Google DeepMind and Google Research have launched Weather Lab, an AI-powered interactive platform designed to improve the forecasting of tropical cyclones. The new model can predict a cyclone’s formation, track, intensity, size, and structure up to 15 days in advance, generating up to 50 possible scenarios. This AI model has demonstrated superior accuracy compared to traditional physics-based global models, reducing track errors by approximately 140 kilometers during the 2023 and 2024 hurricane seasons. The initiative is a collaboration with the U.S. National Hurricane Center (NHC), which is currently testing the model to enhance hurricane forecasts and warnings. The partnership aims to leverage AI to save lives, mitigate property damage, and improve economic outcomes by providing more precise and earlier cyclone predictions. Weather Lab’s public availability offers forecasters and researchers real-time and historical weather data to support disaster preparedness and response efforts. The model uses stochastic neural networks and graph neural networks to deliver its forecasts, marking a notable advancement in meteorological science and AI applications in weather prediction.