Twindex enables dexterous manipulation without physical data

On September 2, 2026, the embodied robotics company Zibian Variable unveiled TwinDEX, a pair of twin three-fingered, nine-degree-of-freedom dexterous hands. These hands are designed for data collection and deployment on robots, marking the first instance of dexterous manipulation driven entirely by non-physical data. Evaluations show that TwinDEX achieves a 5.3x data collection efficiency compared to traditional teleoperation, with non-physical data nearly fully replacing physical data in training models. Zibian Variable claims this redefines the industry's understanding of the 'data pyramid,' as non-physical data now matches the effectiveness of physical data in training. TwinDEX achieves this through consistency in kinematics, contact mechanics, and visual appearance, along with optimized data pipelines and training processes. The system uses wearable exoskeletons for data collection, enabling scalable, parallel data gathering without reliance on physical robots. This approach allows for significantly higher data collection efficiency, with TwinDEX achieving 5.3x the rate of traditional teleoperation. The company states that non-physical data can now effectively replace physical data in training, breaking the long-standing 'impossible triangle' of data quality, dexterity, and scalability. TwinDEX's design enables the first successful integration of these three factors, paving the way for broader, more diverse data sources in robot training.