Enact Intelligence, a Zurich-based physical AI startup, has emerged from stealth with a €26.68 million (about $30 million) pre-seed round — reported by Sifted as the largest pre-seed ever raised in Switzerland — to attack one of robotics’ oldest unsolved problems: giving robots a working sense of touch.
The round was led by redalpine, Notion Capital and firstminute capital. Angel investors include Yann LeCun, the former Meta chief AI scientist now building AMI Labs, Mistral chief executive Arthur Mensch, and Path Robotics founder Andrew Lonsberry.
A Motor That Feels Force
Enact is building its own motors that can sense how much force to apply at each joint — an approach it calls haptic intelligence — so robots can handle work outside standardised factory lines, such as closing a box or lifting a pumpkin without crushing it. The company describes its wider stack as model architectures, data pipelines, actuators and sensors combined into a proprioceptive “nerve system” and “cortex” that give a robot whole-body force awareness.
The bet is that dexterity is a hardware problem before it is a software one. Today’s industrial robots mostly earn their keep on single optimised tasks like sorting and transport, while decades of research have not produced a machine that can reliably peel an orange, turn a key in a lock or untangle a cable. Enact argues that motors which report force back to the model make the underlying AI easier to train, because the robot can learn from touch rather than vision alone.
The Team Behind It
Founder and chief executive Manuel Yves Galliker, an ETH Zurich graduate, previously led the software team for 1X’s NEO humanoid robot in Silicon Valley and worked at the robotics lab Physical Intelligence; a cofounder also comes from Physical Intelligence. The company was founded in March 2026 and is based in Zurich’s Binz quarter, with hiring under way across research, controls and motor engineering roles.
The round sits in the top 1% of seed-stage deals globally over the past 48 months, according to Dealroom data covering more than 47,000 rounds. Swiss robotics has been drawing larger cheques: Zurich’s Mimic Robotics raised $16 million in November 2025 for robot hands, and Gravis Robotics followed a $23 million round with a $200 million Series A led by SoftBank in August 2026. Enact’s raise follows the same pattern as other recent physical-AI rounds covered by GEroundtable, from Eka Robotics’ $150 million Series A to Reliable Robotics’ $115 million raise, while sensing work such as LG Innotek’s robot vision system attacks the perception side of the same problem.
Why It Matters
Enact Intelligence’s $30 million pre-seed matters because of who wrote the cheques and what they signal about where physical AI is heading. LeCun has argued publicly that world models and touch-rich data, not language models alone, are the path to capable machines, and Mensch’s participation ties the round to Europe’s frontier-lab establishment; their backing suggests investors now treat robot dexterity as an AI research problem worth frontier-scale early money rather than a mechanical engineering niche. The test for Enact is whether its force-sensing motors translate into deployable dexterity — robots that can be trained faster because they can feel — rather than another laboratory demonstration.
Background
Enact was founded in March 2026 by alumni of 1X, the Norwegian-American humanoid maker behind the NEO home robot, and Physical Intelligence, the San Francisco lab known for general-purpose robot foundation models; both organisations sit at the centre of the current physical-AI boom, in which investors have poured billions into teaching machines to manipulate the real world after language models conquered text.
0 Comments