• Google DeepMind’s AI model predicts cyclone paths, intensity, and size 15 days ahead.
• Outperformed ECMWF’s model in tests, placing cyclone locations 140 km closer to actual paths.
• Weather Lab platform shares AI models and historical data for research, not official alerts.

Google DeepMind, in collaboration with Google Research, has launched Weather Lab, a new platform showcasing an innovative artificial intelligence (AI) model designed to predict tropical cyclones with unprecedented accuracy.

This experimental model can forecast cyclone formation, track, intensity, size, and shape up to 15 days in advance, generating 50 possible scenarios to aid disaster preparedness. The initiative, detailed in a blog post by Google, aims to enhance forecasting capabilities for extreme weather events, potentially saving lives through earlier and more precise warnings.

 The artificial intelligence model was developed on the basis of stochastic neural networks and trained with a mixture of reanalysis data and a history of almost 5,000 cyclone occurrences over 45 years. The model has been tested superior to the European Centre for Medium-Range Weather Forecasts (ECMWF) Ensemble Prediction System in the North Atlantic and East Pacific areas between 2023 and 2024.

Google DeepMind has partnered with the U.S. National Hurricane Center (NHC) to validate the model’s real-time forecasting capabilities during the 2025 hurricane season. Dr. Kate Musgrave from the Cooperative Institute for Research in the Atmosphere (CIRA) at Colorado State University noted that the model demonstrated “comparable or greater skill than the best operational models for track and intensity.”

Also Read:  Vodafone Idea Expands 5G Services with Samsung

Weather Lab also features other AI models, such as WeatherNext Graph and WeatherNext Gen, alongside historical cyclone track data for researchers to analyze. While the platform is a research initiative and not intended for official weather alerts, it represents a significant step toward integrating AI with traditional forecasting methods. Google’s earlier AI weather model, GenCast, published in Nature, outperformed physics-based models 97.2% of the time, highlighting the potential of AI-driven meteorology.

 The model’s ability to predict cyclone paths with high accuracy could improve evacuation planning and resource allocation, particularly for vulnerable communities.