Flexion has secured $50 million in Series A funding to accelerate development of its autonomy stack, designed to serve as the intelligence layer for next-generation humanoid robots. The round was led by DST Global Partners, with participation from NVIDIA Ventures (NVentures), redalpine, Prosus Ventures, and Moonfire. The raise follows a $7.35 million seed round completed only months earlier.
The Zurich-based company specializes in reinforcement learning and a simulation-to-real platform that enables humanoid robots to perform complex physical tasks with far less human intervention than existing systems. While recent advances in hardware have produced machines with impressive mobility and dexterity, most still depend heavily on scripted behaviors, tele-operation, or costly, large-scale human demonstrations. Flexion’s goal is to replace these brittle approaches with adaptive robot intelligence that can reason, learn, and act autonomously in the real world.
Flexion’s autonomy stack spans three primary layers. Its Command Layer uses language models to interpret natural-language task descriptions and convert them into structured plans grounded in environmental understanding. The Motion Layer is a vision-language-action model mainly trained on synthetic data and fine-tuned for real-world edge scenarios. The Control Layer provides low-latency, transformer-based whole-body control, enabling rapid composition of new behaviors.
The company says this architecture allows robots to be deployed with dramatically reduced human supervision, eliminating the need for extensive scripting or large teleoperation teams. Flexion’s platform aims to unlock broad real-world utility across manufacturing, logistics, disaster response, and other labor-intensive environments.
The funding will support the expansion of Flexion’s Zurich R&D hub, scale its compute and robot fleets, establish a U.S. footprint, and deepen its work with major OEM partners. Flexion’s founding team includes specialists in reinforcement learning, control, perception, and mechatronics, with backgrounds at ETH Zurich, NVIDIA, Meta, Google, Tesla, and Amazon.
The company positions its mission as building the “brain” that turns humanoid hardware into autonomous systems capable of working alongside humans rather than depending on them. With demographic shifts, labor shortages, and increasing demand for automation, Flexion argues that general-purpose humanoid robots are quickly becoming an economic necessity rather than a distant vision.
