In October 2024, Hurricane Milton emerged as one of the fastest-growing storms on record in the Atlantic Ocean. Its rapid intensification took meteorologists by surprise, leading to devastating consequences that resulted in 15 fatalities and US $34 billion in damages across Florida. The inability to predict Milton’s explosive growth stemmed from a critical lack of weather data—data that can only be obtained by flying specialized aircraft directly into the storm. This method requires human pilots to undertake dangerous missions to deploy dropsondes—sensors that gather essential atmospheric measurements.
The challenges faced by meteorologists highlight significant data gaps in storm-prone areas, particularly over vast ocean expanses where most storms develop. To address this problem, WindBorne Systems, a company co-founded in 2019, is pioneering advanced weather prediction techniques that leverage cutting-edge technology.
Revolutionizing Weather Data Collection
Located in Palo Alto, California, WindBorne has developed Global Sounding Balloons (GSBs), which are capable of remaining airborne for weeks. This is a substantial improvement compared to traditional weather balloons, which typically operate for only a few hours. Prior to Hurricane Milton, WindBorne successfully launched six GSBs from Mobile, Alabama, equipped with dropsondes. Within 24 hours, these balloons penetrated the hurricane and released sensors that measured temperature, pressure, humidity, wind speed, and direction—critical data that could have informed more accurate forecasts.
The deployment of dropsondes from a weather balloon marked a significant milestone and demonstrated that it is possible to collect weather data without risking human lives. When WindBorne’s collected data was analyzed using their AI-based forecasting model, WeatherMesh, the predictions of Milton’s trajectory proved to be more accurate than those produced by the U.S. National Hurricane Center. Unfortunately, the results of this experiment could not be shared with the public in real time, as it was primarily a test of WindBorne’s technology.
Building an AI-Driven Forecasting System
WindBorne aims to create what they describe as a “planetary nervous system”—an AI-driven forecasting system that gathers extensive weather data and transforms it into timely, accurate predictions. This system is expected to significantly enhance decision-making in critical areas such as national defense, renewable energy, and agriculture. As climate change exacerbates the frequency and severity of extreme weather events, the need for improved forecasting becomes increasingly pressing.
The origins of WindBorne trace back to a project initiated in the Stanford Student Space Initiative in 2015, where co-founder and Chief Technology Officer Andrey Sushko and fellow students focused on extending the flight duration of conventional weather balloons. It was discovered that traditional weather balloons only monitored around 15 percent of the globe, leaving vast areas—like oceans and deserts—severely under-observed.
To improve data collection, the WindBorne team developed autonomous balloons that can navigate and “surf” the winds, allowing them to remain airborne for weeks. Each GSB weighs less than 2 kilograms and uses a thin, transparent film—only 20 micrometers thick—for its envelope. The balloons are equipped with a ballast system, enabling them to adjust altitude by releasing sand or venting gas to access different wind currents.
Currently, WindBorne’s GSBs form part of a growing global constellation, known as Atlas, which continuously collects more data daily than the U.S. National Weather Service. The data collected is then processed by WeatherMesh, which has demonstrated remarkable accuracy in its forecasts.
By 2028, WindBorne aims to expand its Atlas balloon constellation to approximately 10,000 GSBs, allowing for near-continuous observation of the globe. This ambitious expansion will require around 300 launches per day from approximately 30 launch sites worldwide.
WindBorne’s approach signifies a potential shift in weather forecasting, where traditional physics-based models can coexist with AI-powered systems, each complementing the other’s strengths. As weather-related challenges intensify with climate change, innovative forecasting solutions like those from WindBorne will be crucial in enhancing preparedness and response strategies worldwide.
Reflecting on Hurricane Milton, the WindBorne team views the event as a pivotal moment that validated their technology’s capacity to deliver timely and accurate weather predictions. In a world where a mere 12 to 24 hours of advance warning can save lives, WindBorne’s AI-driven forecasting represents a transformative step forward in meteorology.
