NVIDIA has announced today Project GR00T, a foundational general-purpose model for humanoid robots designed to drive advancements in embodied robotics and artificial intelligence. As part of this initiative, the company has also unveiled a new computer, Jetson Thor, for humanoid robots based on the NVIDIA Thor system on chip (SoC), as well as significant enhancements to the NVIDIA Isaac™ robotics platform, including generative AI foundational models and tools for simulation and AI workflow infrastructure.
“Building foundational models for general-purpose humanoid robots is one of the most exciting challenges to solve in AI today,” said Jensen Huang, founder and CEO of NVIDIA. “Enabling technologies are coming together for leading robotics experts worldwide to take giant steps towards general artificial robotics.”
GR00T-powered robots, which stands for Generalist Robot Technology 00, will be designed to understand natural language and emulate movements by observing human actions, quickly learning coordination, dexterity, and other skills to navigate, adapt, and interact with the real world. In his GTC keynote address, Huang demonstrated several of these robots performing a variety of tasks.
Jetson Thor was created as a new computing platform capable of performing complex tasks and interacting securely and naturally with people and machines. It features a modular architecture optimized for performance, power, and size. The SoC includes a next-generation GPU based on the NVIDIA Blackwell architecture with a transformer engine that delivers 800 teraflops of 8-bit floating-point AI performance to run multimodal generative AI models like GR00T. With an integrated functional security processor, a high-performance CPU cluster, and 100GB Ethernet bandwidth, it significantly simplifies design and integration efforts.
NVIDIA is building a comprehensive AI platform for leading humanoid robot companies like 1X Technologies, Agility Robotics, Apptronik, Boston Dynamics, Figure AI, Fourier Intelligence, Sanctuary AI, Unitree Robotics, and XPENG Robotics, among others.
“We are at a turning point in history, with human-centric robots like Digit on the cusp of forever changing work. Modern AI will accelerate development, paving the way for robots like Digit to assist people in all aspects of daily life,” said Jonathan Hurst, co-founder and robotics director of Agility Robotics. “We are excited to partner with NVIDIA to invest in computing, simulation tools, machine learning environments, and other infrastructure needed to make the dream of robots being a part of everyday life a reality.”
“Embodied AI will not only help address some of humanity’s greatest challenges, but also create innovations that are currently beyond our reach or imagination,” said Geordie Rose, co-founder and CEO of Sanctuary AI. “Such important technology should not be built in silos, which is why we prioritize long-term partners like NVIDIA.”
The Isaac tools used by GR00T are capable of creating new foundational models for any robot embodiment in any environment. These tools include Isaac Lab for reinforcement learning and OSMO, a computing orchestration service.
Embodied AI models require vast amounts of real and synthetic data. The new Isaac Lab is a lightweight, performance-optimized, GPU-accelerated application built on Isaac Sim specifically to run thousands of parallel simulations for robot learning.
To scale robot development workloads across heterogeneous computing environments, OSMO coordinates data generation, model training, and software/hardware loops in distributed environments.
NVIDIA has also announced Isaac Manipulator and Isaac Perceptor, a collection of pre-trained models, libraries, and reference hardware for robotics.
Isaac Manipulator offers next-generation modular AI dexterity capabilities for robotic arms, with a solid collection of foundational models and GPU-accelerated libraries. It provides up to 80 times faster path planning and zero-shot perception increases efficiency and performance, enabling developers to automate a greater number of new robotic tasks. Among the ecosystem’s early partners are Yaskawa, Universal Robots, a Teradyne company, PickNik Robotics, Solomon, READY Robotics, and Franka Robotics.
Isaac Perceptor provides immersive 3D vision capabilities with multiple cameras, increasingly used in autonomous mobile robots deployed in manufacturing and fulfillment operations to enhance worker efficiency and safety, as well as reduce error rates and costs. Early adopters include ArcBest, BYD, and KION Group, aiming to achieve new levels of autonomy in material handling operations and beyond.