Fengqing, an AI global short- and medium-range ensemble weather forecasting system (V1.0), has passed an operational review and entered nationwide weather service in China, according to IT之家 and 36Kr. The system was led by the National Meteorological Center with universities and research institutes, and its full training pipeline ran on a 100,000-card AI4S cluster at the National Supercomputing Center in Tianjin; the developers call it China's first operational weather AI ensemble model with fully self-controlled data, models and compute. It combines a vision Transformer architecture with meteorological physical constraints and reaches an effective global forecast lead time of 11 days, with the developers claiming better core metrics, probabilistic accuracy and spread-error control than the European Centre for Medium-Range Weather Forecasts' (ECMWF) latest AI ensemble system. It will support disaster prevention, transport, energy scheduling and ecological protection.