Google DeepMind has announced a breakthrough in cyclone forecasting with its WeatherNext AI model, as detailed in a paper published in Nature on August 6, 2026. The model achieves state-of-the-art accuracy in predicting a cyclone's track, intensity, and wind structure, giving forecasters an average of one extra day of predictive accuracy compared to previous models.
This improvement is significant: the model's three-day forecasts are as accurate as what prior models could achieve for two days, representing roughly a decade of meteorological progress. The research was a collaboration between AI researchers at Google DeepMind and Google Research, along with expert forecasters from the National Hurricane Center (NHC), the Cooperative Institute for Research in the Atmosphere (CIRA), the UK Met Office, and other weather agencies worldwide.
The model has already shown real-world impact. During the 2025 hurricane season, it helped the NHC make a historic forecast for Hurricane Melissa by predicting rapid intensification and landfall in Jamaica, enabling advance warnings. This year, the team is predicting 1,000 possible scenarios per cyclone to support decision-making.
WeatherNext is a single AI model that bridges the gap between global and local modeling approaches. It was co-trained on nearly 20 terabytes of global atmospheric data and the IBTrACS database of nearly 5,000 historical storms. Using Functional Generative Networks (FGNs), it can generate a 15-day forecast in under a minute on a TPU, allowing forecasters to quickly assess tail risks.
In a move to amplify AI's impact, Google DeepMind is open sourcing the WeatherNext 2 and WeatherNext Cyclones models used during the hurricane season, aiming to empower researchers and build more resilient communities.