A new generation of artificial intelligence models is rapidly emerging as a competitive alternative to traditional physics-based weather forecasting systems. Companies like Google, Nvidia, and Huawei have developed AI models that can generate accurate global weather predictions in minutes, a task that typically takes hours for conventional supercomputers. These AI systems, trained on decades of historical weather data, learn to identify patterns and make predictions without explicitly coding atmospheric physics. While they currently excel at short- to medium-range forecasts, experts note they still face challenges with predicting extreme weather events and very long-term trends where physics-based models retain an edge. The technology promises to make high-resolution forecasting more accessible and cost-effective. Read the full article: https://technologyreview.com/2024/05/15/1092950/ai-weather-forecasting-models/