European collaboration on artificial intelligence (AI) for weather and climate has been recognised with two nominations in this year’s HPCWire Reader’s Choice Awards, in the categories “Best AI Product or Technology” and “Best Use of AI Methods for Augmenting HPC Applications.” Voting for the shortlisted entries is open on the HPCWire website until 3 October 2025.
The nomination in the “Best AI Product or Technology” category recognises Anemoi as an open-source collaborative framework for developing AI-based weather and climate models. Anemoi is developed by ECMWF together with national meteorological services across Europe. It has already enabled several AI weather forecasting models in Europe, including the first operational Artificial Intelligence Forecasting System (AIFS) at ECMWF, the Bris regional model at MET Norway, and the AICON global model at the German National Meteorological Service (DWD).
The nomination for “Best Use of AI Methods for Augmenting HPC Applications” highlights how Anemoi provides portable, scalable workflows that have been deployed on EuroHPC supercomputers such as LUMI, MareNostrum 5, Leonardo and MeluXina to train and run AI weather models at scale. These workflows are central to the development of new AI weather models across Europe and also underpin the AI activities in the Destination Earth (DestinE) initiative of the European Commission, which expand forecasting approaches towards AI-driven Earth system modelling.
These achievements add to a growing list of recognition of the work ECMWF and partners on artificial intelligence and digital technologies in different initiatives: the EMS Technology Achievement Award 2025 for Anemoi; two HPCWire Readers’ Choice Awards in 2024 for DestinE’s Climate Adaptation Digital Twin (Climate DT); and the shortlisting of the Climate DT for the ACM Gordon Bell Prize 2025. Together they underline Europe’s leadership in combining HPC and AI for weather and climate.
Anemoi enabling AI-driven forecasting systems
Named after the Greek gods of the winds, Anemoi has become a shared European framework for AI in weather and climate. It is also an important part of the EUMETNET Artificial Intelligence (E-AI) Programme, which supports the integration of AI into operational meteorology. Contributions from ECMWF, the Royal Netherlands Meteorological Institute (KNMI), the Finnish Meteorological Institute (FMI), MeteoSwiss, Belgium’s Royal Meteorological Institute (RMI), Météo-France, the German National Meteorological Service (DWD), Met Norway, UK Met Office, Geosphere, the Danish Meteorological Institute (DMI), the Spanish State Meteorological Agency (AEMET) and others are driving collective AI innovation across forecasting systems. By making AI models and software solutions openly available through platforms such as GitHub and Hugging Face, Anemoi is democratising access to advanced forecasting tools.
The framework underpins practical applications: its workflows have already enabled the rapid development and operation of AIFS alongside ECMWF’s physics-based Integrated Forecasting System. AIFS can deliver 10-day forecasts in one minute instead of one hour, and in many cases its skill exceeds that of the physics-based model. Such achievements show that Anemoi is not about prototyping research models but about enabling real operational forecasting applications.
Advancing AI weather and climate solutions in Europe
The AI activities in DestinE are advancing AI weather and climate solutions in Europe by expanding towards an AI-driven Earth system model. New AI components are being developed for land, hydrology, ocean, waves and sea ice. At the same time, practical solutions such as Forecast-in-a-Box are lowering barriers to run and exploit AI models. These efforts build on Anemoi and early operational models such as AIFS, while leveraging the capabilities of the European High Performance Computing Joint Undertaking (EuroHPC JU).
The progress in Anemoi, AI forecasting systems and AI Earth system modelling reflects the synergy between national initiatives, ECMWF, the wider European meteorological infrastructure, DestinE, EuroHPC and AI factories, as well as the hugely collaborative effort across Europe. Together, these efforts are pushing boundaries in AI for weather and climate in Europe.
Join the discussion
These themes will be explored further at a high-level meeting in Brussels on 2 October, “AI for Weather and Climate Preparedness and Resilience: From European Vision to Real-World Impact”, organised with the support of the European Commission’s DG CNECT and in partnership with Science|Business. Registration to join in-person or online is open. You can also support these collaborative efforts by casting a vote in the HPCwire Readers’ Choice Awards.