Google DeepMind's WeatherNext Cyclones AI model delivered an extra day of reliable warning lead time on tropical cyclone track, intensity, and wind structure, with proof published in a peer-reviewed Nature study alongside a free, open-source release. During the 2025 Atlantic hurricane season, the National Hurricane Center ran WeatherNext alongside operational guidance and used it to anticipate Hurricane Melissa's rapid intensification and catastrophic Jamaica landfall. Melissa made landfall near New Hope, Westmoreland Parish on October 28, 2025, as a Category 5 storm with 185 mph (298 km/h) sustained winds, killing 95 people and causing approximately $8.8 billion (USD) in physical damage, the costliest storm in Jamaican history. Its maximum winds nearly doubled from 70 mph (113 km/h) to 130 mph (209 km/h) in 24 hours. Tropical cyclones have caused more than 700,000 deaths and $1.4 trillion (USD) in economic losses over the past 50 years. Evaluated on cyclones from 2023 through 2025, WeatherNext's predictions offer an average of a day or more of lead time advantage over leading operational models; its three-day forecast matches what prior models could provide at two days, giving emergency managers a reliable 72-hour window instead of 48 hours.