Google.org funds LF Energy with $2.5M for AI tools

Open source AI projects already used by RTE, TenneT, and Dutch operators are set to scale up thanks to a $2.5 million grant awarded to LF Energy. Beyond the code, the funding covers data, modeling, forecasting, and training and deployment activities across Europe.

Open source building blocks already widely adopted

Several tools supported by LF Energy are already in production. RTE in France and TenneT in the Netherlands are utilizing PowSyBl, and TenneT has built a solution based on this framework in just five months, achieving a tenfold performance increase. Power Grid Model, a calculation engine for distribution, has surpassed 10 million downloads and is used and developed by Alliander, Enexis, and Stedin. On the data side, OpenSynth aggregates 10 million synthetic domestic load profiles generated by AI and incorporates the actual topology of the French transmission network published by RTE.

The $2.5M grant aims for ready-to-use AI

Google.org is dedicating $2.5 million to LF Energy to stimulate the creation and dissemination of open source tools for European networks. This funding aims to strengthen existing AI solutions rather than starting from scratch, with the goal of meeting local needs and facilitating adoption by operators and regions with fewer resources. It targets AI-ready network data, faster modeling, and more accurate forecasting.

System context: doing more with existing networks

European electrical networks are under increased pressure from renewables, electrified heating, transportation, and data centers, beyond their initial design. New lines are not arriving quickly enough. In the short term, operators must maximize existing capacity, which requires reliable observation, modeling, and forecasting technologies.

Four flagship projects targeting data, computation, and forecasting

OpenSynth provides realistic data for researchers and planners without exposing customer information, particularly to study the impact of electric vehicles, heat pumps, and solar energy on local networks. PowSyBl, a modular framework for transmission network analysis, helps manage congestion and plan reinforcements and new renewable connections. Power Grid Model runs thousands of scenarios simultaneously to anticipate the risks of overload in distribution networks. OpenSTEF forecasts load and production from a few minutes to several days ahead, offering lead time to manage congestion and utilize flexible loads like electric vehicle charging; Alliander uses it for congestion management, and Sigholm for urban heating planning.

Stakeholder messages and deployment support

Alex Thornton, Executive Director of LF Energy, emphasizes that AI is only useful if it is based on solid data and models, noting that Google.org's support through INCO specifically focuses on these foundations, with benefits disseminated to all through open source. Adam Elman, Director of Sustainability for EMEA at Google, reminds us that network resilience also relies on open digital tools and mentions that the $2.5 million funding to Linux Foundation Europe is intended to help LF Energy scale AI data, modeling, and forecasting for operators of all sizes. The support extends beyond technical aspects: guides, implementation resources, training, workshops co-located with the LF Energy Summit Europe, and translations into several EU languages are planned.