Recent advancements in artificial intelligence are transforming weather prediction, with new AI models demonstrating the potential to outperform conventional physics-based forecasting systems. These machine learning models, trained on decades of historical weather data, can generate accurate global forecasts in minutes rather than hours, a significant reduction in computational cost and time. While traditional numerical weather prediction relies on complex equations solved by supercomputers, AI systems like Google’s GraphCast and Nvidia’s FourCastNet learn atmospheric patterns directly from data. Early results show these models can match or exceed the accuracy of leading systems for forecasts up to 10 days ahead, particularly in tracking severe weather events. However, experts caution that AI models are not a complete replacement, as they may struggle with unprecedented extreme events not represented in their training data and lack the explicit physical understanding of traditional models. The future likely involves a hybrid approach, combining the speed of AI with the rigorous physics of conventional systems. Read the full article at: https://technologyreview.com/2024/05/15/ai-weather-forecasting-models