Skip to main content

Fogwise® AirBox


Fogwise® AirBox is an embedded AI micro-server with arithmetic power up to 32TOPS@INT8, supports multiple precision (INT8, FP16/BF16, FP32), supports private GPT, text-to-image, and other mainstream AI model deployments, and is equipped with an aluminum alloy casing for deployment in harsh environments.


moduleFogwise® AirBox
sizes104 mm × 84 mm × 50.2 mm
processing unitSOPHON SG2300X SoC,octa-core Arm® Cortex®-A53 (ARMv8) @ 2.3GHz
tensor processing unitTensor Processing Unit, computational power: up to 32TOPS (INT8), 16TFLOPS (FP16/BF16) and 2TFLOPS (FP32)
Support for major deep learning frameworks including TensorFlow, Caffe, PyTorch, Paddle, ONNX, MXNet, Tengine and DarkNet.
memory16GB LPDDR4X
storage64GB eMMC
16MB SPI Flash
High Speed SD Card Slot Available
multimediaSupports decoding 32 channels of H.265/H.264 1080p@25fps video
Fully handles 32 channels of HD 1080p@25fps video, including decoding and AI analytics
Supports encoding 12 channels of H.265/H.264 1080p@25fps video
JPEG: 1080p@600fps, max. Support 32768 x 32768
Support video post-processing, including image CSC, resize, crop, fill, border, font, contrast and brightness adjustment
interface2 Gigabit Ethernet ports (RJ45)
1 M.2 M-key slot for NVMe SSDs.
operating temperature0°C to 40°C
housingsCorrosion-resistant aluminum alloy housing
radiatorPWM speed control fan with custom heat sink

Photograph of sth.

Fogwise® AirBox Overview

Chip block diagram

SG2300X Block Diagram

System block diagram

Fogwise® AirBox Block Diagram