Former Tesla Optimus scientist establishes UMA to create a humanoid robot for Europe.

      A former Tesla scientist involved in developing the intelligence for Elon Musk’s Optimus robot aims to replicate that success in Paris, this time focusing on the European market. Rémi Cadene, CEO and co-founder of the startup UMA, has introduced plans for a lightweight, AI-driven humanoid named Northstar, as reported by Bloomberg.

      This machine is being tailored for use in manufacturing plants, logistics warehouses, and eventually private residences. Cadene informed Bloomberg that UMA is currently engaging with around 50 prospective clients regarding how they might utilize it, prioritizing Europe before considering expansion into the United States or Asia.

      This approach is significant since the continent has been working over the past year to demonstrate its capability to compete in the humanoid robotics sector rather than yield dominance to American and Chinese competitors. Cadene is not the only former member of the Optimus team to venture into independent projects, as others have also left Tesla to create robotic systems and dexterous robot hands.

      Cadene's impressive background is arguably one of UMA's key advantages. He worked at Tesla for approximately three years, from 2021 to 2024, focusing on the AI for Autopilot and constructing the initial neural networks for Optimus before joining Hugging Face to lead LeRobot, an open-source toolkit essential for robotic learning globally.

      He has built a similarly impressive founding team, which includes chief science officer Pierre Sermanet, a veteran from Google DeepMind and NYU, chief technology officer Simon Alibert, a LeRobot co-founder, and chief robot officer Robert Knight, the designer of the widely utilized open-source SO-100 arm.

      UMA, which stands for Universal Mechanical Assistant, emerged from stealth mode in December 2025 with an investor roster featuring prominent organizations in the sector. Its backers include Greycroft, Red River West, Kima Ventures, and Factorial, with advisors such as Meta’s chief AI scientist and Turing Award recipient Yann LeCun, alongside Hugging Face co-founder Thomas Wolf.

      The financial details are less clear. Prior to the public launch, Cadene was said to be seeking around $40 million in seed funding, though UMA has not confirmed any final figures, leaving the size of any closed funding round uncertain.

      However, UMA has clearly outlined its product vision. Launch materials indicate plans for a mobile industrial robot equipped with dual arms for warehouses and assembly lines, along with a more compact humanoid designated for settings that involve human interaction, such as hospitals, laboratories, and homes.

      The company also intends to launch several pilot programs in logistics, manufacturing, and healthcare throughout 2026, aiming to establish Northstar as a machine capable of perceiving, moving, and manipulating objects in real-world environments rather than just in controlled demonstrations.

      This strategy takes advantage of Europe’s industrial landscape and its significant labor shortages, particularly in warehouses facing high turnover and healthcare systems that lack millions of workers. These needs are exactly where humanoid robots intended for homes and factory settings are being most aggressively marketed.

      The market is competitive and well-funded. European counterparts like Germany’s Neura Robotics and Stuttgart-based Sereact have successfully raised substantial funds over the past year, while American companies like Figure and 1X continue to achieve deployment targets alongside impressive valuations.

      UMA’s bet is that a European team rooted in open-source principles can expedite software development over hardware and that European clients might prefer purchasing robots produced locally. Whether UMA can transform a group of top researchers into a deliverable product remains a significant question, which is the same challenge every humanoid startup faces while investing heavily in the future of physical AI. For now, Cadene has established a name, assembled a team, identified a target market, and developed a robot that must still demonstrate its capabilities.