Robotic factories that build robots
HQ: Emeryville, CA, United States | Founded: 2025 | Employees: 7 | Stage: Seed (Y Combinator S26) | Website: https://tensr.com/
Tensr builds robotic factories that make robots, operating today from a 12,000 sqft facility in the Bay Area. It is filling customer orders for some of the largest robotics companies in the United States, including complete humanoids, and has a partnership to manufacture robotics for the International Space Station with NASA's Johnson Space Center in Houston. Its thesis is that reviving domestic manufacturing means driving cost down through automation, and that the missing capability is a factory able to handle a new order end to end on its own — supply chain, design for manufacturing, retooling industrial machines, adapting robots, inspection, packaging and logistics. The stated goal is to go from a placed order to a completed manufacturing run within 24 hours. Technically, Tensr treats the factory itself as a data harness: mistakes such as flawed customer designs, failed supply-chain predictions, slow retooling and imprecise robot handling become learning signal, on the belief that robotics machine learning should improve from its own real-world errors rather than data collected by people. The team comes out of UC Berkeley robotics: chief executive Eric Berndt raced autonomous IndyCars at 160mph, chief technology officer Adith Sundram worked on robotics and controls at Berkeley, and founder C.K. Wolfe holds a computer science PhD in AI and robotics from Berkeley, ran the Berkeley robotics lab programme under BAIR and raised over $17M in DARPA and industry sponsorship.
Tensr sells the unglamorous precondition for the robotics boom: someone has to actually build the machines. It already runs a 12,000 sqft factory in the Bay Area producing for large US robotics companies, complete humanoids, and robotics bound for the International Space Station — real revenue where most competitors in this batch have only demonstrations. The thesis is that domestic manufacturing got expensive because it never automated, and that an automated factory can push cost down toward materials plus electricity. The interesting technical idea is the 'data harness': treating the factory as a system whose mistakes — flawed designs, bad supply-chain forecasts, slow retooling, imprecise handling — become training data. The risk is that this is a manufacturing company with manufacturing economics: capital-hungry and slow.
Browse: Home | Blog | About | All Startups A-Z | View full profile on StartupWiki