Google’s WeatherNext 3 can now predict weather every hour
WeatherNext 3 uses live satellite data and AI to offer sharper, hourly weather forecasts across Google Search, Maps, Gemini, and developer platforms.
Google has introduced WeatherNext 3, its latest AI weather forecasting model, designed to predict local weather conditions every hour. Developed by Google DeepMind and Google Research, the model combines live satellite imagery with traditional weather information to produce forecasts that are more frequent and detailed.
The biggest change is how often it can update. While earlier AI weather models typically worked with longer update cycles and lower spatial detail, WeatherNext 3 can refresh forecasts every hour. That could make it easier to track fast-changing conditions such as sudden rain, storms, snow and temperature shifts.
A closer look at changing weather
WeatherNext 3 uses raw geostationary satellite imagery, which comes from satellites that continuously monitor the same areas of Earth. This gives the model a near-continuous view of clouds, moisture and developing weather systems.
Google says combining these observations with other weather data helps the model respond more quickly when atmospheric conditions change. The model can visualise surface variables such as temperature and moisture at a 5-kilometre resolution.
Other surface variables are available at around 10 kilometres, while some atmospheric variables, including wind speed, are produced at a 25-kilometre scale. Google says WeatherNext 3 provides a global weather picture that is roughly 5 times more sensitive than WeatherNext 2.
The improvement could matter most when weather changes quickly. A forecast that updates every hour can potentially give people and businesses a more current picture than one based on information collected several hours earlier.
From umbrellas to energy grids
For everyday users, more frequent forecasts could help with simple decisions such as planning a commute, carrying an umbrella or organising an outdoor event. The bigger impact could come in industries where weather directly affects operations.
Farmers can use weather information to plan activities, while airlines and logistics companies need it to manage routes and schedules. Emergency services can benefit from better visibility into changing conditions, while renewable energy companies need accurate forecasts to estimate solar and wind generation.
Google says WeatherNext 3 improves precipitation forecasts by up to 50% when looking a day or more ahead. The model also includes clean-energy variables related to solar and wind power, potentially helping energy companies estimate generation as weather conditions change throughout the day.
WeatherNext 3 is being integrated into Google Search, Google Maps & its Platform Weather API, Google Earth Engine, Gemini, BigQuery and Google Cloud Storage.
That makes it more than a consumer weather tool. Developers, researchers and businesses can also use the forecasting system as part of their own applications and workflows.
Still, better AI forecasts do not make weather completely predictable. For severe weather warnings and public safety alerts, official meteorological agencies and local authorities remain the source to rely on.