Nvidia has introduced the Jetson AGX Thor, a $3,499 robotics AI computer designed to accelerate real-time reasoning in robots and AI-powered physical systems. 

The launch follows earlier remarks by CEO Jensen Huang about creating a “new brain” for intelligent machines. Production modules, available in bulk orders, will be priced from $2,999 each, the company said in a blog post.

Built on Nvidia’s Blackwell architecture, Jetson Thor delivers 2,070 FP4 teraflops of compute power. According to Nvidia, the platform offers 7.5 times more AI performance, 3.1 times greater CPU capacity, and double the memory of its predecessor, Jetson Orin. 

These improvements are intended to allow robots to process data from multiple sensors with minimal delay, enabling advanced applications such as visual reasoning and multimodal AI in fast-changing environments.

Early adopters include leading robotics firms Agility Robotics and Boston Dynamics, which plan to integrate Jetson Thor into upcoming humanoid robot models. 

Academic institutions such as Stanford University, Carnegie Mellon University, and the University of Zurich are also testing the platform for projects in navigation, perception, and planning.

Jetson Thor supports a wide range of generative AI frameworks and models, including Llama, Gemini, and Cosmos Reason, alongside robotics-focused platforms such as Nvidia Isaac. It is intended for developers building applications across sectors, from humanoid robots and industrial automation to surgical robotics and smart agriculture.

Nvidia confirmed the hardware is now available through its distribution partners. The company also announced that its DRIVE AGX Thor developer kit for autonomous vehicles is open for preorder, with deliveries expected to begin in September.

Source: Nvidia official blog

Total views: 646