In this keynote, I will present a unified BDIx-driven distributed AI framework that brings autonomous, explainable intelligence to the very edge of 5G/6G networks and nano-grid energy systems. Building on our work in the ADROIT-6G project, I will demonstrate how BDIx agents—combining beliefs, desires, intentions, and explicit “why” explanations—can orchestrate device-to-device connectivity, spectrum, and power resources in real-time, under uncertain and highly dynamic conditions. By embedding these agents into edge and near-edge nodes, we move from static configurations and rigid policies to a living, adaptive control layer that can learn, reason, and justify its actions while respecting latency, reliability, and energy constraints.

The talk then extends this intelligence from communication networks into nano-grid and micro-grid energy management, treating each nano-grid as a “prosumer” agent that must balance local renewables, storage, EV charging and critical loads. I will explain how ANFIS-based plan libraries, linear programming, and swarm-based optimization (PSO) work together with BDIx to co-optimize power flows, costs, and emissions across multiple interacting nano-grids. Using results from detailed simulations and real-world use cases, I will highlight the practical benefits—reduced energy bills, more resilient grids, and greener operations—as well as the broader vision: a unified, AI-driven regional innovation ecosystem where 6G connectivity and sustainable energy infrastructures co-evolve and reinforce each other.