Google DeepMind and Google Research have introduced WeatherNext 3, an AI weather model that refreshes global forecasts every hour at up to 5 km resolution. It is already being used across Google Search, Gemini, Google Maps, the Maps Platform Weather API and Earth Engine, according to Gadget Pilipinas, which summarises Google’s announcement.
Google DeepMind Watch analysis
What happened
The model ingests geostationary satellite mosaics alongside ground-station data. Google says it can produce surface temperature and moisture estimates at 5 km resolution, surface winds at 10 km, and signals aimed at energy operators, including turbine-height wind, cloud cover and solar irradiance.
Google also claims precipitation errors fell by up to 50% on selected benchmarks against traditional numerical weather prediction baselines. That is a Google-reported result, not an independently verified forecast audit, and the supplied report does not establish Philippine-specific rollout dates or pricing.
Key findings
- Hourly refreshes
New forecasts arrive every hour, making the model more useful for decisions that cannot wait for a conventional update cycle. - Finer geographic detail
WeatherNext 3 works at up to 5 km resolution, bringing forecasts closer to local planning. - Energy-ready outputs
Turbine-height wind and solar-irradiance estimates target weather-dependent generation planning.
Why it matters
This is AI moving from impressive forecast charts into the plumbing of everyday decisions and energy systems. More frequent, finer-grained predictions could help people and operators respond sooner, if the claimed gains hold outside Google’s selected tests.
Our read
WeatherNext 3 looks like a meaningful applied-AI upgrade, especially because it is connected to products people already use. Treat the accuracy headline as promising rather than universal, and judge it against local performance during difficult weather.
What to watch
- Whether independent evaluations reproduce the claimed precipitation improvement.
- Performance during tropical storms and other extreme events.
- When the full feature set reaches specific countries and products.
- Whether energy operators report measurable planning gains from the new outputs.
Discussion spark: Would hourly, 5 km forecasts change a real decision you make, or is reliability during extreme weather the more important test?
Sources and evidence
- Google’s WeatherNext 3 Delivers Hourly Forecasts at 5km Resolution – Gadget Pilipinas (11 September 2026, 02:29 UTC)
not affiliated with, endorsed by, or operated by Google or Google DeepMind