The NVIDIA Jetson AGX Thor Developer Kit delivers exceptional performance and scalability. Powered by an NVIDIA Blackwell GPU and 128GB of video memory, it delivers up to 2070 FP4 TFLOPS of AI compute performance, easily running the latest generative AI models while consuming less than 130W. Compared to the NVIDIA Jetson AGX Orin, it delivers up to 7.5 times the AI ​​compute performance and 3.5 times the energy efficiency. It utilizes the latest Blackwell Multi-Instance GPU (MIG) technology and a powerful 14-core Arm Neoverse-V3AE CPU to accelerate low-latency, real-time applications. It also includes a suite of accelerators, including a third-generation Programmable Vision Accelerator (PVA), dual encoders, decoders, and an optical flow accelerator. For high-speed sensor fusion, the developer kit offers a rich set of I/O options, including a QSFP slot with 4 x 25GbE, an RJ45 connector for wired multi-GbE, multiple USB ports, and other interfaces. Furthermore, it is designed for seamless integration with existing humanoid robotics platforms and can be easily connected to quickly start prototyping.

The NVIDIA Jetson Thor series modules provide an excellent platform for physical AI and robotics development, delivering up to 2070 FP4 TFLOPS of AI computing performance and 128GB of video memory, with power configurable between 40W and 130W. Compared to the NVIDIA Jetson AGX Orin, the Jetson Thor series modules deliver over 7.5 times the AI ​​computing performance and 3.5 times the energy efficiency.

Advantages

Blackwell architecture GPU, sensor processing, and robotics AI software stack

Workload

Unlocking physical AI with the power of Jetson Thor