WindBorne Systems Raises $37M Series B to Scale Weather AI Network

WindBorne Systems has raised a $37M Series B to expand the atmospheric sensing network and AI models behind its weather intelligence platform. The oversubscribed round was co-led by Khosla Ventures and Galvanize, with TransLink Capital, Lux Capital, and existing investors participating. The financing brings WindBorne's total funding above $62M.

This is not a bet on another weather app with a cleaner radar screen. WindBorne operates Atlas, a global constellation of more than 600 autonomous, long-duration balloons that collect atmospheric observations from places where traditional systems have limited coverage. Those measurements feed WeatherMesh, the company's deep-learning forecast platform, creating a vertically integrated stack that combines sensors, operations, proprietary data, and AI.

What Happened

WindBorne announced the Series B on August 5, 2026, after a year in which the company says both revenue and its active balloon constellation tripled. Khosla Ventures, which led WindBorne's earlier pre-seed and Series A investments, returned as a co-lead. Galvanize joined as the other co-lead, while TransLink Capital, Lux Capital, and existing investors also participated.

The official announcement puts total funding at more than $62M. WindBorne previously announced a $6M seed round led by Footwork in June 2023 and a $15M Series A led by Khosla Ventures in June 2024. TechCrunch reported that the new financing values WindBorne at $250M after the round.

The capital will expand Atlas, increase compute capacity for the next generation of WeatherMesh models, support new weather intelligence products, and help replace part of the balloon network's satellite communications with a mesh radio system. WindBorne also plans to grow its commercial team as it moves beyond a customer base anchored by government agencies.

Why the Data Layer Matters

AI weather models are improving quickly, but models still depend on observations that describe the atmosphere's current state. Sparse measurements, especially over oceans and remote regions, limit what even sophisticated forecasting systems can see. WindBorne's thesis is that better forecasting begins by collecting observations other systems miss.

Atlas is designed for that job. WindBorne's Global Sounding Balloons can remain aloft for weeks, adjust altitude to navigate through different wind currents, and collect repeated vertical profiles of pressure, temperature, humidity, position, altitude, and wind. A 2026 technical paper by WindBorne researchers reports that the current sensor package has flown on thousands of balloons across more than 1M total flight hours, with external comparisons against radiosondes and established reanalysis datasets.

That hardware becomes more valuable when it feeds the model directly. WeatherMesh combines proprietary observations with other data sources, giving WindBorne a feedback loop that a software-only forecasting company cannot easily reproduce. Each expansion of the sensing network can improve the data available for training and forecasting, while model performance can guide where new observations matter most.

From Government Contracts to Commercial Demand

Government customers gave WindBorne a practical route into a difficult market. The company's observations are assimilated into NOAA's Global Forecast System, and WindBorne works with the U.S. Air Force and U.S. Navy on data and forecasting programs. These are buyers with established weather workflows and a clear reason to pay for better observations.

Commercial expansion is the harder test. Weather data creates value only when a customer can connect a forecast to a decision involving inventory, routing, energy demand, crop exposure, emergency preparation, or financial risk. Trading firms are among WindBorne's early private-sector users because weather can move commodity prices and operating assumptions quickly.

That makes the go-to-market investment as important as the balloon count. WindBorne has to turn technical forecast improvements into products that fit how businesses make decisions. The Series B gives the company room to build those commercial systems without stepping away from the infrastructure advantage that made the opportunity credible.

The Leadership Behind the Build

WindBorne grew out of a Stanford balloon research project begun in 2015 and became a company in 2019. CEO John Dean, CTO Andrey Sushko, and CPO Kai Marshland are the current co-founders and executives most closely associated with the company's public leadership. Historical co-founder Paige Brown was part of the early team that turned the student project into a venture-backed company.

The team's background explains the architecture. WindBorne is not outsourcing the hard physical layer and hoping software can paper over the gaps. It designs and manufactures the sensing systems, operates a global fleet, manages the data pipeline, and develops the AI models. That is a much more demanding company to build, but it also creates more places for technical learning to compound.

What the Series B Signals

The financing reflects investor confidence in a specific kind of AI company: one with differentiated data produced by infrastructure it controls. Foundation models are becoming easier to access, and forecast architectures will continue to spread. A proprietary observation network is harder to copy, particularly when it requires manufacturing, international launch operations, regulatory coordination, and years of accumulated atmospheric data.

WindBorne still has to prove that more accurate forecasts can become a large commercial business. The weather industry is full of technically impressive systems whose value becomes harder to capture once the forecast reaches a customer's workflow. The company's reported revenue growth is encouraging, but the next stage depends on converting technical advantage into repeatable products and durable customer behavior.

The $37M Series B gives WindBorne more than capital for 600 additional dots on a map. It funds an attempt to make atmospheric observation and AI forecasting function as one infrastructure layer. If that integration keeps improving decisions before weather becomes damage, delay, or financial loss, WindBorne will have built something more valuable than a better forecast. It will have built a system that makes uncertainty more actionable.