We spend most weeks tracking the AI and robotics side of the labor story. This week, we wanted the other half from someone who has lived it.

Tom Noctor, Vice President of Product at Breadcrumb, spent two decades as an electrician and data center builder before moving into software. Here, he explains what the skilled labor shortage looks like on mission-critical builds, and what GCs are actually doing about it.

By Tom Noctor, Vice President - Product, Breadcrumb

I worked through the 2008 global financial crisis. I remember crews being told on a Friday that there was no work for them the following week. It was tough. You had a family to feed and no leverage at all.

Nearly two decades later, on the mission-critical builds we work with now, the tables have turned. Owners have money, and GCs have the contracts. What nobody has enough of is skilled people to do the work.

Here’s what the labor shortage looks like on these builds and what the GCs staying ahead of it are doing differently to hold onto their crews.

Data center construction starts hit $77.7 billion in the US last year, and the sector is projected to double by 2030. That demand has landed on top of a labor market that’s already stretched thin.

I spent two decades in this industry before I worked in software, as an electrician from 2006 and then building data centers for hyperscale data centers in Europe. The US construction industry is short roughly 430,000 skilled workers, and mission-critical builds are feeling it the hardest. These projects run huge numbers of workers and subcontractors at once, most of it specialized electrical work that can't be handed to anyone without the right certifications. You can't out-hire this. Retirements are outpacing new entrants, and the people with the right skills go to the highest bidder first.

On top of that, the hyperscalers and Fortune 500 companies funding these builds expect a level of documentation most commercial projects have never needed before. Digital proof of workforce compliance is now a condition of even being allowed to bid, not something you produce after the fact if an inspector asks. That's more paperwork landing on the same stretched workforce, eating into the time of exactly the people who are already too scarce to spare.

Faced with all of that, GCs often reach for the same fix first: pay more.

Mission-critical projects can afford to compete on salary, and for a while that’s been the whole strategy. I've watched GCs stack a bonus on top of pay, offer private healthcare and a wellness stipend, adding another layer every time a competitor matches the last one.

But that race has a ceiling. Margins on these projects are tight, and there's only so much a GC can keep adding before it stops making financial sense. Pay isn't the only thing candidates are weighing either. This generation of builders expects to use technology day to day, not file paperwork. Once every GC on a shortlist can offer a comparable package, what tips the decision increasingly comes down to what the job actually looks like day to day, not just what it pays.

Time is the lever most GCs haven't pulled yet. Think about a superintendent who loses three hours every morning checking workers in on paper and chasing down signatures. They're not managing the build during that time — they're stuck in the job trailer doing paperwork. Digitize that process, you cut 25-50% of admin load, and it frees them up to get back in the field and do what they actually signed up for.

It's not just one person, either. If you reduce the amount of time it takes for builders to capture safety and compliance information in the field, you’ll get them on the tools faster. Do that across a whole workforce, and the productivity compounds fast. 

Top 100 ENR contractor Brinkmann Constructors is a good example of a GC doing it right. They moved their safety documentation and on-site orientations to a digital process with Breadcrumb. Last quarter alone, over 7,000 of Brinkmann’s subcontractors and visitors checked into the jobsites from their mobile devices, and their site teams could see in real time who was on site and who was compliant. With all the data flowing into one place, it keeps the project on schedule.

The way people spend their time at work isn't the only thing that decides whether someone stays. I don't say this lightly: you cannot call a project successful if someone gets injured or killed on it. Nobody wants to work for, or with, an unsafe company. I've seen serious incidents on site over the years, and I've walked off jobs myself rather than do something I judged unsafe. 

It's also a real business risk for the people funding these projects. The owners commissioning data centers are large public companies, and they don't want to be associated with an unsafe jobsite. It affects their reputation and, ultimately, their share price. That's why more of them are asking for digital proof of workforce compliance before a GC is even allowed to bid, not just after something's gone wrong. If a GC can't demonstrate it, they're not in the conversation. Right now in the US, it's private clients setting that bar, not regulation, and the GCs who don't get ahead of it are going to fall behind and lose the trust of workers who have options.

None of this is unique to data centers in principle. Every corner of construction is short on skilled labor. But mission-critical work is where the pressure shows up first, because the projects are bigger, the risk is higher, and the clients funding them demand more and more. What owners are starting to require on data centers is already showing up in healthcare and commercial bids too.

The GCs treating retention as a productivity and safety problem, are the ones keeping their crews around for the next project. Pay is close to maxed out already. Time and safety are the levers still worth pulling.

The labor squeeze Tom describes goes beyond safety documentation. We've seen the same dynamic play out in construction scheduling, where mega-projects are turning to agentic AI for the same reason: there simply aren't enough qualified people to go around, and software is stepping in to close the gap.

POWERED BY: