Nvidia Integrates Chinese Robotic Hands into Humanoid Systems
The chip giant is using components from startups like Wuji and Sharpa to advance its robot training platform.
Nvidia has incorporated components manufactured by companies with strong ties to China into its new training systems for humanoid robots. The devices use mechanical hands developed by startups Wuji and Sharpa to support the Isaac software and the Jetson Thor platform.
CEO Jensen Huang showcased the Agile One SN system, which combines an Agile Robots body with Wuji hands, at a recent event in Berlin. The technology aims to accelerate the development of autonomous robotics and is expected to be available to research institutions like Stanford and ETH Zurich starting in 2027.
Another reference model announced by the company uses a body from the Chinese firm Unitree and hands from Sharpa, which is based in Singapore with research operations in Shanghai. These startups have gained such relevance that Google DeepMind researchers have compared them in technical evaluations regarding strength and motor precision.
This reliance on Chinese-made components for high-tech systems comes as Nvidia expands its robotics infrastructure. The use of these parts aims to solve complex technical challenges in object manipulation, as demonstrated by a robot's ability to spin a pen autonomously.
Spotted an error? Write to suporte@criptoico.com.br.
