The era of humanoid robots could be a significant step closer thanks to a new release from Nvidia.
The computing giant has announced the launch of Project GROOT, the new fundamental model aimed at aiding the development of such robots in industrial use cases.
Unveiled at Nvidia GTC 2024, the company says its new launch will enable robots to be smarter and more functional than ever before – and they’ll do it by watching humans.
We are BIG?
Announcing the launch of Project GROOT (stands for “Generalist Robot 00 Technology”) on stage at Nvidia GTC 2024, Jensen Huang, the company’s founder and CEO, revealed that robots powered by the platform will be designed to understand natural language and emulate movements by observing human actions.
This allows them to quickly learn coordination, dexterity, and other skills to navigate, adapt, and interact with the real world – and absolutely won’t lead to a robot uprising.
Huang then showed a number of demos in which Project GROOT-powered robots performed a number of tasks from XXX, showing off its capabilities.
“Building basic models for general humanoid robots is one of the most exciting problems we can solve in AI today,” says Huang. “The enabling technologies are coming together to allow leading roboticists around the world to make giant leaps toward artificial general purpose robotics.”
The magnitude of the importance for Project BIG was also highlighted by the fact that Nvidia has built a new computing system, Jetson Thor, specifically designed for humanoid robots.
The SoC includes a GPU based on the latest Nvidia Blackwell architecture, which includes a transformer engine capable of delivering 800 teraflops of AI performance, allowing them to run multi-modal generative AI models such as GR00T.
The company also unveiled upgrades to its Nvidia Isaac robotics platform, designed to make robotic arms smarter, more flexible and more efficient than ever – making them a much more attractive choice for factories and industrial applications around the world.
This includes new collections of pre-trained robotics models, libraries and reference hardware aimed at faster learning and better efficiency.