Danish Meteorological Institute Teams Up with Gefion Supercomputer for Precision Weather Forecasting
Copenhagen, Tuesday, 18 March 2025.
DMI’s partnership with Gefion aims to innovate AI-based weather models for faster, more precise forecasts, enhancing national response to extreme weather events through significant computing efficiency gains.
Revolutionary AI Implementation
In a groundbreaking development announced on March 17, 2025, the Danish Meteorological Institute (DMI) has partnered with the Danish Centre for AI Innovation A/S (DCAI) to revolutionize weather forecasting through advanced AI models [1]. The collaboration’s initial success has been remarkable, with DMI’s specialized team developing a regional AI weather model that matches the accuracy of traditional international models in predicting temperature and wind conditions, while dramatically reducing computational time from hours to minutes [1][3].
Supercomputing Power and Efficiency
The Gefion supercomputer, currently ranked 21st among the world’s most powerful supercomputers, serves as the backbone of this innovative initiative [1]. This achievement is particularly significant given DMI’s existing robust computational infrastructure, which includes a supercomputer in Iceland operated in collaboration with UWC-West and access to the European weather center ECMWF’s computing resources [1]. The new AI model’s efficiency represents a paradigm shift in weather forecasting technology, requiring only a fraction of the traditional computing power while maintaining competitive accuracy [1].
Future Applications and Expansion
DMI meteorologist Kasper Hintz highlights the institute’s ambitious plans for expanding AI applications: ‘We have several AI models in our pipeline for predicting precipitation and cloud cover, which will directly benefit from the experience we have gained in just a few months’ [1][3]. The success of the pilot project has led to a formal three-year collaborative agreement between DCAI and DMI [3], focusing on developing more sophisticated weather models that integrate additional variables and uncertainty estimation.
National Security and Emergency Response
The implementation of these advanced AI models is expected to significantly enhance national preparedness against extreme weather events [1]. By enabling more frequent and rapid weather forecasts, the system will provide emergency services and authorities with crucial additional response time during severe weather conditions, potentially saving critical infrastructure and lives [1]. This technological advancement positions Denmark at the forefront of AI-driven meteorological innovation, as emphasized by DCAI CEO Nadia Carlsten in recent statements [1][3].