Horizon Europe 2026: data sharing for AI foundation models in the energy sector

Data sharing to support the training and development of AI foundation models in the energy sector is a 2026 topic of Horizon Europe, the EU research and innovation program. It is part of the Cluster 5 Energy call and funds Innovation Actions. The total budget is €30,000,000, and the Commission expects to fund about three projects of €10,000,000 each.

Projects develop, train and test AI foundation models for the energy system using large datasets. Each project must cover at least one use case, such as grid planning and operation, forecasting, congestion management, anomaly detection, fault diagnosis, predictive maintenance, flexibility management, demand side energy efficiency or smart charging of electric vehicles. Projects also test new ways to share data safely among energy actors, for example data space connectors, federated training or synthetic data.

Consortia are expected to bring together data owners such as utilities and grid operators, application developers such as startups and SMEs, and model deployers such as industry and equipment makers. Proposals must show access to large relevant datasets at the time of submission. Eligible countries are set out in Annex B of the Horizon Europe General Annexes.

Models must be tested in lab demonstrators and real life pilots, building on earlier EU projects on AI testing and experimentation facilities. Activities should reach technology readiness level 7 to 8 by the end of the project. The Commission asks for open source development wherever feasible, results that stay findable and reusable after the project, and models that are transparent, explainable and free of bias.

Applications are submitted through the EU Funding and Tenders Portal, where the submission session opened on 4 August 2026. The deadline is 1 December 2026. Funded projects are expected to join the BRIDGE initiative of EU smart grid and energy projects.

Quick answers

Do the models have to be open source?

The topic asks for open source AI foundation models wherever possible and says open source development practices should be pursued wherever feasible. Results should follow the FAIR principles beyond the end of the project.

Can projects use EU computing capacity?

Yes. Projects can use the computing capacity of the EU AI Factories, in particular the three AI Factories that aim to develop AI applications for the energy sector.