DeepMind's WeatherNext AI Delivers 24 Extra Hours of Cyclone Warning
Google DeepMind published research in Nature on August 6, 2026, introducing WeatherNext Cyclones, an AI model that predicts a tropical cyclone's track, intensity, and wind structure within a single unified system. Trained on nearly 20 terabytes of atmospheric data and records of close to 5,000 historical storms, the model can generate up to 1,000 forecast scenarios in under a minute on a single processor. Benchmarks against 2023–2025 storms showed it provided roughly 24 hours more reliable warning than leading operational systems, with 28% better track accuracy and 6% better intensity accuracy when blended with official forecasts. The model was used operationally by the U.S. National Hurricane Center during the 2025 Atlantic hurricane season to help forecast Hurricane Melissa's rapid intensification and Jamaica landfall. DeepMind has open-sourced the code and pretrained weights, making the system freely available to researchers worldwide.
This is an AI-generated summary. ShortSingh links to the original source for the complete article.
Discussion (0)
Log in to join the discussion and vote.
Log in