Foundry Robotics said in a Business Wire release on 28 October 2026 that it has raised an oversubscribed $100 million Series A led by Marlinspike Partners, and that it has expanded its manufacturing footprint twentyfold to a 130,000-square-foot facility in Emeryville, California.
By our arithmetic, a twentyfold expansion to 130,000 square feet implies an earlier footprint of about 6,500 square feet, and the $100 million round works out to about $769 for each square foot of the new facility.
Short explainer · Foundry Robotics
A vertical news Short narrated from the published headline and opening lines of this piece.
Business Pill · Physical AI
An evergreen Business Engineer explainer on physical AI: AI systems that perceive and act in the physical world. It explains the general idea only and does not discuss Foundry Robotics or this round.
The key insight: By our arithmetic, Foundry Robotics’ twentyfold expansion to 130,000 square feet implies an earlier footprint of about 6,500 square feet, and its $100 million Series A equals about $769 for each square foot of the new Emeryville facility.
What the Company Announced
The release says Khosla Ventures and Red Glass Ventures, described as early backers, took part, along with new investors SE Ventures, Wing VC and Axiom Ventures and existing investors Hanabi and Zero Shot Fund.
Foundry describes itself as building and operating “autonomous factories for complex assembly”, combining AI-powered robotic cells with experienced manufacturing experts. The release says the new money will build out the Emeryville facility, accelerate deployment of AI models into real-world manufacturing, and grow its robotics research, engineering and factory operations teams.
The company says it serves defense and commercial industries. The applications the release lists include drones, autonomous vehicle subassemblies, wire harnesses, underwater vehicles, telecom and data center infrastructure, ground robots, rugged computers, robot compute boxes and satellites.

What Was Posted on X
Michelle Sun maps five layers of manufacturing, from materials and standard items to finished products, and gives a battery pack or wire harness as examples of subassemblies.
How the Model Works
The release describes a manufacturing-as-a-service model in which Foundry owns and operates the factory, “taking customers from design to production in weeks without new production line investments or expensive retooling”.
It says Foundry trains AI models on the building blocks of assembly and deploys them in a grid of robotic cells, with FoundryOS described as “a single agentic system linking the robots, the factory floor and the cloud”. The release adds: “Every production run feeds back into the models”.
Founder and chief executive Adarsh Kulkarni said in the release: “Our approach is to own the entire final assembly stack and integrate intelligence at every layer”.
Who Is Behind It
The release says Kulkarni co-founded Foundry with Emmanuel Lamacchia, described as the former Program Manager for the Tesla Model Y program. It says Kulkarni was employee number 4 at Ghost Robotics and later Scale AI’s inaugural Head of Robotics.
Duane Fotheringham, President of HII Unmanned Systems, is quoted in the release: “Production capacity is a national constraint on getting unmanned systems to the fleet”. Mislav Tolusic, Managing Partner at Marlinspike, said: “Defense and industrial equipment manufacturers urgently need more domestic assembly capacity, and Foundry is delivering”.
A Five-Layer Map of Manufacturing AI
A day after the announcement, Michelle Sun, whose X profile describes her focus as the physical AI value chain, robotics and industrial automation, posted a map of five layers of manufacturing: materials and standard items, parts, components, subassemblies and finished products. She wrote: “The next wave of generational companies will come from manufacturing AI”.
Her post gives “a battery pack or wire harness” as examples of subassemblies. Wire harnesses and autonomous vehicle subassemblies are also on the list of applications in Foundry’s release; her post does not mention Foundry.
The Structural Read
As we read the release, Foundry sells capacity rather than machines: it says it owns and operates the factory and takes customers from design to production in weeks without new production line investments.
As we read the release, the AI claim rests on a loop between the floor and the models: the release says every production run feeds back into the models that run its robotic cells.
As we read the two sources side by side, the subassembly examples in Michelle Sun’s map, such as wire harnesses, overlap with applications Foundry lists; her post does not mention Foundry.
Adarsh Kulkarni, Foundry Robotics founder and chief executive, in the release
“Our approach is to own the entire final assembly stack and integrate intelligence at every layer”
Three Implications
FOR DEFENSE SUPPLIERS The release quotes HII Unmanned Systems President Duane Fotheringham saying production capacity is a national constraint on getting unmanned systems to the fleet.
FOR HARDWARE MAKERS The release says customers can go from design to production in weeks without new production line investments or expensive retooling.
FOR INVESTORS The release does not give a valuation or revenue; by our arithmetic the round equals about $769 per square foot of the new facility.
The Business Engineer Lens
This story maps onto the Business Engineer framework The Four Intelligence Moats.
The framework describes one moat this way: “The Container Moat — built by closed data loops inside a customer’s environment, nascent but the deepest position in the stack.”
As we read it, Foundry’s release describes a closed loop of its own, with every production run feeding back into its models, but inside factories it owns rather than in a customer’s environment. The framework was published in July 2026 and does not discuss Foundry.
What Is Not Established
The release does not give a valuation, revenue, the number of customers or the size of earlier funding. The description of the company as running America’s first autonomous factory for assembly is the company’s own claim. The twentyfold expansion is the company’s figure; the earlier footprint and the dollars per square foot are our arithmetic, and the round also funds teams and AI deployment, not only the building.
Michelle Sun’s five-layer map is her own framework, published as commentary, not as research on Foundry.
The Bottom Line
Foundry Robotics says it raised a $100 million Series A led by Marlinspike Partners and expanded its manufacturing footprint twentyfold to a 130,000-square-foot facility in Emeryville. By our arithmetic, that implies an earlier footprint of about 6,500 square feet and about $769 raised per square foot of the new site.
95,000+ executives read Business Engineer for the AI strategy frameworks cited by ChatGPT, Claude, and Perplexity.
A note on sourcing. Round, investor, facility, application and model details and all company quotations come from Foundry Robotics’ 28 October 2026 release on Business Wire, read via its Morningstar syndication. The five-layer map and its quotations come from Michelle Sun’s 29 October 2026 post on X. The earlier footprint and the dollars per square foot are our arithmetic.
Sources: Foundry Robotics release via Business Wire (28 Oct 2026) · Michelle Sun on X (29 Oct 2026)









