ROHM Semiconductor Unveils AI-Equipped Microcontrollers for Anomaly Detection

ROHM Semiconductor has introduced a new line of AI-equipped microcontrollers (MCUs) designed to predict equipment anomalies and forecast degradation without the need for network connectivity, announced in a press release. These MCUs, identified as ML63Q253x-NNNxx and ML63Q255x-NNNxx, are the first in the industry to perform both learning and inference independently on the device.

The AI MCUs utilize ROHM's proprietary on-device AI solution, Solist-AI™, which employs a simple 3-layer neural network algorithm. This allows for real-time anomaly detection and maintenance efficiency improvements in various devices, including industrial equipment and home appliances. The MCUs are equipped with a 32-bit Arm® Cortex®-M0+ core and feature low power consumption, making them suitable for a wide range of applications.

Mass production of these MCUs began in February 2025, with eight models available in the TQFP package. ROHM has also released an AI simulation tool, Solist-AI™ Sim, to aid in evaluating the effectiveness of learning and inference before deployment. This tool supports pre-implementation validation and enhances inference accuracy, facilitating the adoption of these innovative MCUs.