Google launches WeatherNext 3 AI for high-precision weather forecasting

Google has unveiled WeatherNext 3, an advanced artificial intelligence model developed by Google DeepMind and Google Research to enhance the precision of global weather forecasting. Unlike previous models that relied on processed data from supercomputers, this system utilizes real-time satellite data and observations to provide hourly updates.

The new model offers significantly higher resolution, providing data on temperature and humidity at a 5-kilometer scale, while other atmospheric variables are processed at 10-kilometer and 25-kilometer resolutions. This represents a fivefold increase in detail compared to its predecessor. Testing indicates improvements in precipitation forecasting by as much as 50 to 60 percent for day-ahead predictions.

WeatherNext 3 is being integrated into Google services including Search, Maps, Gemini, and Google Earth Engine. The technology is expected to be particularly beneficial for regions with limited meteorological infrastructure, such as parts of Africa, Asia, and Latin America. Additionally, the model provides strategic utility for the renewable energy sector by predicting wind speeds and solar radiation to assist in energy production planning.