WindBorne Systems has raised a $37 million Series B funding round at a $250 million valuation, with the capital aimed at expanding its autonomous balloon constellation and deepening its AI-driven weather modeling capabilities. The round was co-led by Khosla Ventures and Galvanize, with additional participation from TransLink Capital, Lux Capital, and previous investors, according to TechCrunch. Saloni Multani, a partner at Galvanize, co-led the round.

Founded in 2019 and based in Redwood Shores, California, WindBorne operates what it calls Atlas, described on its website as the largest balloon constellation in the world. The fleet currently comprises about 600 Global Sounding Balloons in the air at any given time, launched from roughly 20 sites around the world. The company launches hundreds of balloons each month, each capable of controlling its altitude and flying autonomously for weeks.

The operational advantage over traditional weather balloons — known as radiosondes — is stark. A conventional weather balloon produces a single vertical atmospheric profile (a snapshot of temperature, pressure, and humidity at various altitudes) over a lifetime of approximately two hours before it is lost. Each WindBorne balloon, by contrast, produces around 50 controlled vertical profiles over its multi-week flight duration, according to WindBorne's website. That density of observation is the raw material the company feeds into WeatherMesh-6, its deep learning forecast model, which WindBorne has announced as officially live.

The business model spans both commercial and government customers. The U.S. National Weather Service purchases data from WindBorne, and the U.S. Air Force and U.S. Navy maintain research partnerships with the company. WindBorne also has a partnership with the Gates Foundation, suggesting applications that extend beyond U.S. defense and domestic forecasting.

WindBorne is also expanding the range of environments its sensors can capture. The company is deploying aerial sensor packages designed to descend into the ocean and continue collecting measurements as floating buoys. That extends the platform's observational reach from the upper atmosphere to the ocean surface, a domain where direct measurement has historically been sparse relative to land-based collection.

The Series B builds on a $15 million Series A raised in June 2024, which was earmarked for scaling the balloon constellation and advancing AI weather modeling. The new round more than doubles the company's total known funding and arrives as the competitive landscape for AI-augmented meteorology grows increasingly crowded. Several organizations, from national weather agencies to commercial players, are applying machine learning to atmospheric prediction, but most rely on satellite and ground-based observational inputs. WindBorne's distinguishing bet is that its own hardware fleet collects vertical, high-frequency, on-location data — think of it as having ground-level reporters throughout the atmosphere rather than relying on satellite photos from afar — which gives its models a structural input advantage that downstream AI can exploit.

Whether that input advantage translates into forecast accuracy gains that customers will pay a sustained premium for is the commercial question the Series B is implicitly underwriting. The National Weather Service as a paying customer lends the dataset credibility; the defense partnerships indicate demand for operational meteorology in domains where the marginal value of better forecasting is high. But scaling a hardware-in-the-loop observational network across global launch sites is capital-intensive, and the path from data collection to proprietary model performance to a defensible commercial position is not a short one.

The company's technology is invented, designed, and manufactured in the USA, a detail that aligns with its government customer base and the regulatory environment surrounding aerospace and meteorological data. CEO John Dean leads the company, though further details about the leadership team were not available in the reported facts.

WindBorne's combination of proprietary sensing hardware, a growing observational network, and a deep learning forecast model places it at an unusual intersection of physical infrastructure and AI. The $250 million valuation reflects investor confidence that owning the data pipeline, not just the model, is where durable value accrues in applied AI. Whether that thesis holds will depend on the quality of forecasts WeatherMesh-6 can produce at scale, the cost economics of the balloon fleet, and the willingness of commercial customers to pay for weather intelligence beyond what government agencies provide for free. The capital to test those questions is now in place.