- Nvidia Jetson Thor is framed as the brain of the physics robot
- The main robotics companies are already listed as the first users worldwide
- Nvidia Jetson Thor offers 2,070 Teraflops FP4 inside a 130 watt feeding envelope
Nvidia has launched the Jetson Agx Thor developer kit, calling it the next step towards robotics systems that can work in real time.
The system, built on the Blackwell GPU line, is part of as a platform for “physical AI” and advanced robotic functions in manufacturing, logistics, medical care, agriculture, retail sale and transport.
Nvidia says that she can deliver up to 7.5 times more AI calculator and more than three times the energy efficiency of your Jetson Orin line, which has been of great use since 2022.
Offers capacity of supercomputer
Nvidia described Jetson Thor as “the final supercomputer to boost the age of physical and robotics.”
“We have built Jetson Thor for the millions of developers who work in robotic systems that interact and give the physical world more and more,” said Jensen Huang, founder and CEO of Nvidia.
“With unmatched performance and energy efficiency, and the ability to execute multiple generative models on the edge, Jetson Thor is the final supercomputer to boost the age of physical the physical and general robotics.”
With a aforementioned figure of 2,070 Teraflops FP4 in an envelope of 130 watts, it is placed as powerful enough to execute multiple generative models at the same time.
Admits Vision Language Models such as Isaac Gr00t N1.5, along with other LLM systems.
The device also integrates 128 GB of memory, which is expected to be able to handle larger AI workflows on the edge.
Several robotics players are already listed as the first users, including Agility Robotics, Amazon Robotics, Boston Dynamics, Caterpillar, Hexagon and Medtronic.
Meta has also been appointed as an early partner, while companies such as John Deere, OpenAi and Physical Intelligence are testing the system.
“Nvidia Jetson Thor offers the computational power and energy efficiency necessary to develop and climb the next generation of robots that can work safely and effectively into dynamic and real world environments, transforming the way we move and manage goods worldwide,” said Tye Brady, chief technologist of Amazon Robotics.
Nvidia points out that more than two million developers already use their robotics battery, with more than 7,000 clients that have implemented Hardware from Jetson Orin in Edge AI projects.
Jetson Thor runs on the Nvidia Jetson software platform, which is designed to admit multiple AI tools at the same time.
The package is integrated with Nvidia Isaac for simulation, Metropolis for Vision AI and Holoscan for real -time sensor processing.
This provision is intended to allow a module system to admit many models and workflows of AI writers, instead of requiring several separate chips.
The developer kit is now available at $ 3,499 and the production systems, including transport tables, will be distributed worldwide through their partners.
