News

Remember the severe skepticism of the protagonist from the movie? Will Smith disliked humanoid robots for their heartless, cold, and calculated actions that he had noticed from past experiences. So, is NVIDIA jumping the gun here to support humanoid development? To be honest, at this juncture, robots still have some ways to go before they mimic the future shown in the movie, where humanoids are at the beck and call in every home. Until then, existing humanoids are primitive and mostly preoccupied with managing their movement. What NVIDIA is doing is progressing humanoid development to make them more useful for understanding natural language, emulating human actions to learn coordination, dexterity, and other skills to navigate and interact with the real world more fruitfully. This is , a general-purpose foundation model for humanoid robot development. At its heart, Project GR00T-based humanoid robots will embrace a new robotics platform powered by the all-new . Powered by the just-released , this high-performance SoC will have the chops to tackle multi-model generative AI models, a necessity for the expectations lined out earlier to progress robots to the next stage. Project GR00T will deploy a variety of upgraded tools required for robot foundation model training and development. Utilizing a version of , NVIDIA has carved up , which is specifically optimized for robot learning. Notable traits required for a successful humanoid deployment include imitation and transfer learning from humans, as well as reinforcement learning within a virtual environment to improve their reliability and facing many other challenges (synthetic data points) that could be time-consuming or dangerous to experience in the real world. To scale robot development workloads across heterogeneous compute, is a new orchestration service that coordinates the data generation, model training and software/hardware-in-the-loop workflows across distributed environments. Additionally, NVIDIA launched , which offers pre-trained models and libraries to speed up the robotics learning process and allow developers to progress towards tackling new robotic tasks. Accompanying it is , a hardware suite of multi-camera, surround-sound vision capabilities. Most of these new Isaac capabilities will be available from next quarter.   1)  2)