Google and DeepMind Unveil WeatherNext 3, Forecasting Weather Directly from Satellite Data
Original title:Google's WeatherNext 3 ditches physics simulations and learns weather directly from live satellite data
Google Research and DeepMind have introduced WeatherNext 3, a weather forecasting model that departs from conventional numerical physics simulations to learn directly from real-time satellite imagery. The model produces hourly forecasts at up to five-kilometer resolution—five times sharper than its predecessor. By bypassing heavy fluid dynamics pipelines, the system aims to supply granular, low-latency forecasts to historically underserved regions across Africa, Latin America, and the Asia-Pacific.
Why it's worth reading
WeatherNext 3 demonstrates a practical shift from classical numerical weather prediction toward end-to-end satellite learning, offering granular forecasts to data-sparse regions without compute-heavy physical simulations.