8 x AI performance enables AI inference with low latency
NVIDIA Jetson AGX Orin module offers up to 275 TOPS, 8 times more than NVIDIA^® Jetson AGX Xavier™ 32GB module. Built in 2048 NVIDIA^® CUDA^® cores and 64 Tensor Cores, NVIDIA Jetson AGX Orin 64GB module enables server-class AI inference at the edge with low latency.

Comprehensive M.2 expansion, PSE and LAN ports
AIE-KX11/12/21/22 is a system with high expansion ability, built in 1x M.2 B-Key for LTE/5G, 1x M.2 E-Key for Wifi/BT and 1x M.2 M-Key for storage. For network, on-board 2 or 4x PSE ports support 2 or 4x POE cameras, and 1x10GbE port offers data speeds up to 10 billion bits per second, 10 times faster than traditional GbE standard.

Wide input power and operating temperature for various embedded applications 
To meet various embedded applications, especially challenging environments in smart factory, smart city, and smart transportation, AIE-KX11/12/21/22 supports operating temperature ranging from -25°C to 70°C and wide input power from 9 to 36VDC.