Foundry Robotics announced on 28 October 2026 an oversubscribed $100 million Series A round to build what it calls America’s first autonomous factory for assembly, a plant in Emeryville, California, where robotic cells, not fixed production lines, assemble complex hardware. The round was led by Marlinspike Partners, with early backers Khosla Ventures and Red Glass Ventures and new investors SE Ventures, Wing VC and Axiom Ventures, alongside existing investors Hanabi and Zero Shot Fund. Foundry has expanded its manufacturing footprint twentyfold to a 130,000-square-foot facility.
Owning The Factory, Not Selling Robots
Foundry’s model differs from most robotics companies in a simple way: it owns and operates the factory, offering manufacturing as a service, rather than selling robots or software to manufacturers. Customers go from design to production in weeks, the company says, without new production line investments or expensive retooling, because Foundry’s robotic cells can be reconfigured for a different product instead of being rebuilt for one.
Traditional assembly lines, Foundry argues, are built around a single product, take months to deploy and are costly to reconfigure. That rigidity is the bottleneck the company is attacking. Defense and industrial companies can design new hardware faster than ever, the company says, but manufacturing remains the constraint, and flexible assembly capacity is what lets a design become a finished product at speed.
How The Factory Works
Foundry trains AI models on the building blocks of assembly and deploys them in a grid of robotic cells with manufacturing-grade precision. FoundryOS, its agentic system, links the robots, the factory floor and the cloud, so every production run feeds back into the models.
The same cell that assembles a drone today can, in principle, learn to build a satellite component tomorrow, and get better at both with every run. The applications span defense and commercial work: drones, autonomous vehicle subassemblies, wire harnesses, underwater vehicles, telecom and data centre infrastructure, ground robots, rugged computers, robot compute boxes and satellites.
That breadth is the point of the model. A factory that can switch products without retooling is valuable precisely because no single product justifies a dedicated line, and the list of products Foundry serves is deliberately wide.
Who Is Behind It
Adarsh Kulkarni co-founded Foundry with Emmanuel Lamacchia, the former programme manager for the Tesla Model Y programme. Kulkarni was employee number four at Ghost Robotics, where he built legged robots for the US military, and later became Scale AI’s inaugural head of robotics.
Lamacchia, the company says, was the single-threaded owner for the Model Y, then the world’s best-selling car, responsible for manufacturing about 1.5 million cars a year. The leadership team also includes a former Tesla software leader and veterans who worked at Raytheon and Northrop Grumman.
Duane Fotheringham, president of HII Unmanned Systems, said production capacity is a national constraint on getting unmanned systems to the fleet, and that Foundry’s approach is exactly the kind of capability the defense industrial base needs. That endorsement, from a customer-side voice in defense, is part of the company’s pitch that this is industrial policy as much as a startup story.
Physical AI, And A Funding Wave
Foundry sits inside a wider physical AI wave, in which investors are backing companies that put AI into the physical world rather than only into software. Physical Intelligence, for example, has been in talks around a large valuation for its robot foundation models, as covered in Physical Intelligence talks.
Foundry’s bet is different: not a better robot brain alone, but a factory designed around robots from the start. The same week also brought another $100 million round, for data platform Hex and its agent-native rewrite, as covered in Hex’s $100M raise.
The pairing is instructive. One company is rebuilding software for agents, the other is rebuilding factories for robots, and both argue that bolting AI onto an old architecture, whether a dashboard or an assembly line, has reached its limits.
Why It Matters
If Foundry’s model works, it changes who can manufacture complex hardware. Small companies and defense programmes that cannot justify a dedicated production line could buy assembly capacity the way they buy cloud computing, by the job rather than by the factory.
That would lower the barrier to building drones, satellites and other hardware, and speed up the cycle from design to deployment. The risks are equally concrete. A factory that assembles many different products must prove quality, security and reliability across all of them, and a single shared facility becomes a single point of failure for its customers.
Foundry has not disclosed a valuation, and its “first autonomous factory” claim is the company’s own description, not an independent ranking. The Emeryville plant, now being built out with the new funding, is where those claims will be tested.
Background
Foundry Robotics was founded in 2026 and is based in Emeryville, California. It builds and operates autonomous factories for assembly, offering manufacturing as a service for defense and commercial industries.
Source: Foundry announcement
0 Comments