Bull’s R&D Spain team  will play a key role in the European Agentic6G project, heading both the task on Agentic AI Multi-Agent System (MAS) Architecture and API Definition and Work Package 3, focused on the Agentic6G Framework and agents for software generation, orchestration, and self-optimisation. The team will also lead the development of the AgenticOps framework, designed to support multiple LLM-based agentic AI agents.

The project is part of the Smart Networks and Services Joint Undertaking (SNS JU) portfolio and will run from May 2026 to April 2029. It brings together a consortium of nine partners, including Orange Romania, the University of Murcia, the Polytechnic University of Valencia, the University of Leicester, Aston University, One Source, eBOS, Bull, and Ericsson.

Agentic6G aims to advance the integration of AI within future 6G systems through an Agentic-AI approach based on Multi-Agent Systems (MAS). It will enhance the interplay between network and computing resources across the Device–Edge–Cloud continuum, potentially leveraging new 6G architectural advancements alongside extreme communication capabilities beyond 5G.

The project will enable autonomous service composition, allowing MAS to compose, coordinate, and optimise B5G/6G and vertical microservices. This will be achieved through the combination of collective agentic AI knowledge, Large Language Models (LLMs), shared memory, and tools.

In addition, Agentic6G will introduce end-to-end Agentic Operations (AgenticOps) to support lifecycle management, along with self-organising capabilities such as spontaneous agent generation, self-protection, self-healing, and auto-scaling.

The project will deliver a set of Agentic6G agents as autonomy enablers and an integrated MAS platform, to be tested and validated through representative technical use cases, including a complex swarm robotics application.