mimic hand M1
About
About mimic hand M1
The M1 has 15 actively driven degrees of freedom arranged in a human-aligned form, with bidirectional backdrivable actuation that lets the fingers yield under contact. Fingertip tactile sensing, proprioceptive feedback and synchronized global-shutter cameras supply the signals needed for contact-aware policies and precise grasp adjustment.
mimic robotics couples the hardware with a data pipeline built around wearable demonstrations and large human-action datasets. That design supports transferring demonstrated motions into learned robot behaviours while preserving the hand strength needed for sustained industrial grasps. The company positions the system for manipulation research and deployment work requiring dexterity, repeatable contact and rich training data.
Capabilities
Source-backed capabilities
- Capability areas: Tactile sensing, Dexterous manipulation
- Workflows: End-effector manipulation
- Tactile interaction sensing
Claimed - 2026-07-16 - mimic robotics describes M1 as a tendon-driven industrial hand with 15 active DoF, backdrivable actuation, fingertip tactile sensing and synchronized vision for learned manipulation.
Source - Robotic grasping
Claimed - 2026-07-16 - mimic robotics describes M1 as a tendon-driven industrial hand with 15 active DoF, backdrivable actuation, fingertip tactile sensing and synchronized vision for learned manipulation.
Source - Dexterous manipulation
Claimed - 2026-07-16 - mimic robotics describes M1 as a tendon-driven industrial hand with 15 active DoF, backdrivable actuation, fingertip tactile sensing and synchronized vision for learned manipulation.
Source - Force-controlled manipulation
Claimed - 2026-07-16 - mimic robotics describes M1 as a tendon-driven industrial hand with 15 active DoF, backdrivable actuation, fingertip tactile sensing and synchronized vision for learned manipulation.
Source
Timeline
- Mimic Robotics launches M1 hand and U1 wearable dexterity stack
2026-07-16 - Product Launch - The M1 uses 15 active degrees of freedom across 21 joints, highly backdrivable bi-directional tendon actuation, dual encoders and fingertip tactile sensing. The U1 passive exoskeleton matches the M1 hand's kinematics and sensing so human demonstrations can be collected without operating a live robot. Mimic combines human video, wearable demonstrations, robot teleoperation and deployment data in one learning stack for industrial manipulation.
Source