I’ve spent a good chunk of this year standing in plants that didn’t exist two years ago. New concrete, new lines, fresh paint on the guarding. Sometimes the plant manager walks you through with a look on his face like he still can’t quite believe he gets to run the place.
I love those walkthroughs. I joined Tractian six months ago, and seeing our new Global Headquarters in Atlanta has given me a firsthand look at the energy that comes with opening a new facility.
Opening a plant and getting it running reliably are two different jobs, though, and the second one is what I've been thinking about most this year. The investment decision gets the headlines. The hiring, training, and maintenance work takes years, much of it after everyone has gone home from the ribbon-cutting.
That second job is what I want to spend a few minutes on.
The announcements are only the beginning
There's no shortage of intent on the record. IoT Analytics counted 227 public industrial companies announcing changes to their U.S. footprint in the second quarter of 2025 alone, 33 of which were pulling production out of China. RTX by itself committed $2 billion to expanding domestic capacity.
The concrete has been slower to follow. That same analysis tracked U.S. manufacturing construction spending down roughly 21% from its June 2024 peak of $239 billion through March 2026, and if you strip out the electronics sector, everything else grew about 2% once you account for inflation. Manufacturing employment has slipped about a percent since the April 2025 tariffs landed.
Some of that gap is capital and some of it is permitting, and a good share of it is reasonable caution. It’s hard to bet a nine-figure line on rules that may not sit still. Underneath all of it sits a constraint that money moves very slowly, and it's the one I'd lose sleep over if I were running a plant this year.
You can build a plant in 18 months. You can't hire 30 years of experience.
Here's the math I keep coming back to. The Manufacturing Institute and Deloitte estimate manufacturers may need as many as 3.8 million additional workers between 2024 and 2033, and that around half of the skilled openings (roughly 1.9 million jobs) could go unfilled.
Numbers that big tend to go numb, so let me put a face on one of them.
Every plant I've ever walked has a guy. He's been there 28 years, and he can tell you which pump is about to go by the sound it makes when the line starts up cold on a Monday. He knows the compressor that trips when it rains, the conveyor that needs a shim nobody ever documented, the exact spot on the floor where you can hear the gearbox before anyone else can. None of it is written down anywhere, because it lives in his hands and his ears. And, he’s retiring (maybe this year, maybe next).
The kid replacing him is sharp and motivated and worth every dollar. He's 6 months in, and 6 months doesn't buy you the compressor that trips when it rains.
That's the part of the reshoring story that’s hard to fit into an announcement. We're standing up new capacity at the same moment the deepest institutional knowledge in the building is heading out the door to go fishing. We're asking crews, that were already stretched thin, to cover more square footage, more assets, and more brand-new equipment they've never met.
What that looks like at 2 a.m.
You probably know what it looks like already, because most plants are living some version of it. A critical asset goes down on second shift and the person who would have caught it three weeks early is gone. The call goes out, somebody's phone lights up at home, and a maintenance lead who was supposed to be at his daughter's game on Saturday spends it standing in front of a failed bearing hunting for a part number instead. Everybody works harder, and the plant still loses the hours.
That's the part that stays with me. The same failure shows up in the KPI deck as unplanned downtime and in somebody's actual life as a missed birthday. Both of those are real costs. Reactive maintenance charges you twice.
Bringing that line back to Ohio doesn't change the challenges of the past. A domestic plant inherits the same aging assets, the same thin crew, the same blind spots, the same 2 a.m. phone calls. Geography gets us the opportunity. Reliability is what turns it into results.
Giving a small crew better information
This is the part where I’m optimistic about, and a good share of why I do this work.
We can't manufacture 30 years of experience, but we can stop asking one person's instinct to be the early warning system for a 400,000-square-foot facility. That's the whole idea behind what we build at Tractian: put sensors on the assets that matter. Watch vibration, temperature, and energy continuously, and let the machines tell the whole team what they used to only tell the guy who'd been there since 1997. The failure signature that lived in his gut becomes a pattern anybody on the crew can read on their phone, three weeks before the thing breaks.
What that does for a stretched crew is worth spelling out. A newer tech walks up to the right machine at the right time, already knowing what to look for and which part to bring. The 28-year veteran gets to spend his last two years teaching instead of firefighting. And the lead makes the ball game, because the bearing got replaced on a Wednesday afternoon on purpose.
Whirlpool is a good example here, and a company that knows plenty about building appliances in this country. Their teams have avoided more than $1 million in costs by acting on early warnings from our platform and intervening before those assets went down. That's a seven-figure number made up entirely of failures that never got the chance to happen, and production hours nobody had to explain in a Monday morning meeting.
Onshoring gives American manufacturing the opportunity, and reliability is what lets us hold onto it. Capacity on the ground only counts when it's running, and running is a people problem long before it's a technology problem. Technology just happens to be the fastest way we know to make a good, thin, green crew genuinely capable.
What I'd plan for
If you're standing up new domestic capacity right now, I hope it goes incredibly well. I sincerely mean that. We need it to work for reasons that go well past anybody's balance sheet, because there are towns riding on these decisions.
The plan deserves the same seriousness you brought to site selection and capex, built around who's running it and what the assets are actually doing. Three things I'd want answered first: who owns it, what the month after that person leaves looks like, and how many of our critical assets we can see today.
Changing the world, one asset at a time...
If you want to talk through what asset visibility could look like in your plant, my inbox is open. I'd enjoy the conversation.