WindBorne Systems, a startup that gathers weather data using the world's longest-flying weather balloons and feeds it into its own AI-powered forecasting model, has raised a $37 million Series B round to help turn accurate forecasting into a profitable business. Advances in AI over the past four years have made it possible for private companies to run sophisticated weather simulations without supercomputers, and WindBorne's challenge now is persuading commercial customers, not just governments, to pay for the resulting forecasts.

The round was co-led by Khosla Ventures and Galvanize, with participation from TransLink Capital, Lux Capital and existing investors, valuing WindBorne at $250 million. Founded in 2019, the company operates 20 launch sites and around 600 balloons worldwide, and is expanding into ocean-based sensor buoys, feeding a proprietary dataset that CEO John Dean calls a "planetary nervous system". Its main customers today are government bodies such as the US National Weather Service, Air Force and Navy, and the new funding will go towards compute, a mesh radio network to replace satellite communications, and building a sales team to break into the private sector, where investment funds using weather data to predict commodity prices are an early target.