In recent years, there has been a growing interest in utilizing Artificial Intelligence (AI) and Machine Learning (ML) techniques for anomaly detection in a wide range of industrial and commercial applications. Furthermore, machine learning on embedded devices rather than in the cloud is perceived as a new approach to Embedded AI. This work focuses on the development of an Embedded AI-based anomaly detection system for a fan using an STM32 microcontroller and an Inertial Measurement Unit (IMU) sensor to capture vibration data. It leverages the processing capabilities of the STM32 microcontroller to collect and analyze the vibration data from the IMU sensor and further deploy an ML algorithm to identify anomalies in the fan’s operation. Real-time output is provided by the system, enabling prompt feedback to users. To evaluate the performance of the system, a testbed consisting of a fan and an IMU sensor was utilized. The results indicate that the system exhibits high accuracy in detecting anomalies in the fan’s operation while maintaining low false positive rates. The proposed system holds potential for diverse industries such as manufacturing, transportation, and energy.