Foundry Robotics Raises $100M for Autonomous Assembly

Changing a connector can change the factory assignment behind a hardware product. Foundry Robotics is building its business around making those changes easier to absorb, with an assembly service that puts robotic cells, manufacturing staff and production infrastructure inside the same supplier relationship.

The Emeryville, California company announced a $100M Series A on October 28, 2026, led by Marlinspike Partners. Co-founders Adarsh Kulkarni and Emmanuel Lamacchia are using the financing to expand a factory-centered approach to physical AI: customers bring the hardware they need built, while Foundry takes on the operation that assembles it. The commercial question reaches beyond how well a robot moves. It concerns how much manufacturing infrastructure a hardware company needs to commit to before its product and demand have settled.

What Foundry Robotics is financing

Foundry's company-issued announcement identifies Khosla Ventures and Red Glass Ventures as early backers participating in the round, alongside existing investors Hanabi and Zero Shot Fund. New investors include SE Ventures, Wing VC and Axiom Ventures. The release describes the Series A as oversubscribed and directs the proceeds toward the Emeryville buildout, deployment of AI models into manufacturing, and growth in robotics research, engineering and factory operations.

The company reports a 130,000-square-foot facility, giving this financing a physical destination as well as a research agenda. Floor space alone cannot explain the business. Foundry also has to bring together the equipment, software and production organization needed to turn that space into a service customers can rely on.

Marlinspike's stated investment focus includes technologies with defense and commercial applications. Foundry's customer positioning spans those markets, giving the lead investor a direct connection to the manufacturing assignment. Its participation establishes investment conviction; production economics will emerge from the work performed inside the factory. The announcement leaves valuation and detailed operating metrics undisclosed, so the round size should be read as financing for the next stage of execution rather than a measure of factory output.

A manufacturing customer buys capacity and responsibility

Foundry describes its offering as manufacturing-as-a-service. Under that arrangement, the company owns and operates the factory, bringing complex electromechanical products into production through a combination of adaptable robotic cells and skilled staff. Its company overview places the ability to change designs and scale production inside the customer proposition.

That allocation of responsibility matters to a hardware business. A dedicated line can tie up capital while the product team is still learning what customers need. Working with an outside assembly operator gives the buyer another way to approach that commitment, although the commercial value will depend on price, quality, delivery and the cost of moving a design into production.

Foundry consequently has an unusually broad product to deliver. Its customer experiences the combined result of engineering decisions, material coordination, assembly and inspection. An impressive robotic action helps only when it fits that wider production sequence. The service model brings those obligations back to the company running the floor.

For an operator evaluating Foundry, the useful conversation starts with a specific product and production requirement. Which operations can the supplier handle, what changes require additional engineering, and how will quality and delivery be measured? Those questions connect the physical AI proposition to the purchase a manufacturing team actually needs to make.

Robotics learns inside a working production system

Foundry's factory description combines autonomous robotics with manufacturing expertise and software coordination. Its official site also describes testing, inspection and digital traceability, placing records of the production process alongside the machines doing the assembly. That combination is relevant when a customer needs to understand how a finished unit was built and checked.

The funding announcement names FoundryOS as the software connecting robots, the factory floor and the cloud. Foundry says production runs feed learning back into its AI models. This gives the company a proposed relationship between the service business and the research program: operating a factory creates experience that can inform the next manufacturing assignment.

The operating consequence is more demanding than collecting more data. A manufacturing business must decide which experience transfers between tasks, where people need to intervene, and how to introduce changes while maintaining the customer's requirements. These are implications of the model, rather than independently demonstrated performance results. The public announcement does not supply yield, utilization or unit-cost figures that would allow an outside reader to calculate the economic improvement.

Foundry's industry page illustrates applications including drones, underwater vehicles and telecom hardware. The range helps explain why adaptable assembly is part of the pitch: the company is pursuing capacity that can serve different hardware needs, rather than describing one dedicated product line as its entire market.

The founders bring the production problem into the company

CEO Adarsh Kulkarni's background includes Ghost Robotics and leading robotics work at Scale AI, according to the company release. Emmanuel Lamacchia previously worked on Tesla's Model Y program. Their combination places experience with robotic systems and vehicle-program execution inside the same founding team, with the service model requiring both technical development and day-to-day manufacturing judgment.

The company's careers page puts engineering and manufacturing in the same workplace. That detail gives the hiring plan commercial meaning: people developing robotics and software can work near the production situations their technology must handle. The Series A finances an organization around the machines as well as the machines themselves.

DevCuration's coverage of Maven Robotics' industrial-robot financing and Re:Build Manufacturing's expansion provides adjacent context for capital flowing into physical production. These are different businesses and different financing records. For Foundry, the distinctive commitment is to take a customer's assembly work into a factory it runs, where the next design revision becomes an assignment for its own engineers, technicians and robotic cells.