AI-native factories for America’s next industrial era
For the last half century, the logic of manufacturing was simple: find cheaper labor, build larger factories, stretch supply chains across the world, and optimize for scale. That logic worked incredibly well for a long time. It also hollowed out large parts of the American industrial base.
Now we need a different model. The next generation of critical technology, from drones and satellites to medical devices, robotics, energy systems, and AI infrastructure, will require a manufacturing base that is faster, more flexible, and closer to home.
The United States is making two massive bets at once. One is on AI. The other is on reindustrialization. They are often treated as separate movements, but they are becoming deeply connected. AI is not just a software story. It may be the tool that lets the U.S. rebuild manufacturing capacity without trying to copy the exact system China spent decades perfecting.
China built its manufacturing advantage through labor density, supplier concentration, state support, and relentless execution. The U.S. is unlikely to win by replaying that game. Our better path is to build around the advantages we already have: software, AI, capital, engineering talent, technical founders, and urgent domestic demand.
That is why we invested in Foundational Industries.
Foundational is building AI-native factories for the next era of American manufacturing. The company is starting with a simple but powerful premise: the factory of the future will not be a traditional production line with some AI bolted onto the side. It will be designed from the beginning as a software-driven, robotic, reconfigurable system.
Most factories are built for repetition. A product is designed, a line is configured, equipment is installed, processes are documented, workers are trained, and the whole system is optimized around producing the same thing again and again. That works beautifully when demand is massive and stable. It works less well when products are changing quickly, volumes are meaningful but not enormous, and customers need to move from prototype to production without waiting years or moving offshore.
That middle ground is where many of the most important new industrial companies get stuck. They can design the product. They can raise the capital. They can find the customer. But when it is time to manufacture hundreds, thousands, or tens of thousands of units, the options are often bad. Small shops can handle prototypes. Large contract manufacturers want bigger, more predictable programs. Building your own factory too early is expensive and distracting. Going overseas may create supply chain, security, and speed problems.
Foundational is attacking that gap.
The company is building a manufacturing platform based on reconfigurable work cells, robotic execution, software orchestration, and data-driven factory operations. Instead of treating every product as a one-off manufacturing challenge, Foundational is building a system that can learn across products, processes, suppliers, and facilities.
That is the part that makes the company interesting. A modern factory should not just produce parts. It should generate data. It should understand what worked, what failed, what took too long, which suppliers performed, which steps created defects, and which processes can be reused the next time. Over time, the factory should become faster and more capable because the system has seen more products and solved more problems.
This is the manufacturing version of compounding.
The first use case is AI infrastructure. That is a smart place to start. The demand for data center equipment is exploding, and every new generation of chips changes the physical requirements around power, cooling, enclosures, cabling, and deployment. The world talks constantly about GPUs, models, and compute clusters. But all of that intelligence still depends on very physical systems: racks, sheet metal, thermal management, power distribution, wiring, testing, and installation.
Those bottlenecks are already real. If the AI buildout is one of the largest infrastructure projects of our lifetime, then manufacturing capacity for that buildout becomes strategic. Foundational can begin with a market where the pain is immediate, the demand is large, and customers are already supply-constrained.
From there, the same capabilities can expand into other categories that need domestic, high-complexity manufacturing: defense systems, drones, robotics, aerospace, medical devices, edge compute, and industrial equipment. These are not markets where speed and supply chain control are nice to have. They are often the difference between winning and losing.
The founding team is a major part of the bet. Jonathan Winer has spent his career building technology-enabled infrastructure businesses. His experience exposed a gap that many founders in hard tech eventually run into: America can have the capital, the customer demand, and the strategic urgency, yet still lack the manufacturing capacity to get the product built domestically.
That is a brutal constraint. It also creates the opportunity.
Foundational has assembled a team that combines software, robotics, physical infrastructure, and large-scale manufacturing experience. A pure software team would likely underestimate the operational complexity. A pure manufacturing team might not rethink the architecture deeply enough. Foundational needs both: people who know how factories actually work, and people who believe factories can work differently.
This will not be easy. Manufacturing punishes abstraction. Real-world production comes with scrap, delays, quality issues, supply chain surprises, underutilized equipment, labor constraints, customer changes, and margin pressure. Foundational will need to prove that its software and robotics create a genuine operating advantage, not just a better story around a hard services business.
If Foundational works, it becomes more than a contract manufacturer. It becomes part of the new industrial base: a network of flexible, AI-driven factories that help the next generation of American companies move from prototype to production without leaving the country.
That matters because the future is becoming physical again. AI may be digital, but the world it enables will be built in atoms: data centers, robots, sensors, satellites, drones, energy systems, medical devices, factories, and machines we have not yet imagined. The countries and companies that can turn intelligence into physical output will have a durable advantage.
For years, software ate the world. Now software has to help build it.
That is the promise of Foundational.
Visit Foundational Industries at https://foundational.industries.