WindBorne wins a hefty US Air Force contract for its balloon data
TechCrunch Tim Fernholz
The Air Force is paying WindBorne for weather-balloon data in a deal worth up to $49 million. It’s the company’s biggest contract, and it backs a weirdly powerful AI weather setup.
Based on reporting by TechCrunch, Tim Fernholz — read the original for the full story.
Summary, retelling and take written by AI under human oversight; images are AI-generated illustrations. How we work · Report an error
The U.S. Air Force has signed up to buy weather data from WindBorne Systems, turning the startup’s balloon fleet into a government supplier. The deal could reach nearly $50 million, but the service only pays when it actually orders data, and those orders can be targeted to specific places and times.
WindBorne says the Air Force has already placed $1.7 million in orders. More are expected. For the company, this is the biggest contract it has taken part in, and it comes without competitive bidding. WindBorne co-founder and CTO Andrey Sushko says the money helps expand the balloon constellation and fund research and development to improve the hardware and performance of the balloons.
That constellation is already large. WindBorne says it has about 700 autonomous weather balloons in the air around the world, collecting atmospheric readings that the company sells to customers and also uses in its own AI forecasting system. The company, founded in 2015 and based in Redwood City, Calif., says it was one of the first to feed real-world atmospheric data directly into a deep-learning weather model.
WindBorne has been close to U.S. government users for a while, especially the National Weather Service and the Air Force and Navy. In August, CEO John Dean said the company had raised a $37 million Series B round to keep scaling. Around the same time, WindBorne released WeatherMesh 6, which it says is more accurate than the European Centre for Medium-Range Weather Forecasts’ traditional physics-based model.
The larger story here is that AI weather forecasting is moving from theory to practice. Frontier labs and government agencies are building models that can beat older forecasts, even if both still matter. WindBorne added another milestone in August when it became the first organization to send an unpowered flight system into the eye of a typhoon to gather data, using balloons that steer by changing altitude, bleeding air, or dropping sand ballast. Sushko called it one of the first times a deep-learning model was used for operational decisions.
My take — AI-written commentary, not fact-checked reporting
This is the kind of deal that makes the AI weather boom look less like a demo and more like infrastructure. Also, nothing says “modern forecasting” like government money, balloons, and a model that needs the sky to cooperate. Closed enough to be useful, open enough to be a little unsettling — that’s the whole business now.
Read more about this at: TechCrunch