NVIDIA today announced general availability of its Jetson AGX Thor developer kit and production modules for robotics and physical AI applications. With demand growing for robots capable of operating in unstructured environments and interacting safely with people, NVIDIA is positioning Thor as the onboard computer to make that happen.
Built on the NVIDIA Blackwell GPU architecture and equipped with 128GB of memory, Jetson Thor delivers up to 2,070 FP4 teraflops of AI compute within a 130-watt power envelope. That represents a 7.5x increase in AI performance and 3.5x higher energy efficiency compared with its predecessor, Jetson Orin. NVIDIA said the performance boost is critical for enabling robots to run multiple generative AI models simultaneously at the edge, unlocking real-time perception, reasoning, and interaction.
“We’ve built Jetson Thor for the millions of developers working on robotic systems that interact with and increasingly shape the physical world,” said Jensen Huang, founder and CEO of NVIDIA. “With unmatched performance and energy efficiency, and the ability to run multiple generative AI models at the edge, Jetson Thor is the ultimate supercomputer to drive the age of physical AI and general robotics.”
NVIDIA said several of the world’s top robotics and industrial technology companies are already integrating Jetson Thor. Agility Robotics plans to use Thor as the compute backbone for its Digit humanoid robot. Agility said Thor will allow it to run larger, more powerful, and more intelligent policies and reasoning models on Digit locally, in customer facilities and deployments. Agility said it’s particular interested in how Thor will help Digit with end-to-end manipulation and advanced scene understanding. From Agility Robotics:
“Digit’s ability to interact with objects in human environments depends on reasoning about context, contact, and control all at once. Running larger, end-to-end manipulation policies onboard will allow Digit to react faster and handle more complex tasks, from stocking shelves to assisting in logistics, and will improve generalization to new environments.”
NVIDIA Jetson Thor developer kit. | Credit: NVIDIA
“For robots working around people, awareness is everything. With Thor, we’ll explore richer onboard semantic perception. Again, with deep learning, scale at the edge is everything – so we’ll be able to test larger and more powerful models that can better handle the variable environments and problems our robots have to act in.”
Boston Dynamics is also using Thor into its Atlas humanoid for compute, AI workload acceleration, and high-bandwidth data handling directly onboard. Last week Boston Dynamics and Toyota Research Institute released a video about how large behavior models are advancing Atlas’ skills. If you haven’t seen the video, it’s incredible. Check it out below:
Other leading companies using Thor include Amazon Robotics, Caterpillar, Figure, Medtronic, and Meta. NVIDIA said John Deere and OpenAI are currently evaluating Thor.
“The future of robotics in logistics depends on the ability to deploy increasingly intelligent and autonomous systems,” said Tye Brady, chief technologist at Amazon Robotics. “NVIDIA Jetson Thor offers the computational horsepower and energy efficiency necessary to develop and scale the next generation of AI-powered robots that can operate safely and effectively in dynamic, real-world environments, transforming how we move and manage goods globally.”
NVIDIA said more than 2 million developers and over 150 hardware, software, and sensor partners are part of the broader NVIDIA Jetson software and hardware ecosystem. Jetson platforms now support a wide range of frameworks and models, including NVIDIA Isaac simulation and development tools, the Isaac GR00T humanoid foundation model, and domain-specific AI for vision, sensing, and manipulation.
The Jetson AGX Thor developer kit is available now starting at $3,499, while Jetson T5000 production modules can be purchased through global distributors and embedded partners. NVIDIA’s network of hardware integrators and sensor suppliers are already developing production-ready systems to help robotics companies accelerate deployment.
Connect with NVIDIA at RoboBusiness
NVIDIA will have a major presence at RoboBusiness, which takes place Oct. 15-16 in Santa Clara, Calif. and is produced by The Robot Report. Deepu Talla, VP of robotics and edge AI at NVIDIA, will deliver the opening keynote on “Physical AI for the New Era of Robotics.” The big bang of generative AI for the physical world is here, according to Talla. Physical AI, where models can perceive, reason, and act in real-world environments, is redefining what’s possible in robotics.
Jim Fan, director of AI and a distinguished research scientist at NVIDIA, will be on a keynote panel about lessons learned from early humanoid deployments. He’ll be joined by Katlyn Lewicke, head of global automation strategy and intel at GXO Logistics, and Pras Velagapudi, CTO of Agility Robotics.
Amit Goel, director of product management for autonomous machines at NVIDIA, will be on a keynote panel about the State of Robotics. Goel and other industry leaders from Cobot, DHL, and F-Prime Capital will discuss the breakthroughs driving progress, the barriers still holding robotics back, and the trends shaping robotics in the years to come.
Founded in 2004, RoboBusiness is the premier event for developers and suppliers of commercial robots. There will be 60-plus speakers in the conference, a startup workshop, a robotics startup competition, networking receptions, and more. Over 100 exhibitors on the show floor will showcase their latest enabling technologies, products, and services to help solve your robotics development challenges.
