Data CenterEmployerEnergy & Infrastructure NewsIndustry InsightsUS

AI Data Centers Are Changing the Build. Is Your Team Ready?

By November 18, 2026No Comments5 min read

AI data center construction is asking far more of engineering and construction teams than the facilities most of them learned on. High-density computing, liquid cooling, and much heavier electrical loads all have to work together, and the people who have delivered that combination are still rare.

Not long ago, building a data center meant putting racks of servers into a well-cooled room and making sure the power was reliable. The rest followed a pattern the industry had refined over decades. The facilities being built for AI today are a different proposition, and that is changing what hiring leaders need to look for.

What makes AI data center construction different

The scale of the shift is easier to see in numbers. Uptime Institute’s 2026 survey of more than 800 operators put the most common rack density at around 11 kW, while NVIDIA’s GB300 NVL72 rack-scale systems reach up to 142 kW per rack, as Capacity reported in September. At those densities, air cooling stops being a realistic option. Liquid cooling moves from a specialist add-on to the core of the design.

That brings new systems into the building. Coolant distribution units, manifolds, leak detection, and the pipework that ties them together all have to be designed, installed, and commissioned. They sit alongside electrical systems carrying far more power per square foot than most teams have worked with before.

The cooling also has to be closed loop. Communities are wary of a large facility drawing on the local water supply, and developers need to show they aren’t putting extra pressure on the area.

All of this sits on top of the power challenge I wrote about in Time to Power, where securing enough electricity is already shaping which projects get built.

What it asks of engineering and construction teams

In practice, the AI build pulls on two groups at once. Design engineers have to work out how cooling, electrical systems, and the wider building fit together. Construction teams then have to deliver those increasingly complex solutions on site, often on aggressive schedules.

The disciplines involved are the familiar ones: MEP (mechanical, electrical, and plumbing) and CSA (civil, structural, and architectural). The coordination between them is far tighter than on a conventional build. A decision on the cooling design can change the electrical layout, which in turn affects the structure, so a team where each discipline works in isolation will struggle.

The full skill set is rare, and that’s normal

It’s tempting to start these searches looking for someone who has delivered a high-density, liquid-cooled facility from start to finish across every discipline. Those people exist, but there are few of them, because the AI-focused facilities we’re seeing now have only been around for a few years. There hasn’t been time for a large pool of engineers to take one from zero to one across the full specification.

What I see in the people who actually get hired is different. They are usually strong in one discipline, and they have picked up genuine exposure to the others along the way. A mechanical engineer who has also done electrical work on the side, even though it was never their main responsibility, is often more valuable on an AI build than someone with a longer CV in a single discipline. That kind of cross-discipline exposure is a niche profile, which is why it’s worth knowing how to spot it.

The practical takeaway is to stop hiring for a profile that barely exists and start building a team. Decide which discipline is essential for the role, look for depth there, and then look for the exposure that will help that person coordinate with everyone else on the project. 

Six questions to separate conventional experience from AI data center experience

If you’re interviewing for an AI-focused build, these questions help separate the two:

  1. What was the rack density on the last facility you delivered, and was any of it liquid cooled?
  2. Which cooling approach did the project use, and what exactly was your role in it?
  3. How did the mechanical and electrical teams coordinate, and where did you sit in that process?
  4. Outside your core discipline, what have you worked on hands-on?
  5. How was water use handled on the project?
  6. What went wrong during commissioning, and how was it resolved?

Listen for specifics, such as a density figure, a named cooling system, or a commissioning problem the candidate personally helped fix. Vague answers usually point to experience on conventional facilities, which is still valuable, just not the same thing.

Talk to Storm4

Our data center team works with operators, developers, design partners, and cooling businesses to find engineering and construction leaders who have worked on high-density builds, through permanent staffing and executive search. Speak to Oli and the Storm4 data center team.

We’ve helped some of the most successful startups in Energy and Infrastructure grow.

— now it’s your turn.

Share