Google DeepMind has introduced WeatherNext Cyclones, an AI system that forecasts tropical cyclone tracks and intensity in the same model.
The Decoder reports that the system can look about one day further ahead than leading operational models, an improvement the article compares to roughly a decade of progress in traditional weather forecasting. The model was built with the National Hurricane Center, the Cooperative Institute for Research in the Atmosphere, and the UK Met Office.
The system has reportedly been running forecasts since June 2025, and DeepMind has released code and model weights on GitHub. That openness matters because weather tools need outside scrutiny before they can be trusted in high-stakes operational settings.
The limitation is interpretability. The article notes that even the developers do not fully know how the model achieves its performance using much coarser data. For forecasters, that makes WeatherNext Cyclones promising, but not a simple replacement for established physical models and expert judgment.