BIOSTAR’s presence at CES 2026 is joined by long-time partner MemryX to highlight various systems and solutions, such as the EdgeComp MU-N150 industrial system, now integrated with the MemryX MX3 M.2 AI accelerator.
The EdgeComp MU-N150 itself is built around Intel’s Twin Lake N150 processor and is designed to be as compact and reliable as possible for industrial environments. It comes with dual 2.5 GbE LAN ports, dual HDMI outputs, USB 3.2 connectivity, and an M.2 Key-M 2280 slot, all packed into a fanless chassis that keeps thermals and power draw in check. As such, it is a natural fit for automation control systems, HMI terminals, smart retail installations, and digital signage, where reliability and low maintenance matter more than flashy hardware.
On the other hand, the MemryX MX3 M.2 module is the key to efficient AI inference, delivering high throughput with low latency while keeping power consumption low. By integrating all memory onboard, the MX3 also avoids common system bottlenecks, which is crucial for real-time workloads at the edge. On the software side, MemryX keeps things flexible, with support for common AI frameworks like TensorFlow, PyTorch, and ONNX, along with tools such as a compiler and performance simulator that work across Linux, Windows, and Android.
Put together, the BIOSTAR EdgeComp MU-N150 and MemryX MX3 combination is positioned as a dependable edge AI platform that can handle tasks like machine vision, robotics, safety monitoring, and real-time analytics without the overhead of cloud dependency or high power requirements.
