Google’s DeepMind and Google Research have developed WeatherNext, an artificial intelligence model designed to predict cyclone trajectories and intensities with unprecedented precision. According to research published in Nature, the model provides an average of one additional day of lead time compared to traditional forecasting methods; specifically, its three-day projections demonstrate accuracy levels equivalent to previous models' two-day forecasts. This advancement was demonstrated during the October 2025 Hurricane Melissa event, where WeatherNext predicted with 80 percent confidence that the storm would strike Jamaica as a Category 5 hurricane five days before landfall. To overcome the scarcity of historical cyclone data—which typically hinders machine learning for rare extreme events—researchers trained the model using extensive general weather datasets alongside cyclone-specific information. This dual-training approach enables the model to master both broad meteorological patterns and the specific dynamics of cyclonic systems. For emergency management, this temporal advantage facilitates critical time-sensitive operations, such as organizing evacuations and staging relief supplies, potentially saving lives by extending the window for preparedness. The researchers note that achieving this level of predictive foresight would typically require a decade of conventional scientific progress.