Large scale operational validation and upscaling of state-of-the-art (Generative) AI tools and models powering a next generation digital energy system

From the Commission's own budget table for this call. Confirm against the topic page before you plan around it: these move when a work programme is amended.

Expected Outcome: Project results are expected to contribute to all the following expected outcomes: Advanced, secure frugal (energy-, resource-, data- and cost-efficient, sustainable) AI tools for real-time grid operations (including protection), market operations (if relevant) and consumer empowerment in distributed energy systems. A demonstrated AI development environment integrated with the upgraded digital spine of the energy system and energy-saving consumer applications. Scaled-up capabilities for learning at edge nodes and federated systems using agent-based architectures. A governance framework and operating model for the Gen(AI)-powered digital spine. Scope: Building up on previous actions in Horizon Europe (e.g. the digital spine of the energy system in cluster 5 as well as cloud and edge continuum and AI data processing activities in cluster 4) and Digital Europe (Common Reference Framework for energy consumer applications across the EU, Common Energy Data Space), projects are expected to: Validate (demo site and operational grid) and scale secure state-of-the-art specialised frugal (energy- and resource-efficient, data- and cost-efficient, locally deployed and sustainable) open-source (Generative-)AI tools and models for real-time grid (and market, if relevant) monitoring, balancing, forecasting and consumer empowerment in a distributed energy system. Demonstrate an AI development environment within the enhanced digital spine of the energy system, including its digital twin, which should incorporate the functionality of the Common Reference Framework for energy saving applications. Prototype and validate at scale machine learning at the edge node level and federated at swarm level in an agent-based architecture, leveraging new mechanisms for scaling down AI models as well as incremental and reinforcement learning. The project's deliverables should be compatible with the principles of the Common European Energy Data Space and consistent with the EU's Apply AI and Competitiveness Compass, ensuring that they promote AI adoption and competitiveness in line with EU policies, as well as the EU Digitalisation of the energy sector initiatives. Where relevant and possible, the selected projects are expected to make use of the European AI factories and AI gigafactories. The selected projects are expected, where available, to make use of and further develop commonly agreed European standards and to actively collaborate and contribute to the work of established and de facto European standards developing organisations and open-source communities and plat…