Flexion is betting that the hardest part of humanoid robotics is not the body but the brain. Founded in 2024 and based in Zurich, the company builds the intelligence layer that makes humanoid robots functional in dynamic, unpredictable environments, deliberately leaving the mechanical hardware to others. Its founding team brings deep credibility: CEO Nikita Rudin and co-founder Marco Hutter come from ETH Zurich's Robotic Systems Lab, while co-founder and CTO David Holler previously led research at Nvidia, where he helped build Isaac Gym, a widely used robotics simulation platform.
The platform centers on adaptive autonomy. Flexion combines natural-language task comprehension so robots can understand instructions, synthetic vision-language-action training to learn behaviors at scale in simulation, and transformer-based full-body control to coordinate complex movement. Together these aim to produce robots that can perceive, reason about a task, and act with minimal human teleoperation or hand-tuning, which is the central obstacle to deploying humanoids usefully outside controlled demos.
By positioning itself as a horizontal intelligence layer, Flexion can in principle power humanoid hardware from multiple manufacturers rather than tying its fortunes to a single robot. This mirrors a broader thesis in robotics that the value will accrue to the software and learning systems, much as it has in autonomous driving and other physical-AI domains. The reinforcement-learning and simulation expertise on the team is core to that approach.
In November 2025, Flexion raised a EUR 43 million (about USD 50 million) Series A led by DST Global Partners, with participation from NVentures (Nvidia's venture arm), redalpine, Prosus Ventures, and Moonfire, following a EUR 6.3 million seed round months earlier. The funding will expand R&D in Zurich, scale compute and robotic infrastructure, establish a US presence, and drive commercial rollout of its autonomy stack. With elite research talent and strong backing, Flexion is one of Europe's most notable entrants in the race to build humanoid-robot intelligence.