Timely detection and corrective action of widespread power equipment defects is essential for high-quality power supply. Previously, manual diagnosis could process only 119 units per person per day and incurred annual costs of USD 8,700,000. The AI-powered automatic optical system for power equipment diagnosis can process up to 2,400 units per day, achieving 20 times greater efficiency compared to manpower inspection. The system, mounted on a vehicle driving at approximately 30 km/h, automatically identifies power equipment (93.9% accuracy), tracks it, captures high-resolution images, analyzes defects in real time (74.4-96.9% accuracy), and stores the results. Automated diagnosis is expected to evolve further with additional learning to improve AI accuracy. This technology helps prevent power outages by avoiding safety accidents and by efficiently and reliably identifying defective equipment.