Google Research and DeepMind have unveiled WeatherNext 3, an innovative AI-driven weather model that reshapes traditional forecasting methods. Unlike its predecessors, it bypasses conventional physics simulations and instead learns directly from up-to-date satellite data. This advanced model delivers hourly forecasts with a remarkable five-kilometer resolution, providing five times the detail of earlier systems. This heightened accuracy is particularly beneficial for regions like Africa, Latin America, and the Asia-Pacific, which have historically faced challenges in weather prediction.

WeatherNext 3 excels in capturing local weather phenomena and topographical features, significantly enhancing precipitation forecasts by as much as 50 percent. It also contributes to renewable energy planning by accurately predicting wind speeds and solar irradiance, essential for balancing energy supply and demand. Integrated into Google Search, Maps, and the Gemini app, WeatherNext 3 revolutionizes how users interact with weather data, ensuring more reliable forecasts for daily planning and severe weather alerts while encouraging users to refer to national weather services for critical warnings.