Your essential services are one surprise failure away from disruption. Consider how physical AI could tackle the crisis

It’s a tense time for workers—and really for anyone who uses infrastructure. If the past few weeks of winter-grid strain taught us anything, it’s that essential systems have less room for failure than most people realize. And we’re still maintaining them through reactive, manual methods, dispatching crews after something breaks instead of preventing failure in the first place.

For all the energy being spent on whether AI will eliminate white-collar jobs, I believe that business leaders are missing a much bigger story: AI won’t replace skilled trades–it will require more of them, and most importantly, make them better.

That’s not a nice-to-have. It’s a necessity as the trades workforce ages, retirements accelerate, and fewer new workers enter the pipeline. I’m a founder with decades of experience in what’s increasingly being called “physical AI,” and I strongly believe that AI should be used to support, not replace, workers. As labor constraints tighten and the cost of downtime climbs, the fastest path to resilience isn’t automating people, it’s up-leveling them: giving frontline teams continuous visibility and decision support where work actually happens.

Across boardrooms and earnings calls, leaders are fixated on which white-collar roles will disappear first, even as they race to deploy tools that promise to write faster, analyze quicker, and compress desk work even further. Nearly every AI debate centers on productivity gains for desk-bound knowledge workers, yet this conversation overlooks where the real workforce pressure lies and where AI can deliver its most meaningful economic impact.

Here’s the reality: the AI boom is physical.

Civilization doesn’t run on build-outs alone. It runs on maintenance: keeping power on, water flowing, equipment safe, and facilities operating every day. Data centers, power grids, battery plants, EV infrastructure, water systems, and industrial facilities all depend on the constant work of electricians, utility crews, HVAC technicians, construction laborers, and field service workers.

The big problem is the deterioration of the physical systems that keep society running. The big opportunity is applying Physical AI to prevent and reverse it. More than 45% of U.S. infrastructure is rated in poor or mediocre condition and somehow, while much of the economy has been digitized, we largely maintain our infrastructure with methods from the Roman era. We dispatch field crews “blindly,” sending them into the field to conduct manual, reactive and infrequent inspections, resulting in either a waste of time with no problem being found, or a missing major issues that go undetected for years. This leads to costly failures that put even more pressure on an already strained workforce, often requiring repeat visits with different tools or parts. In many industries, more than half of truck rolls are unnecessary. This would all be preventable in a world of physical AI.

In other words, AI isn’t just a software story. It’s a labor story—and one that depends heavily on skilled trades. This is the workforce that can gain the greatest productivity benefits from AI.

The real labor crisis

America already faces a critical shortage of skilled tradespeople. A large share of the workforce is nearing retirement: roughly 40% of skilled trades workers are 45+ (and nearly half of that group is 55+). Meanwhile, the entry pipeline is thin, with less than 9% of the workforce in the 19–24 range. In plain terms, more people are aging out of these jobs than entering behind them—and decades of hard-won field knowledge are walking out the door with them.

This isn’t just inefficient. It’s dangerous, demoralizing, and unsustainable. These roles require navigating hazardous, physically taxing conditions–from extreme temperatures to confined spaces. Because the tasks are highly repetitive and frequently hampered by inefficient processes, the work can feel unfulfilling. It’s not surprising that burnout is common, and young people aren’t jumping to take on these roles.

Physical AI is the solution to these problems. Real-time understanding of the state of infrastructure eliminates the cost and frustration of unnecessary truck rolls and gives predictive insight to help field teams solve real problems before critical failures. Guidance and contextual awareness of worksites help new job entrants to come up to speed faster and more safely, bridging the labor gap. Most importantly, for all of us, we can create many new jobs while ensuring that the essential services we all rely on are delivered to the quality the world needs.

The opinions expressed in Fortune.com commentary pieces are solely the views of their authors and do not necessarily reflect the opinions and beliefs of Fortune.

This story was originally featured on Fortune.com