WeatherNext 3
VerifiedGoogle DeepMind's global AI weather model with hourly, high-resolution forecasts
Product preview
WeatherNext 3
What it is
WeatherNext 3 is the latest global weather forecasting model from Google DeepMind and Google Research, announced in September 2026. It is an ensemble model that draws on raw satellite imagery alongside traditional analysis data, which lets it produce forecasts for every hour of the day. Google says it predicts surface variables such as temperature and humidity at up to 5 km resolution, with other variables like wind at coarser grids, and it covers energy-relevant fields such as cloud cover and solar radiation. The model feeds forecasts into Google Search, Maps and Gemini, and enterprises can access its data through BigQuery, Earth Engine, Google Maps Platform and Cloud Storage.
Key features
- Hourly global ensemble forecasts
- Direct use of satellite observations
- High-resolution surface variables
- Data access via BigQuery and Earth Engine
What we liked
- + Backed by Google DeepMind research
- + Reaches users through Search, Maps and Gemini
- + Covers variables useful for renewable energy
Worth noting
- · Enterprise data access runs through paid Google Cloud services
- · Not a standalone consumer app
Editor's notes
More context from AI BirdView
Weather forecasting has traditionally relied on expensive physics simulations run a few times a day. WeatherNext 3 shows how far AI models have moved that field, generating forecasts every hour and pulling satellite observations directly into the model rather than waiting for them to be processed into standard analysis data.
- Produces an ensemble of forecasts, which helps express uncertainty rather than a single guess
- Delivers fine-grained surface forecasts, according to Google, at up to 5 km resolution
- Covers cloud cover and radiation for wind and solar operators
- Flows into everyday Google products as well as enterprise data services
- Can be explored through Google's interactive Weather Lab
WeatherNext 3 is most relevant for data scientists, energy and logistics teams, researchers and developers who want high-resolution forecast data inside tools they already use, such as BigQuery or Earth Engine. Everyday users benefit indirectly through Search and Maps. A practical tip: start in Weather Lab to compare forecasts for locations you care about, then move to the BigQuery or Earth Engine datasets once you know which variables your project needs.
Reader reviews
From people who've used WeatherNext 3. Emails are never shown publicly.
No reviews yet. Be the first to share your experience.
Alternatives to WeatherNext 3
Other tools we like in this space.
