Ex-Tesla Optimus scientist unveils European humanoid robot startup

Rémi Cadène, a former scientist for Tesla’s Optimus project, has launched a Paris-based robotics startup called UMA to develop the Northstar lightweight humanoid robot. The venture aims to address labor shortages and high operational costs in Europe’s manufacturing and logistics sectors by providing general-purpose, AI-powered assistants. This move signifies a strategic shift in the humanoid robotics race, positioning a European-led firm against established American and Chinese competitors like Tesla, Figure, and Unitree.

UMA, which stands for Universal Mechanical Assistant, emerged from stealth in December 2025 under the leadership of CEO Rémi Cadène. Cadène brings significant expertise from his three-year tenure at Tesla’s Autopilot group, where he worked on AI systems for both driver-assistance software and the Optimus humanoid robot. He later led the development of the open-source robotics library LeRobot at Hugging Face, which has since become a core infrastructure in the field with over 12,000 GitHub stars. The startup’s founding team also includes former Hugging Face engineer Simon Alibert and robot designer Rob Knight, supported by high-profile investors such as Greycroft, Relentless, and Unity Growth, alongside angel investors like Yann LeCun and Datadog CEO Olivier Pomel.

The company’s primary product, a lightweight humanoid named Northstar, is designed for use in manufacturing plants, logistics warehouses, and eventually residential settings. Unlike many competitors focusing on the U.S. or Chinese markets, UMA is prioritizing Europe as its initial beachhead. Cadène cites the continent’s aging workforce, high labor costs, and dense industrial base as primary drivers for demand. The startup claims to be in discussions with 50 potential customers and aims to launch industrial pilot programs as early as this year, leveraging a talent pool drawn from DeepMind, Tesla, Nvidia, and Hugging Face to build a homegrown European champion.

UMA enters a highly competitive landscape where established players are struggling to transition from prototypes to practical application. While Tesla CEO Elon Musk has positioned Optimus as a future cornerstone of the company, he recently admitted the robots are not yet performing useful work at scale and have no external customers. In contrast, competitors like Figure have already deployed robots at BMW’s Spartanburg plant, and Boston Dynamics continues its partnership with Hyundai. Although UMA’s early prototypes are viewed as being in the nascent stages of hardware development, the company’s focus on the software bottleneck—enabling robots to interact autonomously with physical environments—could be a critical differentiator in the race for commercial viability.