Days before Hurricane Melissa made landfall in Jamaica in October 2025, forecasters at the National Hurricane Center used a new Google DeepMind AI system, WeatherNext Cyclones, to predict its rapid intensification from a Category 1 to a Category 5 storm with unprecedented lead time. NHC Director Michael Brennan said, “That’s something we have not historically been able to do with that kind of lead time.” Melissa struck as one of the most intense Atlantic hurricanes on record. Google’s study, published in the peer-reviewed journal Nature, compared WeatherNext Cyclones against the AI model GenCast and the traditional Ensemble Prediction System. At the 5-day forecast horizon, WN-C’s average track error was 230 kilometers (143 miles), compared to 370 kilometers (229 miles) for ENS and 335 kilometers (208 miles) for GenCast. On intensity, its three-day forecast was 3.75 knots more accurate than the Hurricane Analysis and Forecast System. Co-author James Franklin said, “It’s simply much more accurate than any other piece of track guidance that forecasters currently have — and by a substantial margin.” The model also beat traditional models at predicting wind radii. Google says its model runs on atmospheric data about 100 times coarser than specialized high-resolution regional models.