job creation" alone does not justify all burdens.

Even if thousands of jobs are created during the construction period, permanent employment after completion can be significantly less than during construction. Local communities take on the burden of electricity, water, and land use for a long time in exchange for short-term construction booms.

Therefore, it becomes important to determine how much of the power equipment enhancement costs data centers will bear, whether it will affect household electricity rates, whether water usage will be disclosed, and how to design profit returns to the community.

The Biggest Anxiety Is "After the Boom"

A recurring topic in the skilled worker community is the anxiety about what happens after AI investment slows down.

While some believe data center construction will continue for more than a decade, not all planned facilities will be completed, and if AI companies do not earn sufficient revenue, construction plans may be scaled back.

If a large number of apprentices are hired in line with the construction rush and projects end simultaneously in a few years, there could be an oversupply of skilled workers. Those who bought homes based on high wages or moved to remote areas are more likely to be affected.

However, even if data center demand decreases, fields requiring electricians, such as power grid updates, home construction, domestic manufacturing return, renewable energy, storage batteries, and transportation infrastructure, remain.

The issue is not the number of people but the transferability of skills.

If companies truly aim for sustainable human resource investment, they must develop craftsmen who can work in multiple industries, not just workers who can only be used in the data centers in front of them.

Employment in the AI Era Cannot Be Discussed Only in Terms of "Disappearing Jobs"

Discussions about AI and employment have so far focused on "which occupations will be replaced by machines."

However, what the data center construction sites are showing is a more complex reality.

The more advanced the software becomes, the more physical facilities are needed to operate it. As the digital space expands, so does the work of people building power plants, power lines, substations, cooling devices, communication cables, and buildings.

While AI automates some jobs, it requires a large amount of physical labor elsewhere.

And the profits and burdens generated by that demand do not fall in the same place.

While there are craftsmen earning high wages, there are residents who can no longer request repairs. While there are municipalities welcoming investment, there are regions worried about electricity rates and water shortages. While there are companies wanting to secure personnel in a short period, there are training institutions that have taken years to cultivate safe craftsmen.

The true challenge of AI infrastructure is not simply how many data centers can be completed.

It is about maintaining local housing and public facilities even during the boom. It is about allowing workers to work without sacrificing health and life. It is about ensuring skills and employment remain even after construction ends. And it is about not unilaterally passing the costs of electricity and water onto local residents.

The future of AI is supported not only by invisible algorithms.

It is supported by people who wear helmets, pull cables, assemble distribution boards, and operate equipment at the risk of danger.##HTML_TAG