Reasons behind this project
The rapid transformation of power systems driven by renewable energy integration, electrification of demand, and the growing presence of distributed energy resources is increasing the complexity of grid operations. System operators must manage highly dynamic systems with millions of connected assets, bidirectional power flows, and increasing operational uncertainty.
At the same time, modern grid operations generate massive volumes of heterogeneous data, including time-series measurements, network topology data, market signals, weather forecasts, and operational reports. Traditional analytical tools and task-specific machine learning models struggle to exploit this data efficiently and often remain limited to isolated operational tasks.
AI.grids was initiated to explore a new paradigm: the development of a foundation AI model for power system operations. Inspired by advances in large-scale AI models in other domains, the initiative aims to develop a shared model capable of learning from diverse grid datasets and supporting multiple operational applications, ranging from forecasting and congestion management to planning and asset monitoring.
By leveraging European initiatives such as the Common European Energy Data Space (CEEDS), Energy AI Testing and Experimentation Facilities (TEFs), and EuroHPC AI Factories, AI.grids aims to create a scalable and collaborative framework for AI innovation in the energy sector.
Objectives
AI.grids aims to build the first Pan-European AI foundation model for power grid operations through a collaborative ecosystem of system operators, research organisations, and technology providers.
The main objectives of the initiative are:
• Develop a foundation AI model capable of supporting multiple power system operational tasks across transmission and distribution networks.
• Leverage diverse European grid datasets to train AI models that capture structural, temporal, and operational characteristics of power systems.
• Enable secure and interoperable data sharing frameworks aligned with the Common European Energy Data Space.
• Integrate high-performance computing resources from European AI Factories to support large-scale model training.
• Validate AI capabilities on real operational use cases, including forecasting, congestion management, asset monitoring, planning support, and flexibility services.
• Create an open collaborative ecosystem where system operators, research institutions, and industry partners jointly contribute to the development and validation of AI solutions for grid operations.
Project partners
The project team is currently being defined and the project is open to additional partners.