Silica

Robots today are built with actuators that cannot sense what they touch, forcing software developers to write extensive workaround code to compensate for this lack of physical feedback. Most existing actuators estimate force indirectly by measuring current in the motor, which is an imprecise proxy for actual contact force. This limits what robots can safely and accurately do in tasks like manufacturing, lab automation, and training, where precise sensing of light or delicate forces matters.

Silica builds a nervous system directly into robotic actuator hardware so every joint can feel the world around it. Instead of estimating force from motor current, Silica's actuators measure torque at the output, at the exact point where the joint meets the load, so every demonstration, rollout, and contact is recorded as real force. The company states its actuators can sense forces as light as a paperclip, compared to today's actuators which can only detect something as light as a tennis ball. The product is showcased as the Q Series, with use cases mentioned including manufacturing lines, lab automation, training, and toy bipeds.