WindBorne Systems will sell weather data collected by its autonomous balloon fleet to the U.S. Air Force under a contract that could be worth up to $49 million. The Air Force orders data when it needs it, choosing particular regions and times, and has ordered $1.7 million worth so far.
A contract tied to demand
The contract is the largest WindBorne has participated in. It does not guarantee the company the full $49 million: the Air Force pays as it places orders for data. WindBorne executives expect more orders to follow.
WindBorne co-founder and CTO Andrey Sushko told TechCrunch that the agreement can support continued growth of the balloon network and research and development to improve balloon hardware and performance. He said the contract was awarded without competitive bidding, with the government determining that WindBorne was the only company able to provide global data collection in a way that met its operational needs.
The arrangement links the Air Force’s purchases to the information its operations require, while giving WindBorne a potential source of support for expanding its network and improving its equipment. The contract’s maximum value describes what it could reach, while the orders placed so far show the amount currently committed through purchases.
What the balloons measure
WindBorne has about 700 weather balloons flying around the globe. They collect atmospheric sensor readings, which the company sells to customers and also uses in its own AI weather forecasting model.
Founded in 2015 and based in Redwood City, Calif., WindBorne has worked with the U.S. government, including the National Weather Service and the U.S. Air Force and Navy. The new agreement builds on that work by making the company’s data available through orders for specific places and times.
Weather balloons can also be directed through the atmosphere. WindBorne adjusts their altitude, releases air, or drops sand carried as ballast. Those adjustments help send balloons into winds that carry them toward desired locations.
Forecasting with data and deep learning
Weather forecasts have traditionally relied on expensive supercomputers running detailed equations that approximate how the atmosphere behaves. AI forecasting offers another approach: models can learn from decades of weather data and produce predictions more quickly.
WindBorne is among the first organizations to feed empirical data directly into a deep-learning weather model. Earlier this summer, it released WeatherMesh 6, its latest forecasting model. The company says it provides more accurate forecasts than the traditional model run by the European Centre for Medium-Range Weather Forecasts, which simulates the atmosphere using the laws of physics.
AI forecasts are not a complete replacement for traditional ones. The source describes both kinds of forecast as necessary for useful understanding of the weather, even as models from frontier labs and government agencies prove more accurate. The value of WindBorne’s data contract therefore sits within a broader shift toward combining new forecasting methods with established approaches.
Putting forecasts to work in flight
In August, WindBorne became the first organization to send an unpowered flight system into the eye of a typhoon to collect data. The system had no engine; its balloons relied on altitude adjustments and the surrounding winds for navigation.
Sushko described the flight as a novel step made possible by two things: useful wind and weather models, plus a large enough balloon network to have vehicles in the right places. He also said it was among the first times a deep-learning model had been used for actual operational decisions.
That example shows how the company’s work joins data collection, forecasting and navigation. The balloons gather observations; models help interpret weather and winds; and the vehicles can be steered by changing altitude and ballast. The Air Force contract centers on buying data, while WindBorne says the agreement can also support work to improve the hardware and balloon performance.
WindBorne has also raised a $37 million Series B round to continue scaling its business, CEO John Dean told TechCrunch in August. Together, the funding and the Air Force contract provide context for the company’s effort to grow its balloon network and forecasting work, though the contract’s ultimate value depends on the Air Force’s future orders.