Developed by WiSDON Lab at National Yang Ming Chiao Tung University (NYCU), the AI-native Network Energy Saving (NES) rApp optimizes 5G RAN energy use by intelligently controlling cell on/off states based on users’ RSRP measurements, collected every 10 seconds. Designed for indoor private 5G networks, the solution has been validated across four scenarios with varying traffic loads and mobility patterns. Leveraging AI/ML, it predicts the impact on throughput and handover performance before making adjustments, ensuring QoS and mobility robustness are maintained. Under these scenarios, the NES rApp achieved up to 28% energy savings with less than 2% degradation in downlink throughput.