Weather balloons meet AI: WindBorne raises $37 million to sell smarter forecasts

Key points

• WindBorne Systems raised $37 million in a Series B funding round, valuing the company at $250 million.
• The startup operates roughly 600 weather balloons from 20 launch sites and sells forecast data to the U.S. National Weather Service, Air Force, and Navy.
• AI-based forecasting tools now let WindBorne run atmosphere simulations that once required a supercomputer, cutting costs dramatically.
• The company is developing sensor packages that drop into the ocean and keep collecting data as floating buoys.
• WindBorne plans to use the new funding to expand sales to private-sector buyers, starting with investment funds that trade on weather data.

For most of history, predicting the weather accurately required a room-sized supercomputer crunching billions of calculations. New AI models, the same family of technology that powers chatbots, have changed that. A forecast that once demanded a national government's computing budget can now run on a laptop.

WindBorne Systems, a startup founded in 2019, wants to turn that shift into a business. As first reported by TechCrunch AI, the company has raised a $37 million Series B round, co-led by venture firms Khosla Ventures and Galvanize, with participation from TransLink Capital and Lux Capital. The round values WindBorne at $250 million.

What does WindBorne actually do?

It collects weather data from places that satellites and ground stations miss, using cheap, long-endurance balloons.

The company currently keeps about 600 balloons in the air at any given time, launched from 20 sites worldwide. The balloons are built to stay aloft far longer than standard weather balloons, reaching places like the eye of a typhoon. CEO John Dean calls the network a "planetary nervous system." Alongside the balloons, WindBorne is now testing sensor packages that parachute into the ocean and then float on the surface, continuing to collect readings as buoys.

All that raw data feeds into WindBorne's own AI forecasting model, which also pulls in publicly available data from government weather agencies. Dean says the balloon data makes those forecasts meaningfully more accurate than satellite data alone.

Current paying customers are almost entirely government bodies. The U.S. National Weather Service buys WindBorne's data directly. The Air Force and Navy fund research partnerships, including work on forecasting models that can run on a ship with a patchy internet connection, where downloading a full forecast from shore is not always possible.

Why is the private market so hard to crack?

Selling weather data to businesses has tripped up many startups before. Turning a raw forecast into a useful business decision takes experience and existing workflows that most companies do not have.

That is why private weather companies have mostly made money repackaging government forecasts for narrow needs: news broadcasts, aircraft de-icing schedules, ship routing. The customers who want raw data tend to be investment funds that use weather patterns to predict crop yields or commodity prices.

Galvanize partner Saloni Multani, who co-led the round, put it plainly: "Integrating weather forecasts into broader business decision-making has traditionally been expensive and difficult. We think AI changes that equation."

The argument is that AI tools can now connect a forecast directly to a business decision without an army of specialist analysts in between.

What happens next?

WindBorne will use the new money on four things: computing costs, replacing the balloon network's satellite links with a mesh radio network (a web of radios that pass signals between balloons instead of routing everything through a satellite), building a sales team, and expanding into the private sector.

The bet is that AI lowers the barrier enough that businesses outside finance start acting on weather data routinely. Whether food companies, energy traders, or logistics firms will pay for that remains to be seen.