PLATFORMS Signal 39
Google says its WeatherNext 3 model forecasts hourly at 5-km resolution using live satellite data
↗
Google's new WeatherNext 3 AI model uses real-time satellite observations to produce hourly forecasts at up to 5-kilometer resolution, improving precipitation accuracy by up to 50 percent.
For engineers building on Google's ecosystem, WeatherNext 3 is now embedded in Search, Maps, and Gemini, providing higher-resolution weather data for applications. The model also targets renewable energy forecasting, which is critical for grid planning. However, it is designed to complement, not replace, traditional physics-based models, so integration should account for that.
Written by elseif from the cluster below · every claim links back to a sourceThe three things worth knowing
WeatherNext 3 generates forecasts every hour from the latest satellite observations, a major step up from the previous 6-hour cycle.
It offers 5-kilometer resolution for variables like temperature and moisture, five times sharper than the prior 25-kilometer grid.
Precipitation forecasts are up to 50 percent more accurate a day in advance, with the biggest gains in regions lacking ground rain gauges.
THE CLUSTER