Nvidia has launched Jetson Thor, a powerful new robotics computer that promises to act as the brain of the next generation of the AI-powered robot. The chip, which is now usually available, gives a larger jump in performance compared to its predecessor, Jetson Oin, and is designed to run real -time robotic applications, which require rapid processing of data from several sensors. According to Nvidia, Jetson Thor provides 7.5 times more AI calculation, doubling the memory of CPU performance and Jetson Oin. This means that developers will be able to process high-speed sensor data and run directly advanced visual regioning tasks on the device, a shift that can perform the robot in a dynamic environment.

Forest Robotics, the company behind the humanoid robot digit, has already integrated the Jetson model in its system and will adopt Jetson Thor for the sixth generation digit. Paigi Johnson, CEO of Chaplata Robotics Chief Executive Officer said: “The powerful edge processing given by Jetson Thor will take points to the next level, increase their real -time accountability and expand their abilities to a broad, more complex skills.”

Points have already been deployed commercially in warehouses and factories, where it helps in tasks such as stacking, loading and palating. With Jetson Thor, agility hopes that the robot will take even more complex responsibilities.

Boston Dynamics, which is known to develop some of the most advanced humanoid robots, planned to integrate Jetson Thor into its Atlas model. NVIDIA stated that it would allow Atlas to use server-level calculation power, speed up AI workload and handle high-bandwidth data on the device, reduces dependence on cloud processing.

Cases of potential use expand well beyond the humanoid robot. Jetson Thor Surgical Assistant, delivery robots, smart tractors, industrial manipulators and AI agents can accelerate the development of video analytics, which are all running large and more complex AI models in real time.

The chip has been adapted to the generative regional region model, including large transformers and vision-language-action systems, allowing robots to intelligent intelligence with the world around them. Nvidia states that Jetson Thor supports a wide range of popular AI Framework such as Cosmos Region, Dipsek, Lama, Mithun and Quven, Isaac GR 00T with robotics-specific models.

Research institutes are also starting to find out its ability. Teams at Stanford University, Carnegie Melon University and Jurich University are already using Jetson Thor for advance perception, navigation and plan models for applications such as discovery-and-rescue and medical triaies. Carnegie Mellon’s Associate Research Professor Sebastian Sheerr said that upgrading Jetson Thor will “be able to handle more fine tasks to the robot” by improving perception and decision making on the edge “.

The Jetson Thor Developer Kit is priced at $ 3,499, while the Jetson T5000 production module starts at $ 2,999 for bulk order. NVIDIA also announced that its drive Agx Thor Developer Kit for autonomous vehicles is now available for pre -order, with delivery starting in September.