\n\n
Key developments: The National Hurricane Center (NHC) has begun basing its Hurricane Isaias forecasts on Google DeepMind’s WeatherNext 3 AI model, moving away from sole reliance on classic physics‑based systems such as the European Centre and GFS models. This shift follows successful use of the same AI system in 2025, when it helped the European Centre for Medium‑Range Weather Forecasts (ECMWF) launch the AIFS model that delivers finer‑grained extreme‑event probabilities while consuming less electricity than conventional forecasts. The NHC’s confidence grew after DeepMind’s AI correctly forecast the rapid intensification of Hurricane Melissa in the 2025 Atlantic season, a prediction that physics‑only models missed.
Why it matters: AI‑driven forecasts are proving more consistent and faster, allowing forecasters to issue earlier, more aggressive intensity warnings for rapidly intensifying storms—critical as climate change increases the frequency of such events. The lower energy footprint of AI models also mitigates the growing carbon cost of climate‑related computing, aligning predictive capability with sustainability goals.
What to watch next:
1. Expanded deployment of AI models like WeatherNext 3 across other NHC storm seasons, potentially becoming the dominant tool for track and intensity predictions.
2. Further integration of AI with traditional physics models to create hybrid systems that leverage the strengths of both approaches.
3. Ongoing evaluation of AI energy use versus accuracy gains, especially as data‑center demand rises.
4. International adoption of AI forecasting in regions with limited observational infrastructure, addressing equity gaps highlighted in recent studies of AI model performance across wealthier and poorer nations. [[1, 2]]