Most of the conversation about reindustrialization is at too high a level of abstraction. People talk about reshoring, supply chains, critical materials, strategic autonomy. These are real things, but the abstraction makes it hard to see what reindustrialization actually looks like as work. So I want to talk about the concrete version of it, which is what's involved in making a hardware company work in 2026 and why we ended up doing it in the Midwest.

We're headquartered in Chicago. We manufacture our membrane at Kodak Park in Rochester, New York. The pilot facility is in the upper Midwest. The supply chain for our raw materials runs through Texas and Louisiana. Our customers are in Detroit, Cleveland, Akron, Pittsburgh, Indianapolis, Huntsville, Houston, and a few coastal cities. None of this is by accident. It's roughly the geography you'd end up with if you asked yourself where in the country it's still possible to actually make industrial product at scale, and the answer is mostly the middle of the country.

The thing California is genuinely good at is software. It has been good at software for so long that the rest of the country forgot that California used to do hardware too. The reason California stopped doing hardware wasn't that the climate changed. It's that the talent and capital concentrated on software companies because software companies had the best returns for a generation. The cost of running a deep tech company in San Francisco is roughly twice what it costs in Chicago, and that's before you count the talent premium. The supply chain you need — fabricators, machine shops, materials suppliers, contract manufacturers — is two thousand miles away. The customers you sell to, which are industrial customers with industrial deal cycles, are in the middle of the country. The investors are local; nothing else is.

For a company that ships physical product, this is the wrong arrangement.

The Midwest has the opposite arrangement. The infrastructure that supported the American industrial era — the auto plants, the chemical plants, the steel mills, the heavy equipment makers — didn't disappear. It contracted. A lot of capacity got repurposed for niche manufacturing rather than commodity production, which means that within a day's drive of Chicago you can find roll-to-roll coating lines, precision converting, contract chemical synthesis, large-scale extrusion, machine shops that hold tolerances that are essentially impossible to find on the coasts, EPC firms that have built cleanrooms, and a labor market full of people who know how to run a factory because they've run a factory. The skill base didn't leave. It got quieter.

The Kodak Park facility where we coat our membrane is an example of what I mean. Kodak used to coat film for the entire world. The same coating lines that produced Kodachrome — color film consistent to roughly one part in ten thousand across acres of substrate — are now available for use by companies that need to coat thin polymer films at industrial scale. We didn't have to build a coating line. We had to find one and use it. The capital we would have spent building the line went into the chemistry instead. The yield curve had already been climbed by Kodak, thirty years ago, on the same equipment we're using now.

This is the part of the reindustrialization story that I think gets under told. The American industrial base didn't vanish. It got quieter, and it got more dispersed, and a lot of the surface area is now waiting for somebody to need it again. Finding that surface area and putting it back to work is what a founder building physical product can actually do — and it's much cheaper, both in time and in capital, than starting from scratch.

The other thing the Midwest has, which I think is the most underrate industrial asset in the country, is the national labs. We were spun out of Argonne, which sits forty-five minutes west of downtown Chicago. The national labs are not evenly distributed across the country. A disproportionate share of them — Argonne, Fermilab, Ames, Oak Ridge, Idaho National Lab — are in the middle of the country, not on the coasts. The capability that lives inside these facilities, both in instruments and in expertise, is on a different scale from what's available at any university. When we needed to characterize them olecular structure of a new polymer at very high resolution, we drove a few miles to Argonne and used the most advanced X-ray scattering facility on the continent. The Bay Area has something similar in Berkeley Lab and SLAC. The difference is access. There are far fewer startups in the Midwest competing for the same deep tech resources than there are in the Bay Area. Maybe this changes after people read this essay.

So the case for building here, in practical terms, is roughly: the supply chain we need is here, the manufacturing partners we need are here, the customers we sell to are here or one flight away, the national labs we depend on are here, the talent pool that wants to work on physical product is here, and the cost of running the company is half what it would be on the coast. None of these advantages are romantic. They're logistical.

The romantic version of the argument, which I'll make briefly because I think it's also true: the era we're entering is an industrial era. The companies that matter for the next twenty years are going to build physical things at scale in the United States. They're going to be capital-intensive. They're going to have long deal cycles. They're going to need access to factories, to engineers who have spent careers in factories, to industrial customers, and to a culture that respects the work of making things. That culture exists, in a way it doesn't really exist anywhere else, in the middle of the country. The Midwest never stopped being the place where America builds things. The bones are still good. We're here to rebuild the muscle.

And as a small bonus, the way you compose a manufacturing strategy in 2026 doesn't require building everything from scratch, which means a founder can get further faster on less capital than the previous generation could. We did. The model works.

The jobs that come out of this aren't theoretical. Every megawatt of membrane production capacity employs operators, technicians, chemists, engineers, salespeople, and the surrounding ecosystem of suppliers and customers it pulls in. A domestic substrate industry compounds. An imported one doesn't. The way to have the kind of skilled industrial jobs the country has spent thirty years saying it wishes it still had is to build the substrate of the next industrial era at home, and that's what we're doing. The whole argument, when I look at it from the ground up, comes down to this: it's possible to do the work here. So we are.

‍