Robot hand SharpaWave with sensitive robotics
The robotics company Sharpa focuses its work on AI robots on grippers and sensors for human-like hands rather than locomotion mechanics.
The robotics company Sharpa develops sensitive, human-like robot hands.
(Image: heise medien / André Kramer)
With the SharpaWave robot hand, the company Sharpa takes inspiration from nothing less than the human model. It sees a technical gap for robotics manufacturers in this area. After all, the locomotion of robots is largely technically solved, as Agibot and Boston Dynamics, among others, demonstrate.
However, handling objects still presents a hurdle. Unlike autonomous driving, robots for industry and households must constantly interact with objects and people.
Videos by heise
Hands as the key to physical intelligence
According to Sharpa, the biggest technical bottleneck in robots lies in fine motor skills. Hands and lips account for a large part of sensory processing in the human brain. Accordingly, a powerful gripping technology is a prerequisite for trainable machines.
The company points out that while simple grippers can handle transport tasks, complex operations such as opening doors, operating tools, or precisely gripping delicate objects require more sensitive technology.
(Image: heise medien / André Kramer)
SharpaWave with 22 degrees of freedom
Sharpa is showcasing the SharpaWave robot hand at CES 2026. It features 22 actively controlled axes of motion and is similar in size and design to the human hand. Tactile sensors at the fingertips capture pressure and contact with high resolution.
According to the manufacturer, the hand can move quickly and grip forcefully. This combination is intended to enable it to operate scissors, grasp small objects, and turn pages. Sharpa states that the hand is designed for more than a million gripping cycles.
Data quality over data quantity
Sharpa emphasized the importance of high-quality training data for trainable robots. Real-world interaction data from the physical world is more valuable than large amounts of unspecific image or text data. Hardware like the SharpaWave robot hand facilitates access to such data and improves its usability for learning processes.
In the company's view, the quality of the mechanics determines the experiences a robot can gather. Only with precise sensor technology and robust construction can a reliable foundation for advanced AI systems be created.
heise online is the official media partner of CES 2026
(akr)
