Axiomtek AIE110-ONA-4GB NVIDIA Jetson Orin Nano 4GB Edge AI Developer Kit with 15 W PoE and NVMe

Entry-level vision projects often start with one or two cameras and a model that does not yet need the larger Orin Nano memory tier. The AIE110-ONA-4GB fits the NVIDIA Jetson Orin Nano 4GB module to the Axiomtek AIE110-ONA edge AI developer kit, an enclosed, palm-sized unit that sits on a bench or in a small cabinet.

The Orin Nano 4GB pairs a 512-core NVIDIA Ampere GPU with 16 Tensor Cores and a 6-core Arm Cortex-A78AE CPU, with 4 GB of 64-bit LPDDR5, and Axiomtek rates it at 20 TOPS. Axiomtek supports Super Mode on JetPack 6.2 and above, which raises the AI throughput of the same module. The AIE110-ONA-8GB doubles the GPU to 1024 cores with 32 Tensor Cores and 8 GB of 128-bit LPDDR5 at 40 TOPS, and suits models or camera counts that need more than 4 GB. The Orin Nano has no eMMC, so an M.2 Key M 2280 NVMe SSD on PCIe x4 is the boot device.

For cameras and networking, the AIE110-ONA-4GB has one GbE port from the module and a second GbE port on an Intel I210-IT controller that supplies IEEE 802.3af PoE at up to 15 W, enough to power a single IP camera without a separate injector. One USB 3.2 Gen2 port, one USB 2.0 port and an HDMI output cover peripherals and a local display. An 8-channel programmable digital I/O header, supplied with its cable, handles triggers and status signals, and a full-size Mini PCIe slot with a SIM slot takes Wi-Fi, Bluetooth, LTE or GPS modules.

The AIE110-ONA-4GB runs from 12 V DC on a screw-type jack and is rated from -10 to +60 °C, with an active heatsink, 3 Grms vibration and 50 G shock tested with an SSD fitted. A micro USB port and recovery switch on the front let JetPack be reflashed without opening the unit, and Axiomtek lists JetPack 5.1.1 or later with Ubuntu 20.04 or later. Allxon support adds remote monitoring, control and over-the-air updates once units are deployed.