Enact Intelligence raises $30M to turn robot motors into force sensors

Zurich-based robotics startup Enact Intelligence has raised $30 million in pre-seed funding to develop a system that transforms every motor in a robot into a force sensor. The round, announced on October 22, was led by redalpine, Notion Capital, and firstminute capital, with backing from high-profile angels including Yann LeCun, Arthur Mensch (CEO of Mistral AI), and Andrew Lonsberry (founder of Path Robotics). The company argues its technology could enable "whole-body force awareness" without relying on external tactile sensors, addressing a key limitation in industrial assembly tasks where vision-based robots often struggle.

Instead of depending on traditional force-torque sensors or fingertip-mounted tactile hardware, Enact’s approach interprets contact forces through the electrical signals of a robot’s actuators. Earlier research in the field has suggested this method could be effective, with studies indicating that motor current readings can correlate closely with contact forces—even without additional sensors. One such study reportedly improved task success rates significantly when this feedback was integrated into manipulation policies, though the exact conditions and hardware used in testing remain specific to controlled environments.

The startup’s innovation targets two persistent challenges in robotics: limited coverage and hardware fragility. Conventional force-torque sensors, in use since the 1970s, typically provide data at a single point, such as the wrist, while vision-based tactile sensors cover only small contact areas. These external sensors also introduce vulnerabilities, as they are often mounted in high-stress locations prone to failure. Enact’s solution avoids these issues by leveraging the robot’s existing actuators, turning every joint into a sensor. This could allow robots to adapt in real time to resistance—whether from inserting a cable or pressing a seal—without waiting for visual confirmation.

Founder and CEO Manuel Yves Galliker has described the system as creating a "proprioceptive nervous system" for machines, drawing a parallel to how humans sense effort through their entire body rather than just their fingertips. The architecture is designed to be versatile, working across different robot form factors, from arms to humanoid bodies, as long as they use similar actuators. To enhance signal quality, Enact is developing custom actuators optimized to reduce interference from factors like gearbox friction, backlash, and cable tension, which can otherwise obscure force readings.

The $30 million round is reportedly the largest pre-seed ever raised in Switzerland, signaling strong investor confidence in Enact’s vision. Kamil Mieczakowski of Notion Capital framed the investment as a bet on robots needing more than just visual perception to generalize effectively, while redalpine’s Harald Nieder emphasized the potential of a "new data modality" for dexterity. LeCun’s participation aligns with his long-standing advocacy for AI systems that learn through physical interaction rather than language alone—a position he has championed since leaving Meta in 2025 and raising over $1 billion for AMI Labs earlier this year.

Enact’s emergence comes as robotic manipulation sees heightened activity. Just days before its announcement, Eka Robotics, a Cambridge-based startup, raised $150 million at a $1.45 billion valuation for its Vision-Force-Action architecture, which combines cameras with force-sensitive grippers. While Eka’s approach retains some reliance on tactile hardware, Enact’s model eliminates it entirely, potentially reducing costs and failure points. Another competitor, Tangent Robotics, raised $4.5 million in September to tackle fine motor skills, but its focus remains on integrating external sensors rather than actuator-based sensing.

The critical question is whether Enact’s technology can perform reliably in real-world industrial settings. While lab tests have shown promise, industrial environments introduce variables—such as temperature fluctuations, electrical interference, and mechanical wear—that could affect signal accuracy. If the startup overcomes these challenges, the implications could extend beyond industrial assembly. A robot capable of sensing force through every joint might detect mechanical issues preemptively, adapt to unexpected resistance, or enable new applications in unstructured environments where vision alone falls short.

For now, Enact’s substantial funding suggests investors are betting on its ability to turn this vision into a scalable, industrial-grade solution. If successful, the startup could redefine how robots interact with the physical world—not just observing it, but truly feeling it.

Sources: techtimes.com