ECMWF’s AI Weather Forecasts Go Operational: A New Era in Prediction Begins

By Dale C. S. Destin – Published 18 April 2025 |

In a historic move for global weather forecasting, the European Centre for Medium-Range Weather Forecasts (ECMWF) has officially launched its Artificial Intelligence Forecasting System (AIFS) as an operational tool. This is the first AI-powered weather model of its kind to be openly available and fully supported — and it’s now running side-by-side with ECMWF’s traditional physics-based system.

Why does this matter? Because AI is already outperforming traditional models in some areas — including tracking tropical cyclones, and it does so with up to 1,000 times less energy use and much faster processing.

What is the AIFS?

The Artificial Intelligence Forecasting System is a machine learning model trained on how the weather has evolved over the past. It uses the same starting data as ECMWF’s standard system — a combination of satellite observations, ship and plane reports, sea buoys, and more — but instead of running complex physics calculations, it uses AI algorithms to quickly predict how weather will evolve.

Currently, it produces deterministic forecasts (a single, most likely scenario), but ECMWF is already working on rolling out ensemble forecasts (multiple scenarios) in the near future.

Why it’s important for the Caribbean

This development is more than just a scientific breakthrough — it has real-world applications for the Caribbean and small island developing states (SIDS):

• Tropical cyclone tracking improvements mean better early warning and planning.
• Renewable energy sectors benefit from AI-driven forecasts of sunshine and wind levels.
• Marine safety, insurance, shipping, and fisheries sectors can get faster, potentially more accurate information.
• Lower energy requirements mean that in the future, countries with limited computing resources may benefit from lighter AI forecasting models.

What’s next?

The AIFS will soon expand to include:

• Ensemble forecasts for better risk-based decision making.
• Sub-seasonal forecasting (weeks ahead).
• Potential hybrid models combining physics and AI.

ECMWF sees this AI system not as a replacement, but as a complement to its traditional methods — allowing users to choose what best suits their needs.

Why it matters for 268Weather followers

As weather prediction continues to evolve, tools like AIFS will help enhance regional capacity to respond to severe weather, climate extremes, and sector-specific needs — from energy to agriculture and coastal protection. 268Weather will be watching closely as this new AI system grows, especially as we look for innovative tools that can support resilience in the Caribbean.

Want to dive deeper into how it works? Read the full ECMWF release here.