WindBorne Systems launches AI model to improve weather forecasting

WindBorne Systems has released WeatherMesh-6, an AI-driven weather forecasting tool that offers more frequent and accurate predictions than the European Centre for Medium-Range Weather Forecasts (ECMWF), an intergovernmental organization. According to chief product officer Kai Marshland, the model is as accurate five days out as a traditional forecast is the day before, particularly regarding surface temperature measurements. While traditional models generate forecasts every six hours, WeatherMesh-6 produces them hourly, with a resolution of 3 km in the continental U.S.

The model’s performance relies on WindBorne’s data collection infrastructure, which currently has about 400 balloons in flight launched from 15 sites globally. The company, founded in 2019, feeds this sensor data directly into its deep learning models. CEO John Dean emphasized the importance of this dataset, stating, “I don’t understand, personally, the business model of being [an] AI based weather company without a dataset advantage.” Dean noted that while the company previously relied heavily on the ECMWF for data assimilation—the process of turning disparate sensor readings into a comprehensive, machine-readable picture of the world—he predicted that if they removed the ECMWF’s initial conditions today, WindBorne would still perform well.

WindBorne has raised $25 million in venture funding, with a reported valuation of $85 million in 2024. The company sells its balloon data to the U.S. National Oceanic and Atmospheric Administration (NOAA), the U.S. Air Force, and the U.S. Navy, but Dean is focused on building model and data infrastructure rather than a software-as-a-service (SaaS) product. The company suffered a scare last year when a United Airlines jetliner flew into one of its balloons. To mitigate future risks, WindBorne monitors air traffic to maneuver its balloons, though it does not currently use ADS-B (a global aviation surveillance system) transponders on its sensor platforms.

Why it matters

WindBorne’s approach highlights a shift in weather forecasting where proprietary sensor data, rather than just model architecture, becomes the primary differentiator against established government-run systems.