ROHM Unveils Revolutionary AI-Enhanced Microcontroller

ROHM Semiconductor has launched its first AI-equipped microcontrollers (MCUs), the ML63Q253x-NNNxx and ML63Q255x-NNNxx, designed for fault prediction and degradation forecasting across various devices, particularly in industrial settings like motors. These MCUs can perform both AI learning and inference independently, without requiring a network connection, thus enhancing operational efficiency and reducing maintenance costs. Unlike typical AI models that depend on cloud connectivity, ROHM’s solutions utilize a three-layer neural network to enable real-time anomaly detection on-site. With low power consumption and a 32-bit Arm® Cortex®-M0+ core, these MCUs are suitable for applications in industrial equipment, home appliances, and more. The lineup includes 16 models with various memory sizes and package specifications, and mass production of eight models began in February 2025. To support developers, ROHM offers tools for AI evaluation and has established an ecosystem for model development and integration.

Source link