X Square Robot Launches TwinDEX Exoskeleton to Accelerate Robotic Skill Training

2026-09-02 companies

San Francisco, Thursday, 3 September 2026.
X Square Robot unveiled TwinDEX, a wearable exoskeleton system that drastically speeds up robotic training data collection without using physical robots, achieving 5.3 times higher efficiency.

Launch Overview and Strategic Timing

X Square Robot unveiled TwinDEX, a wearable exoskeleton system that drastically speeds up robotic training data collection without using physical robots, achieving 5.3 times higher efficiency [2][6][7]. On 1 September 2026, the Shenzhen-based company introduced the platform to address persistent bottlenecks in manufacturing and logistics automation [3][4][8]. The system pairs a wearable exoskeleton interface for human operators with a twinned three-finger deployment hardware, significantly reducing training overhead for complex 20-plus degree-of-freedom robotic tasks [2][8]. This innovation targets labor-intensive industries seeking advanced automation tools by enabling robot-free data collection in diverse environments such as offices and kitchens [4][5].

Technical Architecture and Isomorphism

TwinDEX features a co-designed, three-finger, nine-degree-of-freedom (DoF) architecture comprising seven active and two passive degrees of freedom [3][5][8]. The twinned design shares identical kinematics, contact surfaces, visual appearance, and sensors across collection and deployment to keep observations and actions aligned end to end [2][5]. By aligning kinematic chains, joint axes, link proportions, and sensor placement, the system eliminates the need for complex hand-to-robot retargeting [3][8]. The platform synchronizes multi-view RGB camera inputs, 6-DoF wrist poses, finger joint states, and fingertip tactile signals to ensure consistent policy learning between the wearable device and the robotic end effector [3][4].

Efficiency Metrics and Market Position

In performance evaluations, TwinDEX achieved up to 5.3 times the effective throughput of traditional on-robot teleoperation [3][6][7]. The system was validated via a standardized chemistry experiment involving 24 sub-actions, including tool switching and bimanual coordination, all executed autonomously in a single uncut run [4][8]. X Square Robot is currently positioned against competitors like Dyna Robotics and Sunday Robotics, with emphasis on multi-finger articulation supported by funding from Xiaomi and Meituan [5]. The company aims to scale data collection to tens of thousands or millions of episodes, removing the dependency on real-robot hardware that has been the central bottleneck to scaling dexterous manipulation data [2][5].

Sources

Industrial Robotics Automation Trends