NVIDIA presented the Project BIGa model that aims to enhance the development of humanoid robots for industrial applications. This new model promises to make robots not only more intelligent but also more functional, learning directly from the observation of humans.
NVIDIA focuses on humanoid robots with Project GROOT
Jensen Huangfounder and CEO of NVIDIA, announced at NVIDIA GTC 2024 (via Techradar) Project GROOT, an acronym for “Generalist Robot 00 Technology“. The robots that use this platform will be capable of understand natural language e di emulate human movements. And they will be able to improve coordination and dexterity in interacting with the surrounding environment by observing us.
To support Project GROOT, NVIDIA has created Jetson Thor, a system on chip (SoC) designed specifically for humanoid robots. This SoC is based on the architecture NVIDIA Blackwell and can provide up to 800 teraflop of AI performance, essential for running multimodal generative AI models like GROOT.
In addition, NVIDIA has announced updates to its Isaac robotic platformwhich aims to improve the intelligence, flexibility and efficiency of robotic arms. These updates include new collections of pre-trained modelslibraries and reference hardware that accelerate learning and optimize performance.
At the moment it is difficult to say whether the future of robots will depend on NVIDIA’s platform, but it seems clear that the company thinks that robots are fundamental to its future. AI and robotics could play a fundamental role in the near future — and NVIDIA wants to be a protagonist.
