US Air Force Signs Up to $49 Million Deal for WindBorne's Balloon Data

The U.S. Air Force will pay WindBorne Systems up to $49 million for atmospheric data gathered by its fleet of roughly 700 autonomous weather balloons. It's the single largest contract the Redwood City, California startup has ever landed — and the government awarded it without competitive bidding.

The reason for the no-bid award is blunt, at least in the company's telling. WindBorne co-founder and CTO Andrey Sushko said the contract is "a determination by the government that for this capability of collecting global data, we were the only company able to provide it in the way that met their needs for these operations."

The deal is structured as a pay-as-you-go arrangement. The Air Force orders data for specific regions at specific times, and only pays when it does. So far it has ordered $1.7 million worth, according to the company. Executives expect more orders to follow.

What does the contract actually cover?

The $49 million ceiling is the headline figure, but the mechanics matter. The Air Force isn't buying hardware or balloons outright. It's buying sensor readings — the atmospheric data WindBorne's balloon constellation collects as it circles the globe.

That data serves two purposes inside the company. WindBorne sells it to customers, including the U.S. government. It also feeds the readings into its own AI weather forecasting model, WeatherMesh.

Sushko framed the contract as more than a revenue line. "This is largely a vehicle through which continued growth of the constellation can be supported, as well as some work on research and development to continue to improve the hardware and performance of the balloons," he told TechCrunch.

WindBorne, founded in 2015, isn't new to government work. The company has worked closely with the National Weather Service, the Air Force and the Navy. But nothing in its history approaches this scale. The new contract instantly makes the Pentagon one of the startup's most significant customers.

Why does balloon data matter for AI forecasting?

WindBorne occupies an unusual position in the AI weather boom. It is one of the first organizations to directly feed empirical, real-world sensor data into a deep-learning weather model.

That distinction matters because of how the field has evolved. Since transformer model architectures arrived, and the European Centre for Medium-Range Weather Forecasts (ECMWF) released a massive weather dataset a few years ago, AI weather forecasting has become one of the more productive applications of deep learning.

The old approach relied on expensive supercomputers running detailed equations that approximate atmospheric behavior. The new approach lets end-to-end learning systems ingest decades of weather data and issue predictions faster — and increasingly, more accurately.

WindBorne's latest model, WeatherMesh 6, was released earlier this summer. The company says it delivers more accurate forecasts than the ECMWF's traditional physics-based model, which simulates the atmosphere using the laws of physics.

The stakes run beyond one startup's benchmarks. Models built by frontier labs and government agencies are now beating traditional forecasts in many cases, though the frontier remains jagged — meteorologists still need both types of forecasts to build a useful picture of the weather. Google DeepMind's latest weather model offers new capabilities, and its cyclone forecasting tools are becoming increasingly central to how the U.S. government responds to hurricane season.

A company that both collects the raw data and trains the model on it controls a vertically integrated pipeline that few competitors can match. The Air Force contract funds exactly that loop: more balloons, more data, better models.

What did WindBorne do inside a typhoon?

In August, WindBorne notched a first that underlines why the military wants its capabilities. The company became the first organization to send an unpowered flight system — a vehicle with no engine — to collect data in the eye of a typhoon.

The balloons steer without propulsion. WindBorne maneuvers them by adjusting altitude, bleeding air, or dropping sand carried as ballast, riding winds that carry them where they need to go.

Sushko explained the two requirements behind the feat: "That's a pretty novel step, and the two ingredients for that are one, having good wind and weather models, and knowing how to navigate them; and two, having a large enough constellation that you have balloons in the right places."

He added: "It's one of the first times that a deep-learning model was used for actual operational decisions."

That last point is the quiet headline. Most AI weather models make forecasts that humans consult. WindBorne's models are making navigation decisions for its own hardware, in real time, inside a typhoon's eye.

What's next for the company?

The Air Force contract lands on top of an already strong year. In August, WindBorne CEO John Dean told TechCrunch the company had raised a $37 million Series B round to continue scaling its business.

Add the numbers up and the trajectory is clear. A $37 million funding round, a $49 million contract ceiling, roughly 700 balloons in the air, and a typhoon penetration in a single year — all pointed at expanding a constellation the Pentagon now depends on.

The open question is order volume. The Air Force has committed $1.7 million so far against a $49 million ceiling. Whether the Pentagon exercises the rest will test whether a fleet of unpowered balloons can become core weather infrastructure for the U.S. military — or remain a niche, if genuinely unmatched, source of atmospheric data.