The NeoEdge Jetson Nano Orin Edge AI Box is the initiative to introduce compact industrial computing to the real world where it is required to execute high-performance inference and execute large models. It implements NVIDIA Jetson Orin Nano and Orin NX modules to execute LLMs, VLMs, ViTs, and other high-load models on the edge. Robotics developers can deploy this platform without fear of thermal throttling or unstable operation, machine vision developers can do the same, or smart manufacturing system develpers can do the same.
This system focuses on stable multi-protocol connectivity, reliable 24/7 operation, and strong compute density in a fanless enclosure. It supports performance levels up to 157 TOPS in SUPER mode, which allows users to run multi-modal and transformer-based AI workloads on-site. The NeoEdge design keeps everything compact while offering room for storage, wireless modules, and camera inputs.
Previously, we covered the CamThink NeoEyes 301 Edge AI Camera, which runs YOLOv8 at 25 FPS on the STM32N6 MCU. Feel free to check it out if you’re interested in related products.
NeoEdge Jetson Nano Orin Edge AI Box Specifications:
The NeoEdge Jetson Nano Orin Edge AI Box supports four NVIDIA Jetson options: Orin Nano 4GB, Orin Nano 8GB, Orin NX 8GB, and Orin NX 16GB. Performance scales from 20 to 100 TOPS in standard mode and reaches up to 157 TOPS in SUPER mode. The Orin Nano variants use 512 CUDA cores, while the Orin NX modules use 1024 CUDA cores with Tensor Core acceleration. This range covers AI workloads such as LLMs, VLMs, multi-modal models, robotics, and machine vision tasks.
NeoEdge Jetson Nano Orin Edge AI Box front ports
The system has two Gigabit Ethernet, four USB 3.1 Type-A, and one USB 3.2 Type-C high-speed peripheral ports. RS232, RS485, CAN, and eight digital I/O pins can be connected with the help of industrial connectivity. Storage is increased with 2 M.2 Key M slots, one PCIe x4 NVMe SSD and one PCIe x1. M.2 Key E is used as a 4G/5G wireless and M.2 Key B is used as a Wi-Fi/ Bluetooth wireless. There is also the 4K HDMI output, two MIPI CSI camera interfaces, mic in, audio out, numerous SMA antenna ports and a Nano SIM slot in the box.
NeoEdge Jetson Nano Orin Edge AI Box back ports
The design withstands industrial environments with a –25°C to +60°C operating range, fanless passive aluminum cooling, and 12V–36V DC power input. Power peaks at 120W during heavy workloads. The device measures 160 × 125 × 75 mm, weighs 1.9 kg, and supports desktop, wall, and DIN-rail mounting for flexible deployment across factories, edge AI gateways, inspection systems, and autonomous machines.
The NeoEdge platform runs the NVIDIA JetPack 6.0 SDK and supports CUDA, TensorRT, and DeepStream for fast AI inference pipelines. Developers can run PyTorch, TensorFlow, ONNX, VLLM, Ollama, and llama.cpp without compatibility issues. The system supports optimized deployment workflows for both vision models and large language models, making it suitable for AI gateways, robotics controllers, and industrial inspection systems.
The NeoEdge Jetson Nano Orin Edge AI Box is available in 4GB and 8GB versions, priced at $899 and $1,099. More details are available on the product page.
Niladri Chowdhury