Organizations across Europe are using NVIDIA’s open-source AI models to generate high-resolution, real-time weather and climate predictions for specific industries.
Key points
- NVIDIA Earth-2 provides open AI models for global and regional weather prediction.
- StormScope EU predicts cloud conditions across Europe at 10-minute intervals.
- Companies like AXA and TotalEnergies use these tools for risk and energy forecasting.
European organizations are deploying NVIDIA Earth-2 open models to power real-world weather and climate prediction. These tools allow developers to customize forecasts with local data, improving accuracy for energy, insurance, and transport sectors.
What happened
NVIDIA released Earth-2, a family of open AI models and tools for weather and climate prediction. The platform supports global forecasts, detailed regional predictions, and environmental research using NVIDIA computing infrastructure.
StormScope EU is the latest model added to Earth-2. It is a regional version of the StormScope model, trained on European weather and climate data to provide short-term forecasting capabilities.
The model uses EUMETSAT satellite observations to predict evolving cloud conditions. It covers 16 Meteosat channels and updates predictions every 10 minutes, offering high-frequency data for rapid decision-making.
At GTC Berlin, NVIDIA showcased a live StormScope EU pipeline in partnership with Earthmover. The demo uses Earth2Studio to run forecasts and Earthmover’s platform to serve real-time results as map tiles.
Developers can combine Earth-2 models for regional needs. They use StormScope for short-term forecasting and CorrDiff to add local detail to coarser weather data, tailoring outputs to specific environmental applications.
AXA uses Earth-2’s FourCastNet 3 to study regional cyclone risk. The company simulates over 1,000 seasons to generate insights for risk assessments and prevention efforts beyond historical records.
Vaisala Xweather combines Earth-2 with proprietary observations and lightning data. This enhances its weather intelligence platforms for government, energy, and maritime customers, advancing short-term forecasting lead times.
University of Manchester researchers adapted the CorrDiff model for U.K. air-quality modeling. They trained it on the Isambard-AI supercomputer and ran it locally on NVIDIA DGX Spark systems.
TotalEnergies explores Earth-2 for solar-energy forecasting. The company evaluates StormScope-based solar-radiation nowcasts at photovoltaic sites to anticipate changes in solar-power generation at subhourly intervals.
Space42 developed SARDiff4Height using CorrDiff and NVIDIA PhysicsNeMo. This research estimates surface heights from single radar images and produces uncertainty maps for analyst scrutiny.
Why it matters
Weather AI shifts forecasting from broad global patterns to hyper-local predictions. Traditional models often lack the resolution needed for specific industrial decisions. Earth-2 allows organizations to train models on their own data, blending satellite inputs with proprietary observations. This customization enables sectors like insurance and energy to predict risks and outputs with greater precision. Teams can now run complex environmental simulations on local hardware, reducing reliance on centralized supercomputing resources for routine tasks.
The fix
No security vulnerability or patch is involved in this release. This is a product deployment announcement for AI-driven weather modeling tools.
Why it matters
Organizations adopting AI for critical infrastructure should verify data provenance and model accuracy. Teams usually validate AI outputs against historical records before operational use. Check whether local data integration meets regulatory standards for environmental reporting. Ensure that proprietary data used for training remains secure within the chosen platform. Monitor for updates to model architectures that may improve prediction reliability.
What to do and how to stay safe: NVIDIA
- Verify that AI weather models are trained on data relevant to your specific geographic region.
- Ensure proprietary observational data remains secure when integrated into external AI platforms.
- Validate AI-generated forecasts against traditional meteorological data before operational use.
- Monitor vendor updates for new model versions that may improve prediction accuracy.
Step-by-step guide: What is password policies and how they protect systems
General security guidance from the Crypt Weekly newsroom. It is not confirmed advice from the organisations named in this story.
Frequently asked questions
What is NVIDIA Earth-2?
NVIDIA Earth-2 is a family of open AI models and tools for weather and climate prediction, supporting global forecasts and regional research.
How often does StormScope EU update its predictions?
StormScope EU updates its predictions every 10 minutes using EUMETSAT satellite observations to track evolving cloud conditions.
Which companies are using Earth-2 models?
Organizations including AXA, Vaisala Xweather, University of Manchester, TotalEnergies, and Space42 are using Earth-2 capabilities.
