Zurich-based robotics software startup Flexion has raised $50 million in Series A funding to develop what it calls an “intelligence stack” for humanoid robots, betting that advances in generative AI can finally push machines beyond tightly scripted tasks.
The round was led by DST Global Partners and included NVentures, the venture arm of Nvidia, as well as redalpine, Prosus Ventures and Moonfire Ventures, Flexion said.
The company previously collected $7.35 million in seed capital from Frst, Moonfire and redalpine earlier this year.
Flexion is part of a growing wave of firms trying to translate the success of large language models into the physical world.
Investors have poured money into “physical AI” this year, including San Francisco-based Physical Intelligence, which recently raised $600 million, underscoring expectations that robots will take on more real-world work.
Some backers are also looking for picks-and-shovels software plays rather than capital-heavy manufacturing bets.
Unlike hardware-focused humanoid developers, Flexion aims to sell software that can run on different robot bodies.
Its platform combines language-model planning, vision-language-action models trained largely in simulation, and low-latency whole-body control designed to learn new skills without extensive human demonstrations.
The startup says current humanoids remain impressive in demos but are rarely deployed at scale outside controlled settings, because training them to handle edge cases is expensive and slow.
Flexion argues that simulation and reinforcement learning can narrow that gap and reduce reliance on remote operators.
The new funds will be used to expand Flexion’s research team in Zurich, increase computing capacity and robot test fleets, and open a U.S. presence as it works with original equipment manufacturers on commercial pilots.
Flexion did not disclose its valuation.
Demographic shifts and labour shortages are adding urgency. By 2050, about one-third of people in developed economies will be over 60, raising costs for industry and care work, analysts say, potentially widening the market for autonomous robots over time.
