On September 3, 2026, Google DeepMind and Google Research jointly launched the high-precision global meteorological model WeatherNext 3. This model is based on real-time geostationary satellite data and sparse meteorological station observations, with updates every hour. Its spatial resolution covers variables from 5 kilometers above the ground to 25 kilometers in the atmosphere, providing high-resolution forecasts of wind speed, cloud cover, and solar radiation, and supporting accurate estimation of wind and solar power generation. The model is now integrated into Google Search, Gemini, maps, and Earth Engine, aiming to enhance local forecasting capabilities in regions with limited computing resources, such as Latin America, Africa, and Asia-Pacific. Additionally, the model aims to address the issue of insufficient frequency of traditional meteorological data through frequent updates, thereby optimizing supply-demand balance and trading decisions in the electricity market. Currently, this project is still in the early development stage, and no specific technical details or commercialization timeline have been announced.
The WeatherNext 3 model was released on 2026-09-03. It supports wind and solar energy estimation but is still in the early development stage.
Coverage · reports per dayLANGUAGE SPLIT
Coverage mixSOURCE MIX
Chinese media 1English media 1
Entity relations
Integrated timelineUNIFIED TIMELINE
2026-09-03
Release of WeatherNext 3 model
Google DeepMind and Google Research jointly launched WeatherNext 3, which generates global forecasts every hour based on satellite and weather station data, and supports wind and solar energy estimation.
On September 3, 2026, Google DeepMind and Google Research jointly released WeatherNext 3, a high-precision global meteorological model. Based on real-time geostationary satellite data and sparse weather station observations, this model generates forecasts every hour. Its spatial resolution ranges from 5 kilometers for surface variables to 25 kilometers for atmospheric variables, providing high-resolution predictions of wind speed, cloud cover, and solar radiation. It supports accurate estimation of wind and solar power generation. Starting today, the model is available in products such as Google Search, Gemini, maps, and Earth Engine, aimed at enhancing local forecasting capabilities in regions with limited computing resources, such as Latin America, Africa, and Asia-Pacific.
DeepMind has launched a new artificial intelligence weather model that provides hourly forecasts for the electricity market. This model aims to address the issue of insufficient update frequency of traditional meteorological data, thereby supporting more accurate decisions regarding electricity supply and demand balance. Currently, this project is still in the early research stage, and no specific technical details or commercialization timeline have been announced.