Every day, Earth generates more than 100 terabytes of satellite imagery, offering a wealth of geographic information. However, transforming that raw data into practical insights remains a major challenge. Answering seemingly simple questions—like how many fire breaks exist in California and how they’ve changed since the last fire season—still requires extensive manual work or expensive custom models. This is the problem LGND is working to solve.
Founded by Nathaniel Manning and Bruno Sánchez-Andrade Nuño, LGND is a startup using advanced AI and spatial embeddings to make geographic data far more accessible and efficient to analyze. Traditionally, analyzing fire breaks or other landscape features required data scientists to build custom models trained on massive datasets—an approach that could cost hundreds of thousands of dollars per use case. According to Manning, “You probably sink a couple hundred thousand dollars… to try to create that dataset, and it would only be able to do that one thing.”
LGND aims to change this by offering a flexible, scalable platform that generalizes geospatial understanding through a method called vector embeddings. Rather than working with pixels or basic map features like points, lines, and areas, LGND’s embeddings summarize spatial data into compact, high-dimensional representations. These make it easier to spot patterns or relationships—like identifying a fire break based on visual clues such as absence of vegetation or minimum width—without training a new model every time.
“We’re not looking to replace people doing these things,” says Nuño. “We’re looking to make them 10 times more efficient, 100 times more efficient.” These efficiencies come from pre-processing spatial data in a way that frontloads most of the computation, allowing users to query it rapidly and with greater flexibility.
To support its mission, LGND recently secured $9 million in seed funding led by Javelin Venture Partners. Other backers include AENU, Space Capital, Clocktower Ventures, MCJ, and angel investors such as Keyhole founder John Hanke and Salesforce executive Suzanne DiBianca. With this funding, LGND is developing an enterprise application and a robust API, enabling both companies and technical teams to tap into geographic insights with ease.
One envisioned use case is an AI travel assistant that could process complex spatial queries—like finding a rental near good snorkeling, on a white sand beach, with low seaweed in February, and no nearby construction. Such queries would be extremely time-consuming using conventional geospatial methods but are made tractable through LGND’s embedding system.
The ultimate goal is to democratize spatial intelligence across sectors like environmental monitoring, infrastructure planning, and logistics. With the geospatial analytics market estimated near $400 billion, LGND is positioning itself as a foundational provider—what Manning calls the “Standard Oil for this data.”
By redefining how geographic information is stored, queried, and understood, LGND is building the tools to turn Earth’s massive flow of spatial data into a new layer of intelligence—scalable, fast, and open to a far broader set of users than ever before.
www.techcrunch.com/2025/07/10/lgnd-wants-to-make-chatgpt-for-the-earth/