The Snowflake Startup Challenge judges started off this year’s finale by describing what they were looking for in a winner: innovative ideas, technology and products that deliver real business value, and team chemistry, grit and passion to tackle big problems. LGND AI delivered on all fronts, becoming the new 2026 Startup Challenge Champion!
LGND’s goal is to make the entire planet queryable through imagery. Earth observation data has immense potential value — but it’s also expensive and difficult to use, and almost entirely absent from the AI revolution because LLMs are trained on language. LGND builds Large Earth Models, or LEMs, trained on 800 petabytes of pictures of the planet, so that when you ask an agent “where has there been deforestation in the Amazon this year,” you get an answer that’s grounded in imagery rather than relying solely on written reports.
By leveraging the power of the Snowflake platform, LGND is able to work across multiple types of imagery and modalities, giving them more flexibility at the operational level and at scale.
“We didn't just build the model,” said Jeff Albrecht, Head of Engineering at LGND AI. “We built the entire stack. Models, infrastructure, the agents, all powered by a deep integration into Snowflake across services like Snowpipe, Dynamic Tables, and Cortex Search, to name a few. And we built the application on top.”
