Physix Frontier · News Briefing Card (IT Home · Sep 3, 2026)
Google Launches WeatherNext 3 for Hourly AI Forecasts
KEY FACTS
- Google DeepMind has introduced the WeatherNext 3 model, enabling hourly updates for high-resolution global weather forecasts.
- The model achieves a surface variable resolution of 5 kilometers and an atmospheric variable resolution of 25 kilometers, trained on real-time satellite data.
- By leveraging sparse observation data to overcome computational limits, it delivers localized, precise forecasts for regions including Latin America and Africa.
- New features include renewable energy predictions that estimate wind speeds at 100 meters and solar radiation levels.
- The model is integrated into core Google products such as Search, Maps, Gemini apps, and Earth Engine starting immediately.
KEY DATA
5 kmMax Spatial Resolution
HourlyForecast Update Frequency
100 mWind Prediction Height
PHYSIX OBSERVATION
WeatherNext 3 marks a shift in AI meteorology from 'general prediction' to 'vertical scenario empowerment.' Its core value lies not just in improved accuracy, but in filling data gaps in the Global South at low cost while directly addressing pain points in energy trading. When weather forecasts can precisely guide grid dispatch, the service upgrades from information provision to critical infrastructure, signaling that tech giants are reshaping traditional industry decision-making logic through foundational data capabilities.
Source: IT Home report
Physix Frontier