Windborne Systems releases WeatherMesh 6, an AI weather model the startup says is more accurate than the European Centre for Medium-Range Weather Forecasting's forecasts

Windborne Systems, founded by Stanford students in 2019, released the sixth version of WeatherMesh today. The company says it offers more accurate forecasts than the ECMWF's traditional and AI systems across several variables, and produces forecasts every hour rather than every six hours, with resolution down to 3 km in Europe and the continental US.

WeatherMesh 6 improves through better data assimilation (the process of turning sensor readings into machine-readable forecasts). Windborne operates about 400 balloons in flight launched from 15 sites globally, and feeds their direct sensor data into the model—a capability that previously required reliance on ECMWF data. The company's head of AI said direct data ingestion from the balloons was key to the improvement.

Windborne has raised $25 million in venture funding with a reported valuation of $85 million in 2024, and sells balloon data to NOAA, the U.S. Air Force, and the Navy. The company also sells forecasts to investors and commodity traders, though leadership states focus remains on building model and data infrastructure over commercial SaaS products.