With two NVIDIA deep learning accelerators (NVDLA), along with a 32GB 256-bit system memory, the BOXER-8260AI will provide the device for vision-based AI applications. Moreover, its I/O options include 12 RJ-45 slots for PoE, along with DB-9 slots for CANbus and six DIO.

The BOXER-8261AI is equipped with 64 GB 256-bit system memory and 64 GB eMMC storage. The BOXER-8261AI also offers 2 Giga LAN RJ45 ports, alongside 5 USB slots and a micro-SD card slot for additional storage. The AI performance capabilities of the Jetson AGX Orin, based on the NVIDIA Ampere GPU architecture, are 8x that of the Jetson AGX Xavier, giving users unprecedented access to more powerful and efficient AI inferencing capabilities, along with increased processing speed. Further boosting the Jetson AGX Orin is the architecture’s up to 2,048 CUDA cores and 64 third-generation Tensor Cores, in comparison to its predecessor’s Volta architecture, which had 512 CUDA and 64 Tensor Cores.

As is already apparent from the BOXER-8260AI and BOXER-8261’s potential applications, AAEON customers will benefit tremendously from NVIDIA’s innovation, with AAEON’s ODM and OEM expertise utilizing the adroitness of the Jetson AGX Orin in facilitating AI application deployment.