WindBorne raises $37M to turn balloon-fed AI forecasts into a business

New deep-learning techniques behind large language models have also produced weather simulations that can run on laptops instead of supercomputers, changing meteorology — and the bigger challenge now may be making those forecasts useful to businesses. WindBorne Systems, a startup that collects data with long-flying weather balloons and feeds it into its own forecasting model, has raised a $37 million Series B to take on that challenge, CEO John Dean told TechCrunch. The round was co-led by Khosla Ventures and Galvanize, with participation from Translink Capital, Lux Capital, and previous investors, and values the company at $250 million after the raise.

Founded in 2019, WindBorne started by building a novel weather dataset with low-cost sensors and endurance balloons; AI forecasting models developed over the past four years let it generate its own forecasts, something previously out of reach for private companies given the cost of the supercomputers such simulations required. The company now operates 20 launch sites worldwide and keeps about 600 balloons aloft at any time, collecting data from hard-to-reach places like the eye of a typhoon, and has begun deploying aerial sensor packages that can fall into the ocean and continue gathering data as floating buoys. That proprietary dataset — what Dean calls a “planetary nervous system” — feeds a model that also ingests government weather-agency data, creating what he says is a moat. “We demonstrated that when you add balloons to the forecast, you get more accurate forecasts, and the value per data point is much stronger than satellites,” Dean said.

WindBorne’s main customers today are government agencies: the U.S. National Weather Service buys its data, while the U.S. Air Force and Navy pay through research partnerships, including an effort to build forecasting models that can run onboard ships with intermittent connectivity. The new funding will go toward compute spending, replacing the balloon network’s satellite communications with a mesh radio network, and building a go-to-market team to expand into the private sector — mainly, for now, investment funds that use weather data to predict commodity prices and other business outcomes. Galvanize partner Saloni Multani, who co-led the round, said integrating weather forecasts into business decisions has traditionally been expensive and difficult, but that AI is changing that calculation by making it easier to connect forecasts to the decisions businesses need to make.

Why it matters

Previous earth-observation startups have struggled to convert government contracts into private-sector revenue, since extracting value from that data requires workflows most commercial customers don’t have; WindBorne’s bet is that AI forecasting tools lower that barrier enough to finally open a private weather market beyond news media and speculators.