TL;DR
- Google has announced “WeatherNext 2,” its most advanced weather forecasting model to date.
- The company claims the new AI-powered weather prediction model outperforms the previous version on 99.9% of variables, resulting in more accurate and timely reports.
- Google is integrating WeatherNext 2 in the Pixel Weather app, Google Search, Gemini, and Google Maps.
Google is giving its weather tools a major upgrade. The company has announced WeatherNext 2, its most advanced AI-powered weather forecasting model yet. According to Google, WeatherNext 2 will make weather information in the Pixel Weather app, Google Search, Gemini, and Google Maps more accurate and more useful.
That massive forecasting boost comes from WeatherNext 2, Google’s most advanced, AI-powered weather forecasting model, developed by Google DeepMind and Google Research.
It represents a huge leap in speed and accuracy for all of Google’s consumer-facing products, including the dedicated Pixel Weather app, Google Maps, Google Search, and Gemini.
The Power of WeatherNext 2
WeatherNext 2 is a state-of-the-art neural model that shifts away from traditional, time-intensive physics-based simulations to deliver faster, higher-resolution forecasts.
1. Speed and Resolution
-
8x Faster: The model can generate forecasts up to eight times faster than its predecessor.
-
Hour-Level Resolution: It provides highly detailed, hour-by-hour predictions out to 15 days, offering a level of precision crucial for planning daily activities or identifying sudden weather changes.
-
Efficiency: It can generate these complex forecasts in less than a minute using a single TPU chip, a process that used to take supercomputers hours.
2. Accuracy and Reliability
-
Superior Performance: Google claims WeatherNext 2 surpasses the previous model on 99.9% of weather variables (e.g., temperature, wind, humidity) across all lead times.
-
Predicting the Extreme: A core strength of the model is its ability to simulate hundreds of possible weather scenarios from a single starting point. This ensemble approach provides a fuller picture of uncertainty, making it far better at predicting low-probability but high-impact events like severe storms or cyclones.
How This Improves Pixel and Maps
This new AI model is integrated into the core forecasting system, leading to practical, real-world improvements for users:
| App/Service | Benefit from WeatherNext 2 |
| Pixel Weather App | More accurate and timely hour-by-hour updates on precipitation (when rain will start/stop), temperature shifts, and wind changes, allowing for better micro-planning. |
| Google Maps | Provides more reliable and swifter real-time weather layers and hazard alerts during navigation (driving, running, or cycling), helping users avoid sudden adverse conditions on their route. |
| Google Search / Gemini | Faster, more accurate answers and AI-generated summaries when users ask general weather questions (e.g., “What will the weather be like this weekend?”). |
The upgrade is primarily server-side, meaning users should start seeing improvements in their apps without needing to download a new app update.