Microsoft says its Aurora AI can accurately predict air quality, typhoons, and more | TechCrunch
May 23, 2025 - techcrunch.com
Microsoft has unveiled Aurora, an advanced AI model capable of predicting weather-related phenomena such as air quality, hurricanes, and typhoons with high accuracy. Detailed in a paper published in Nature and a blog post, Aurora is trained on over a million hours of data from various sources and can be fine-tuned for specific weather events. Microsoft claims Aurora outperforms traditional meteorological methods in precision and speed, having successfully predicted Typhoon Doksuri’s landfall and bested the National Hurricane Center in forecasting tropical cyclone tracks for the 2022-2023 season.
Aurora's efficiency allows it to generate forecasts in seconds, a significant improvement over traditional systems that require hours on supercomputers. Microsoft has made Aurora's source code and model weights publicly available and plans to integrate the AI model into its MSN Weather app for hourly forecasts. This development positions Aurora as a leading tool in weather science, potentially benefiting research labs and enhancing public weather forecasting services.
Key takeaways:
Microsoft's AI model Aurora can predict weather phenomena like air quality, hurricanes, and typhoons with high precision and speed.
Aurora was trained on over a million hours of data and can be fine-tuned with additional data for specific weather events.
In experiments, Aurora outperformed expert predictions and the National Hurricane Center in forecasting certain weather events.
Microsoft has made Aurora's source code and model weights publicly available and is integrating it into the MSN Weather app for hourly forecasts.
