I-Pi OSM i.MX93 — Compact Edge AI Powerhouse with eIQ Integration

The I-Pi OSM i.MX93 brings next-generation edge capability into an ultra-compact OSM form factor. Designed for developers building AI-driven products, it combines NXP’s i.MX93 dual-core Arm® Cortex-A55 + Ethos-U65 NPU with ADLINK’s production-ready Yocto BSP—now enhanced with NXP eIQ™ AI Toolkit support out of the box.

This platform gives customers a fast, low-risk path to build smart, connected, and power-efficient devices across industrial, medical, retail, and IoT markets.

Edge AI Made Practical

With integrated eIQ ML support, customers can go from model development to deployment without friction.

Built-in eIQ capabilities include:

  • TensorFlow Lite, Arm NN, and ONNX Runtime
  • Model quantization, optimization, and benchmarking tools
  • Deployment-ready ML pipelines
  • Pre-tested eIQ demo application in ADLINK Yocto image
    (per: ADLINK’s ML Demo procedure)

This dramatically shortens the “prototype → PoC → product” cycle by eliminating manual integration and toolchain setup.

Key Hardware Advantages

• i.MX93 CPU + NPU:
Dual Cortex-A55 (up to 1.7 GHz) with Ethos-U65 microNPU for real-time ML acceleration.

• Ultra-compact OSM Size-S module:
Reliable, solder-down design ideal for high-volume products.

• Rich Connectivity:
USB-C OTG, Ethernet, Wi-Fi/BT, MIPI-CSI/DSI, I²C, SPI, UART—ready for sensors, cameras, and displays.

• Industrial-grade reliability:
Robust thermal performance and long lifecycle availability.

• ADLINK-quality Yocto BSP:
Stable, production-ready, and maintained by ADLINK’s embedded Linux team.

Fast ML Deployment Workflow

https://www.nxp.com//videos/poster/TIP-EIQ-SOFTWARE-DEVELOPMENT-ENVIROMENT-INTRO.jpg

  1. Develop & Train ML Models
    Build classification, detection, anomaly, or semantic models using the eIQ Toolkit.
  2. Optimize & Quantize
    Convert models to low-footprint formats suitable for the NPU.
  3. Package for Yocto
    Use ADLINK’s pre-configured Yocto layer with eIQ already enabled.
  4. Deploy to I-Pi OSM i.MX93
    Run your optimized ML models directly on the NPU for low-latency performance.

Try Now eIQ® Toolkit on ADLINK OSM-i.MX93

If you are curious about what ADLINK NXP has to offer, you can try the demos on our development right now by visiting our ADLINK NXP repository wiki page.

Prev