EXPLORINEER
en

Archive — US — techcrunch

This AI weather startup is out-forecasting government agencies

WindBorneSystems, an AI weather startup founded in 2019, releases WeatherMesh 6, a tool providing more frequent and accurate weather forecasts than traditional models used by government agencies like ECMWF. Enhanced data ingestion from about 400 weather balloons drives its success, offering forecasts every hour at a 3 km resolution, significantly improving accuracy compared to existing services.

Key points

  • WindBorneSystems was founded by Stanford students in 2019, initially aiming to sell weather data.
  • The release of deep learning weather forecasting models in 2022 led WindBorne to build their own predictive model.
  • WeatherMesh 6 offers forecasts every hour (versus every six hours), achieving higher accuracy and resolution (3 km) on surface temperature.
  • Traditional weather forecasting relies on complex physics models, while AI models are increasingly integrated but often lack resolution.
  • WindBorne utilizes 400 balloons for data collection, which feeds directly into their models, increasing forecast reliability.
  • Despite its success, WindBorne is cautious about its business model, focusing more on model refinement than immediate monetization.
  • After a near-miss incident with an aircraft, WindBorne has implemented safety measures like transponders on its balloons.

Why it matters

  • AI weather models are becoming increasingly important as climate change intensifies, demanding more accurate predictions.
  • The competition between traditional meteorological organizations and startups highlights a growing trend in the tech industry towards integrating AI and data analytics.
  • The rise of AI in weather forecasting reflects a broader trend in 2025 IT Trends where AI applications are utilized in diverse sectors for efficiency and precision.

WindBorneSystems' advancements in AI weather forecasting mark a significant shift in how weather data is gathered and utilized, pushing boundaries beyond traditional models and spotlighting the potential of AI capabilities in the industry.