Yann LeCun Backed This Zurich Robot Startup—Here's Why

Jan de Vries · 2026-10-22

Most robots today are basically all eyes and no sense of touch. They can see the world around them, but they can't really feel it. That's the gap a small Zurich company called Enact Intelligence is trying to close, and some very big names in AI think it's onto something. Enact just came out of stealth with a €26.68 million ($30 million) pre-seed round. According to the company, that makes it one of the largest pre-seed raises in Europe to date. Not bad for a team of twelve people. ### Who's Backing This Thing? The investor list reads like a who's-who of AI and robotics. Institutional money came from redalpine, Notion Capital, and firstminute capital. Then there were the angels: - Arthur Mensch, founder of Mistral - Yann LeCun, founder of AMI Labs and former Chief AI Scientist at Meta - Andrew Lonsberry, founder of Path Robotics Dr. Harald Nieder, General Partner at redalpine, explained the appeal pretty simply. Dexterity is the next frontier in robotics, and Enact is going after it with a new kind of data. Instead of relying on fingertip-style force sensors, the company pulls high-precision force readings straight from custom-built actuators. As Nieder put it, they're "rethinking the entire stack from the hardware up" to build a real data moat. ### The Idea: Robots That Feel Through Their Muscles Enact was founded in March 2026 by Manuel Yves Galliker. Before this, he led the software team behind the humanoid robot NEO at 1X, helped build that company's AI team in Silicon Valley, and spent time at the robotics lab Physical Intelligence. So he's seen how the sausage gets made. His core argument is straightforward. Vision-only robots struggle to manipulate objects reliably or learn from experience. And existing haptic solutions tend to be either too expensive or too fragile for real industrial work. So Enact rebuilt the whole stack, from model architectures and data pipelines down to the actuators and sensors. Rather than point-based force sensors, its robots capture force data from the actuators themselves. Think about how you sense tension through your muscles, not just your fingertips. > "Unlike the early days of LLMs, physical AI has the advantage of starting from strong reasoning, language, and vision capabilities already in place," Galliker said. "We can build on top of existing models, like video models, and adapt them to output robot actions for interacting with the real world." His goal is to build what he calls "a proprioceptive nerve system and the cortex that makes sense of it." In plain terms: solve dexterity through whole-body force awareness, at a cost that actually works for industrial scale. ### A Zurich Talent Play The twelve-person team pulls engineers from DeepMind, 1X, Amazon Robotics, Helsing, Physical Intelligence, and ETH Zurich. Along with the funding, Enact brought on Philip Kneis, a former DeepTech investor at redalpine, to lead growth. Galliker's take on the geography is refreshing. Silicon Valley is great at software and AI, he says, but thin on robotics and hardware talent. Plenty of money is chasing very few people. Enact's plan is to bring the AI experience from Silicon Valley and pair it with Zurich's deep local talent pool. "That's the winning formula," he added. ### What Happens Next The fresh capital will go toward growing the team fivefold by the end of 2027, speeding up hardware and R&D work, and rolling out robots with industrial partners in manufacturing, infrastructure, retail, and food processing. For anyone watching the physical AI space, Enact is worth keeping an eye on. The bet is bold, the backers are serious, and the timing—as robots move from demos to factory floors—couldn't be more interesting.