Skip to main content

Radxa AICore AX-M1

Product overview​

The Radxa AICore AX-M1 is a high-performance M.2 accelerator module based on the AXERA AX8850 SoC. It features high compute capability and power efficiency, designed for edge AI computing and inference applications.

The Radxa AICore AX-M1 integrates a multi-core high-performance CPU and a high-compute NPU with excellent multimedia processing capability, providing efficient and flexible hardware support for various edge AI scenarios.

Key features​

  • High-performance SoC platform — Powered by the AX8850 SoC with an octa-core Cortex-A55 CPU up to 1.5 GHz, delivering strong general-purpose compute and multi-tasking capability.

  • Powerful AI compute — Integrated high-performance NPU supporting up to 24 TOPS@INT8 neural network inference to meet high-speed inference needs of complex AI models.

  • Professional-grade multimedia processing

    • Supports up to 8K@30fps video input and image processing for high-resolution video analytics.
    • Equipped with a video processor supporting H.264/H.265 HD encoding/decoding for diverse streaming and intelligent video analytics scenarios.

  • Hardware compatibility — Supports multiple host platforms including Intel, AMD, and Rockchip.

  • OS compatibility — Compatible with multiple operating systems including Ubuntu, Debian, and CentOS.

  • Standardized interface and compatibility — Standard M.2 M Key 2280 size, compatible with mainstream industrial and embedded boards.

Product appearance​

Radxa AICore AX-M1 module (front)

Radxa AICore AX-M1 module (back)

Specifications​

CategorySpecification
CPUOcta-core Cortex-A55, up to 1.5 GHz
Memory8 GB LPDDR4x
NPU24 TOPS@INT8; supports matrix compute units and intelligent vision engine
VPUH.264/H.265 8K@30fps encode/decode and 16-channel 1080p@30fps decode support
Hardware compatibilitySupports Intel, AMD, Rockchip host platforms
OS compatibilitySupports Ubuntu, Debian, CentOS and other mainstream Linux distros
Form factorM.2 M Key
Dimensions22 mm x 80 mm
Operating voltage3.3 V
System power≤ 8 W

Application scenarios​

  • Smart cameras
  • Autonomous driving
  • Consumer electronics
  • Security monitoring
  • Edge computing
  • Smart mobility

    You need to be logged into GitHub to post a comment. If you are already logged in, please ignore this message.

    Radxa-docs © 2026 by Radxa Computer (Shenzhen) Co.,Ltd. is licensed under CC BY 4.0