The Startups Rebuilding American Industry | The AD50

Decades of industrial hollowing-out have weakened America’s ability to supply battlefields, with consequences for U.S. and allied forces. The AD50 companies are rebuilding capabilities across the industrial chain, from raw materials and factories to finished systems. Speakers say the effort is gaining urgency as global conflicts intensify and entrepreneurs and investors turn toward physical industries. Defense demand can help establish production capacity and know-how that later support commercial markets, but startups risk losing focus if they pursue too many markets before mastering one difficult product.

The speakers describe government funding as a way to cover upfront costs that commercial buyers may not accept, helping new production capacity become viable. A defense buyer, for instance, could help pay for equipment at a U.S. printed-circuit-board factory whose output could also serve robotics and hearing aids. They point to DJI’s movement from consumer drones into defense production and robotics as evidence that capabilities can transfer: drone motors and robot actuators are similar, and manufacturing knowledge can serve both markets. But the first bespoke version of a product is often expensive, and commercial markets ultimately press for greater capability at lower cost.

Rebuilding industry also means improving many connected steps, not just finding a single software breakthrough. At Hadrian, work runs from quoting and order intake through production, inspection, and transfer; Mariana Minerals applies software to capture data across its asset development, mining, refining, and sales operations and feed it back into the process. One speaker says a manufacturing engineer using Claude can retool a line in about 20 minutes, versus the traditional three months and hundreds of thousands of dollars. More adjustable lines could make high-mix, low-volume production more economical, speed prototyping, and support scale-up. The speakers stress that mine or factory improvement usually takes 15 or 20 changes, and that designing for manufacturing can prevent costly subsystem redesigns after production starts.

A stronger domestic supplier base could make U.S. robotics more viable and reduce reliance on Chinese factories. One speaker estimates a U.S.-based robotics company has a 99% chance that most of its purchases come from China, while arguing that quality, cost, and faster iteration—not only geopolitics—can motivate onshoring. Defense programs also create demand for components such as actuators, motors, valves, and switches. New U.S. suppliers are already improving: one portfolio company once expected half its circuit boards to fail quality tests, while recent batches from new domestic suppliers have had 70–90% pass. Speakers see an opening for manufacturers to co-engineer components with customers rather than merely build to specifications, particularly as new aircraft, military systems, and vehicle designs demand different parts.

The speakers say private capital can let startups develop systems before a government request arrives, addressing procurement’s separation of buyer, end user, and appropriators. Castilian used venture funding to develop a reliable hypersonic missile before receiving government contracts and sales; its team spent two and a half to three years on that first product instead of branching into a small drone. They expect newer defense companies and “neoprimes” to gain share through this approach, while cautioning that complex physical products require sustained focus. Delivering one hard product can build manufacturing expertise, supplier relationships, process knowledge, and customer trust that help with later products. SpaceX and Tesla are cited as examples of vertical integration when U.S. suppliers could not reliably meet their needs, though integration is less necessary when good suppliers exist.

The companies are building large facilities in places including Ohio and Baltimore, with products already coming off production lines and potential output ranging from thousands to hundreds of thousands of systems. Speakers say these buildouts can quickly fill warehouses with machines and workers, while increasing demand for power. They describe capital markets as increasingly willing to fund factories, defense, energy, robotics, and critical minerals: one speaker said that in the first half of 2026, more risk capital entered the defense ecosystem than in all of the prior year, whose total was itself many times the year before. This funding may narrow the gap between building capacity and winning contracts, but government choices, power constraints, and investor patience remain risks. Aaron Connor says nearby residents should have a say in infrastructure projects and share in their economic upside; he points to AMCA’s work with small machine shops to grow their businesses. He expects AD companies to bring jobs, energy, and security to Middle America, while acknowledging that trust and participation take time to build.