U1 Context

Market context

Real DexUMI system positioned for embodiment-agnostic dexterous hand data collection in real-world manipulation.

• Target environment: Robot-learning labs, dexterous manipulation research, embodied-AI data collection and teleoperation setups.
• Workflow context: Dexterous data collection, transfer learning, manipulation training, deployment testing and real-world task execution.
• Customer context: Researchers, AI teams and robotics developers building dexterous manipulation datasets and policies.
• Deployment model: Data and training system for dexterous manipulation rather than a general end-user robot.
• Commercial maturity: Emerging research/product surface.
• Adoption constraints: Adoption depends on embodiment match, data quality, hardware compatibility and training workflow fit.
• Market position: Data bridge for dexterous manipulation moving UMI-style workflows beyond gripper-centric setups.
• Adjacent products: DexUMI, PIKA, Pika Pro, DM-EXton, tactile hand systems

Audience

AI labs

Workflow

Dexterous data collection, transfer learning, manipulation training, deployment testing and real-world task execution.

Deployment environment

Robot-learning labs, dexterous manipulation research, embodied-AI data collection and teleoperation setups.

Specifications

• Mass Kg: 680 g
• Total Joints: 11 joints
• Active Joints: 6 active joints
• Passive Joints: 5 passive joints
• Payload: 5 kg maximum payload
• Max Dynamic Tracking Error Mm: Less than 1 mm maximum dynamic tracking error
• Average Tracking Error Mm: Less than 0.1 mm average tracking error