The Talent Pipeline Is Infrastructure

AI campuses require more than a construction workforce. They need regional systems that continuously produce, attract and retain technical talent.

A construction workforce gets an AI campus online.

An operating workforce keeps it there.

After the cranes leave and the ribbon is cut, a different phase begins. Electrical systems must be monitored. Cooling equipment must be maintained. Fiber connections must remain available. Security threats must be managed. Hardware must be replaced, software updated and failures diagnosed before they become outages.

This work continues every hour of every day.

The largest AI facilities may be highly automated, but automation does not eliminate the need for skilled people. It changes what those people must know.

Tomorrow’s data-center workforce will need to understand equipment, software, controls, energy, networks and operational risk often at the same time.

That creates a different kind of infrastructure challenge.

A region cannot wait for an AI campus to open and then begin looking for the people capable of operating it. The education programs, apprenticeships, employer partnerships and career pathways must already be producing talent.

The workforce pipeline must come online before the AI campus does.

The Short Answer

A regional talent pipeline is the system that prepares people for careers supporting AI infrastructure and helps them remain in those careers as the industry grows.

It includes universities and community colleges, but it also includes apprenticeships, technical schools, high-school career programs, employer training, military-transition programs and paid work-based learning.

A Strong Pipeline Connects Five Elements:

• Visible employer demand

• Relevant education and credentials

• Work experience and apprenticeships

• Access to jobs through housing and mobility

• Long-term career progression and retention

When those elements operate separately, regions may train people without connecting them to jobs, or attract employers without producing the workers they need.

When they operate together, workforce capacity begins to compound.

That is when talent becomes infrastructure.

The Signal

Data-center operators are already feeling the pressure.

Uptime Institute’s 2025 Global Data Center Survey found that 46% of operators experienced difficulty finding qualified candidates for open positions. Thirty-seven percent reported difficulty retaining staff.

The retention finding is especially interesting.

Uptime found that many departing employees were not leaving the data-center industry. They were being hired by competitors offering better compensation, schedules or advancement opportunities. Uptime Institute

That means the workforce problem is not only a shortage of people entering the industry.

It is also a competition for the people already inside it.

As new facilities open, employers may fill positions by recruiting from neighboring campuses. One company solves its immediate hiring problem while creating another opening down the road.

The workers remain in the regional ecosystem, but the operating pressure moves from building to building.

This is why counting graduates is not enough.

A serious workforce strategy must examine recruitment, preparation, placement and retention as one continuous system.

Who Keeps Intelligence Running?

A data center is a building filled with systems that cannot afford to stop working.

Critical-facilities technicians monitor power and cooling equipment. Electricians maintain switchgear, backup generators, uninterruptible-power systems and electrical distribution. Mechanical technicians service pumps, chillers, cooling towers and liquid-cooling equipment.

Network technicians maintain the connections moving information into, through and out of the facility. Controls specialists manage the sensors and automated systems governing its physical environment. Security professionals protect both the campus and the digital systems operating inside it.

Operations managers coordinate people, maintenance windows, emergency procedures and vendor relationships. Supply-chain teams ensure that critical replacement parts are available before equipment fails.

These occupations do not always appear under one recognizable data-center label.

A worker may enter the industry from a utility, factory, hospital, telecommunications company, military installation or large commercial building. The equipment and operating environment may differ, but many of the underlying skills are transferable.

That creates an opportunity for regions that know how to identify adjacent talent.

The next data-center technician may already be working nearby.

They may simply not know that the career exists.

The Job-Description Problem

The AI industry has a visibility problem.

Most people understand what a software developer does. Far fewer understand what a critical-facilities technician, controls specialist or data-center operations technologist does.

Job titles vary between companies. Similar roles may have different educational requirements. Some employers emphasize degrees, while others prioritize certifications, military experience or demonstrated mechanical and electrical capability.

This makes it difficult for students, displaced workers and career changers to see a clear entrance into the industry.

Economic-development messaging can make the problem worse when every AI-related position is described as a technology job.

The work may support advanced computing, but much of it happens beside electrical panels, cooling equipment, fiber racks and control systems.

These are hybrid careers.

They combine the physical and digital worlds.

Communities Need to Translate Those Careers Into Language People Recognize:

• What does the worker do during a normal shift?

• Which skills are required on the first day?

• Which skills can be learned on the job?

• Which certifications matter to employers?

• What does the starting position lead to?

• Can experience from another industry qualify someone?

• What can the worker expect to earn over time?

A talent pipeline becomes stronger when people can see the path through it.

The Systems Translator

One of the most valuable workers in the Intelligence Economy may be the person who understands how several infrastructure systems affect one another.

This is the Systems Translator.

A systems translator does not need to be the leading expert in electricity, cooling, networking and software simultaneously. The value comes from understanding the relationships between them.

A cooling alarm may begin as a sensor problem. A network interruption may affect the visibility of mechanical equipment. A power-quality issue may appear as a server failure. A maintenance decision in one system may create operational risk somewhere else.

As data centers become denser and more automated, those relationships become more important.

The worker who can recognize them helps technical teams communicate, identify the source of a problem and respond before it spreads.

This changes how workforce programs should be designed.

Electrical students need exposure to controls and networking. Information-technology students need to understand the physical environment supporting the systems they manage. Mechanical technicians need experience with sensors, automation and digital monitoring.

Specialization will remain essential.

But the highest-value talent may increasingly be specialized workers who can also see the system around their specialty.

A Degree Is One Entrance - Not the Entire Door

Universities will remain important to the AI workforce.

They produce engineers, computer scientists, cybersecurity specialists, researchers and managers. They also support innovation, internships and employer partnerships.

But a four-year degree cannot be the only entrance into AI infrastructure.

Many operating roles can be reached through community-college programs, industry certifications, military experience, apprenticeships and paid work-based learning.

AWS, for example, has operated a Data Center Work-Based Learning Program offering 12 weeks of paid on-the-job training. The program highlighted by AWS did not require a college degree or previous data-center experience. AWS

That matters because the scale of the workforce challenge requires more than one pathway.

A technician leaving the military may understand electrical or mechanical systems but need help translating that experience into civilian credentials. A manufacturing worker may already know industrial maintenance and safety procedures. A fiber installer may need additional data-center training rather than an entirely new education.

The goal should not be to lower standards.

It should be to recognize multiple ways of reaching them.

When employers define the competencies they actually need, educators can build shorter and more precise routes toward those jobs.

The Talent Pipeline Has Geography

A worker may be qualified for a position and still be unable to accept it.

The job may be too far from affordable housing. The shift may begin before public transportation operates. Childcare may not be available during overnight or weekend hours. A household may be unwilling to relocate without employment opportunities for a spouse or partner.

These conditions are easy to dismiss as personal concerns.

At scale, they become infrastructure constraints.

Data centers operate continuously. Their workforce must be able to reach the facility during nights, weekends, severe weather and emergencies.

That Means Workforce Readiness Is Connected To:

• Housing availability

• Transportation and commuting times

• Childcare options

• Reliable broadband

• Public services

• Quality of life

• Employment opportunities for entire households

A region can produce qualified graduates and still lose them to another market where it is easier to build a life.

This is why workforce development cannot end at the classroom door.

The talent pipeline must connect education to the practical conditions that allow workers to participate in the regional economy.

Retention Is Part of Capacity

Hiring receives most of the attention.

Retention determines whether the workforce compounds.

When experienced technicians leave, employers lose more than one employee. They lose operating knowledge, mentorship capacity and institutional memory.

New workers need experienced people around them. An expanding industry cannot build a sustainable workforce if every senior technician is being moved between employers to fill the next urgent opening.

The 2025 Uptime Institute survey found that staff hiring and retention remained as challenging as they had been during the previous two years. It also found that workers leaving one operator were often being recruited by another data-center employer. Uptime Institute

That creates a regional retention challenge.

Companies compete individually, but the consequences are shared across the market.

Compensation matters. So do predictable schedules, management quality, advancement opportunities, training and a visible career ladder.

A technician should be able to see how an entry-level position can lead to senior technical work, operations leadership, engineering, commissioning, cybersecurity or another specialty.

If that progression is unclear, the employee may need to change companies to keep moving forward.

A strong talent market does not prevent workers from changing jobs.

It gives them enough opportunity to continue growing without leaving the region.

The Regional Talent Flywheel

Workforce readiness is often treated as a sequence.

A school trains a student. The student applies for a job. The employer hires the student.

In a strong AI infrastructure market, the process operates more like a flywheel.

Employers communicate future demand to educators. Schools create programs around verified skills. Students receive hands-on experience through apprenticeships and internships. Graduates enter local companies. Experienced workers become instructors, mentors and managers.

Employer growth creates more demand, which supports more training, which attracts more workers and employers.

This is the Regional Talent Flywheel:

Employer Demand + Education + Apprenticeships + Worker Mobility + Retention = Workforce Readiness

Each part reinforces the others.

Employer demand makes the career path visible.

Education turns that demand into teachable skills.

Apprenticeships convert classroom knowledge into demonstrated capability.

Worker mobility allows people to reach the opportunity.

Retention preserves experience and creates the mentors needed to train the next group.

When one part is missing, the flywheel slows.

A community college may create a program without enough employers offering internships. A company may announce hundreds of jobs without giving educators time to prepare students. A region may graduate qualified workers who leave because housing is unaffordable.

The goal is not to launch more disconnected workforce initiatives.

It is to build a system in which each initiative strengthens the next one.

Institution Watch: Northern Virginia Community College

Everyone is watching the universities building the next AI model.

I’m watching the community colleges training the people who keep the infrastructure running.

Northern Virginia Community College - better known as NOVA - offers one of the clearest examples of a regional institution responding directly to the data-center economy around it.

NOVA offers an associate degree in engineering technology with a data-center operations specialization. It also offers a one-year, 26-credit Career Studies Certificate designed to prepare students for industry credentials including BICSI copper and fiber installation certifications and OSHA 10. NOVA

The program is focused on the operating environment, including servers, networking, cloud infrastructure and IT systems. Its ET Career Scholars initiative can provide eligible students with a free year at NOVA, a summer bridge program, a one-year data-center operations certificate and a guaranteed internship interview with an industry partner.

The employer connection is what makes the model especially important.

In 2023, AWS contributed $300,000 to support NOVA’s Information and Engineering Technologies Fund as the college expanded its Data Center Operations program to a third campus. NOVA

AWS and NOVA have also collaborated on fiber-optic fusion-splicing training, giving participants hands-on experience and opportunities to meet local employers. AWS

The signal is larger than one program.

Northern Virginia’s data-center cluster created visible employer demand. That demand helped educational institutions build specialized programs. Employer participation brought equipment knowledge, internships and clearer hiring pathways into the system.

This is what a regional talent flywheel looks like when it begins to turn.

The Community College Advantage

Research universities receive much of the attention in conversations about AI leadership.

Community colleges may become just as important to AI infrastructure readiness.

They are positioned close to regional employers. They can serve recent high-school graduates, working adults, veterans and people changing careers. Their programs can often be aligned with certifications and practical technical skills.

They can also respond to the needs of a specific regional economy.

A community near data centers may need critical-facilities technicians and fiber installers. A region supporting new power generation may need electrical and mechanical operators. A manufacturing corridor may require industrial-maintenance and controls specialists.

The most effective institution is not always the one with the most prestigious AI research laboratory.

Sometimes it is the one capable of preparing 40 qualified technicians before a new facility opens.

That is my hot take on the AI workforce conversation:

The community college may be one of the most undervalued assets in the Intelligence Economy.

Automation Will Raise the Skill Floor

Data centers will become more automated.

Software will monitor equipment, identify anomalies, optimize cooling and help predict maintenance requirements. Remote operations may allow smaller teams to oversee more infrastructure.

That does not make workforce planning less important.

It changes the workforce being planned.

Routine monitoring may decline, while demand increases for people capable of interpreting alerts, managing automated systems and responding when the technology encounters something it was not prepared to handle.

The worker becomes less focused on watching one piece of equipment and more focused on understanding the operating environment.

Automation may reduce certain tasks.

It can also raise the minimum level of judgment required from the people who remain.

This is why experienced workers and systems translators will matter. When the automated process fails, the facility still needs someone who understands what the equipment is doing, why it is doing it and what should happen next.

AI infrastructure will not operate without human judgment.

It will concentrate that judgment in fewer, more consequential moments.

Building Careers - not Just Filling Openings

Workforce announcements often focus on a number.

Five hundred jobs. One thousand trainees. Several hundred internships.

Those numbers can be useful, but they do not reveal whether a sustainable talent system is being created.

A stronger assessment asks what happens to people over time.

Do trainees complete the program?

Do they receive interviews?

Are they hired into positions related to the training?

Do they remain employed after one or two years?

Are they promoted?

Do wages increase as skills grow?

Do experienced workers eventually mentor the next class?

These questions separate workforce activity from workforce capacity.

A program should not be judged only by how many people enter it.

It should be judged by whether those people can build durable careers through it.

Workforce Across the Eight Pillars

The talent pipeline touches every pillar of the Infrastructure of Intelligence™.

Power systems require utility engineers, generation operators, electricians and lineworkers. Connectivity requires fiber technicians, network engineers and cybersecurity professionals. Land development requires planners, surveyors, equipment operators and construction managers.

Water and cooling require mechanical technicians, treatment specialists and engineers. Sustainability creates demand for energy managers, environmental professionals and people capable of operating new resource-efficiency systems.

Policy determines which credentials are recognized, how apprenticeships operate and how workforce funding is distributed. Capital determines whether training programs can purchase equipment, hire instructors and expand quickly enough to match employer demand.

Workforce is where the eight pillars become human capability.

A region may possess excellent physical infrastructure.

People determine whether it can be operated, maintained and improved.

What Leaders Should Start Asking

Communities, employers and educators should move beyond asking whether a region has enough workers.

They Should Ask:

• Which operating occupations will AI campuses require?

• How many people will be needed for each shift?

• Which jobs require degrees, certifications or demonstrated technical experience?

• Which skills can transfer from utilities, manufacturing, telecommunications or the military?

• Which regional institutions currently teach those skills?

• Are employers helping design the curriculum?

• Do programs use equipment comparable to what workers will encounter on the job?

• How many students complete the programs?

• How many graduates receive interviews and relevant job offers?

• Are paid internships, apprenticeships and work-based learning available?

• Can workers reach the facilities during nights and weekends?

• Is housing affordable within a realistic commuting distance?

• Are childcare and transportation compatible with 24-hour operations?

• What percentage of experienced employees leave each year?

• Why are they leaving?

• What career paths exist beyond entry-level positions?

• How will automation, liquid cooling and higher-density computing change future skill requirements?

• Who is responsible for coordinating the regional talent pipeline?

These questions reveal whether a workforce strategy is designed to generate announcements or produce careers.

The Golden Nugget

The workforce pipeline must come online before the AI campus does.

A multiyear construction period is also a multiyear training opportunity.

Communities that wait until the facility opens will be forced to recruit from other markets or compete for workers already employed by neighboring operators.

Communities that begin earlier can use the development timeline to align schools, employers, apprenticeships and career pathways with the positions that will eventually become available.

The campus opening should not start the workforce strategy.

It should be the moment the strategy begins producing results.

The IOI Take

The strongest AI regions will not simply attract talent.

They will continuously produce it.

They will connect employer demand to classroom instruction, instruction to work experience and employment to long-term career growth. They will identify transferable skills and create several credible entrances into the industry.

They will also understand that talent retention depends on more than wages. Housing, transportation, schedules, career progression and quality of life all influence whether skilled workers remain in the market.

This is the deeper workforce opportunity of the Intelligence Economy.

AI infrastructure can create more than a temporary construction cycle or a collection of isolated technical jobs. It can become the foundation for durable regional career systems spanning energy, construction, telecommunications, operations and advanced technology.

But that outcome will not happen automatically.

It must be designed.

One Last Thought

An AI campus may be constructed once.

Its workforce must be renewed continuously.

New people must enter. Experienced workers must remain. Skills must evolve as equipment, operating models and computing systems change.

The building may be filled with automation.

The intelligence keeping it operational will still be human.

Workforce

Policy

Capital

Sustainability