From RAN Control to Agentic Intelligence: Architecture and Vision for Energy Efficient AI-RAN

The article introduces an agentic AI-native Radio Access Network (RAN) architecture designed to enhance energy efficiency in future 6G networks by integrating AI with existing Open RAN frameworks. It emphasizes the use of semantic intent abstraction and coordination driven by Large Language Models (LLMs) to enable adaptive orchestration and energy-aware multi-objective optimization across diverse applications. This approach addresses the challenge of increasing energy consumption in RANs, making it significant for practitioners aiming to develop sustainable and efficient AI-driven telecommunications infrastructure.