Europe Deploys NVIDIA Earth-2 Open Models for Weather and Climate Prediction

NVIDIA is expanding the use of its Earth-2 open weather and climate AI models across Europe, introducing StormScope EU as the latest addition to the platform. The regional model is designed for short-term forecasting and has been trained on European weather and climate data, using EUMETSAT satellite observations to predict evolving cloud conditions across 16 Meteosat channels at 10-minute intervals. The capability is central to a live demonstration at GTC Berlin, where NVIDIA is showcasing a StormScope EU pipeline built with Earthmover. The interactive weather demo uses Earth2Studio to generate forecasts and Earthmover’s platform to serve real-time results as map tiles, allowing users to explore updates as new satellite observations arrive without downloading full datasets. The broader Earth-2 framework gives researchers and developers access to open models and tools they can customize with their own data. Teams can combine models such as StormScope for nowcasting and CorrDiff for downscaling coarse weather information to local detail. The open source NVIDIA PhysicsNeMo framework also supports training and adaptation, enabling organizations to integrate proprietary data and domain expertise into regional environmental applications, including air quality and climate-related use cases. Several organizations are applying Earth-2 to real-world decisions. AXA is using FourCastNet 3 to study regional cyclone risk beyond the historical record by simulating more than 1,000 seasons, supporting risk assessment and prevention planning. Vaisala Xweather is combining Earth-2 with its own observations, lightning data and forecasting expertise to improve weather intelligence for government, energy, maritime, transportation and insurance customers. Its work includes adapting StormScope with lightning data to extend short-term warning lead times and deliver insights through software-as-a-service products and AI-assistant integrations. University of Manchester researchers have adapted CorrDiff for U.K. air-quality modeling, training the model on the NVIDIA-powered Isambard-AI supercomputer and running it locally on DGX Spark systems, bringing high-resolution environmental modeling from national infrastructure to desktop systems. TotalEnergies is evaluating StormScope-based solar-radiation nowcasts at two photovoltaic sites and developing subhourly weather forecasting to help anticipate changes in solar-power generation. Space42 has developed SARDiff4Height using CorrDiff and PhysicsNeMo to estimate surface heights from a single radar image and generate uncertainty maps for analyst review. Collectively, these deployments show Earth-2 moving beyond global weather prediction into localized, operational applications for energy, insurance, transport, public safety and environmental research.