Horizon Energy: AI Foundation Models for Grid Data

Developing, training and testing AI foundation models for the energy sector, built on large shared datasets rather than one company's own data. Funded work covers two linked pieces: a data-sharing and data-governance strategy that makes large-scale, secure data pooling possible across energy actors (data space connectors, federated training on confidential data, synthetic data to fill gaps, or another approach such as an 'AI gym'), and a foundation model trained on that data, addressing at least one — not necessarily all — of these energy use cases: grid planning and operation (including static power-flow and dynamic EMT modelling), forecasting, congestion management, anomaly detection, fault diagnosis, predictive maintenance, flexibility management, demand-side energy efficiency, or smart bidirectional EV charging. Models have to be tested and validated in both lab demonstrators and real-life pilots, reaching TRL 7-8 by the end of the project.