Google's AI weather model gets sharper with real-time satellite data

google has launched an updated ai weather model called weathernext 3, which utilizes real-time satellite observations to make forecasts with unprecedented resolution. the new model can produce a global picture that's five times sharper than google's previous model, weathernext 2, by learning from fresher and richer observational data sets. weathernext 3 can generate forecasts every hour on a 5-kilometer grid, compared to weathernext 2's 25-kilometer grid every 6 hours. this increased speed and resolution significantly improves predictions for rain and snowfall stemming from fast-moving weather systems, especially in areas with limited ground-based weather monitoring. google claims that users will see precipitation forecasts that are up to 50 percent more accurate when looking at least a day in advance. the model also produces forecasts for renewable energy generation, including predictions on wind speed at 100 meters. this development is crucial for google, as the company seeks to increase its use of renewable energy to power its growing data centers.