The innovative deep learning methods utilized in LLMs have also enabled weather simulations to be conducted on laptops instead of supercomputers, transforming the field of meteorology. However, the greater challenge for AI might be simplifying the process for individuals and organizations to utilize those forecasts effectively.

WindBorne Systems, a startup that gathers data using the world’s longest-flying weather balloons and integrates it into a robust forecasting model, has secured a $37 million Series B funding round to tackle that challenge, CEO John Dean informed TechCrunch.

This new round was co-led by Khosla Ventures and Galvanize, with contributions from TransLink Capital, Lux Capital, and prior investors, valuing the company at $250 million following this funding round.

Established in 2019, WindBorne began with a strategy to gather a unique set of weather data using its affordable weather sensors and long-endurance balloons. The advancements in AI weather forecasting models over the past four years have enabled them to produce their own forecasts, which was previously unattainable for most private firms due to the prohibitive costs of supercomputers that were needed to model the atmosphere.

Currently, the company operates 20 launch sites globally and has around 600 balloons airborne at any time, collecting data in inaccessible regions, such as the center of a typhoon. The company is now starting to deploy aerial sensor packages that can descend into the ocean and persist in collecting data as floating buoys.

The unique data set produced by this “planetary nervous system,” as Dean refers to it, establishes a competitive edge for their weather model, which also incorporates data sets generated by governmental weather agencies worldwide.

“We showed that when you include balloons in the forecast, you achieve more precise predictions, and the value of each data point is significantly higher than that from satellites,” Dean stated. “We’ve also been increasing our revenue while doing this, which mitigates the risk for VCs regarding the demand signal.”

Currently, the company’s primary clients are government entities. The U.S. National Weather Service acquires the company’s data, while the U.S. Air Force and U.S. Navy engage with WindBorne through research collaborations, including initiatives to develop forecasting models operable onboard vessels that may experience sporadic connections to the broader world.

The next step involves entering the commercial sector — primarily targeting investment funds that leverage weather data to forecast commodity prices and other business results. In addition to expenditures on computing and efforts to transition from the balloon network’s satellite communications to a mesh radio system, this funding round will enable WindBorne to grow its go-to-market team to widen its customer reach in the private sector.

However, that is not always straightforward. Over the last ten years, various startups have attempted to scale sensing operations like earth observation satellite networks but struggled to penetrate the private sector, as extracting value from that data necessitates expertise and established procedures. Many revert to government agencies accustomed to utilizing that data already.

Private weather forecasting companies exist but primarily generate revenue by repackaging or refining government forecasts for media use, specialized applications such as aircraft de-icing and maritime navigation, or the aforementioned speculators. However, this may shift as AI tools enhance data analysis efficiency.

Saloni Multani, a partner at Galvanize who co-led the funding round, noted that the private weather market has remained constrained because “integrating weather forecasts into larger business decision-making has historically been costly and complex. We believe AI alters that dynamic. Superior forecasts make the endeavor worthwhile, and AI significantly simplifies connecting those forecasts to business decisions.”