WindBorne Systems Secures Major U.S. Air Force Contract for Weather Balloon Data
The U.S. Air Force has entered into a significant contract with WindBorne Systems, a Redwood City-based startup specializing in autonomous weather balloons. Under this agreement, which could be worth nearly $50 million, the Air Force will purchase atmospheric data gathered by WindBorne’s extensive global fleet. This contract marks the largest single engagement for the company, underscoring its growing influence in the weather data and forecasting industry.
Founded in 2015, WindBorne currently operates approximately 700 weather balloons worldwide. These balloons collect crucial atmospheric sensor readings, which are sold to various customers and simultaneously integrated into WindBorne’s proprietary AI-driven weather forecasting model. The company is among the pioneers in directly feeding empirical data into deep-learning models for weather prediction, aiming to enhance the accuracy and timeliness of forecasts.
A Strategic Partnership with the U.S. Military
WindBorne has a longstanding relationship with U.S. government agencies, including the National Weather Service, the U.S. Navy, and the Air Force. However, this latest contract is notable for its scale and flexibility. Valued at up to $49 million, the Air Force pays only for the data it requests, which can be region-specific and time-sensitive. To date, the Air Force has placed orders totaling $1.7 million, with expectations of further purchases as operational needs evolve.
Andrey Sushko, WindBorne’s co-founder and CTO, explained to TechCrunch that the contract supports not only the expansion of the balloon constellation but also ongoing research and development efforts to improve the balloons’ hardware and performance. He emphasized that the contract was awarded without competitive bidding, reflecting the government’s recognition that WindBorne uniquely meets its requirements for global atmospheric data collection.
Innovating Weather Forecasting with AI and Empirical Data
WindBorne’s CEO, John Dean, revealed in August that the company had successfully raised $37 million in a Series B funding round to scale operations further. Earlier that summer, WindBorne launched WeatherMesh 6, its latest forecasting model. According to the company, WeatherMesh 6 outperforms the widely respected European Centre for Medium-Range Weather Forecasts (ECMWF) traditional model, which relies on physics-based atmospheric simulations.
The transformation of weather forecasting through AI has accelerated with the adoption of transformer architectures and the availability of large-scale weather datasets like those released by ECMWF. Traditional methods required complex supercomputing resources to run detailed atmospheric equations. In contrast, modern AI models can ingest decades of historical weather data and produce faster, often more accurate forecasts.
Frontier research labs and government agencies are increasingly validating these AI models’ accuracy, although both AI-driven and traditional forecasting methods currently complement each other. Notably, Google DeepMind’s latest AI weather model has introduced new capabilities, particularly in cyclone and hurricane forecasting, becoming essential tools for U.S. government response during severe weather events.
Breaking New Ground in Autonomous Weather Balloon Navigation
In a landmark achievement this August, WindBorne became the first organization to send an unpowered flight system—one without an engine—into the eye of a typhoon to collect atmospheric data. The company maneuvers its balloons by adjusting their altitude, releasing air, or dropping ballast like sand, which allows the balloons to catch specific wind currents and navigate toward targeted locations.
Andrey Sushko highlighted the novelty of this approach, noting that it depends on two main factors: accurate wind and weather models to guide navigation, and a sufficiently large balloon constellation to ensure presence in critical areas. This operation represented one of the first instances where a deep-learning model directly informed operational decision-making in real time.
When you purchase through links in our articles, we may earn a small commission. This doesn’t affect our editorial independence.
Source: Here
