LGND Wants to Build “ChatGPT for the Earth” With $9M in New Funding
Startup Aims to Unlock the Power of Geospatial Data for Everyone
LGND, a San Francisco-based startup, is on a mission to make the world’s geospatial data as accessible and searchable as text—essentially, building a “ChatGPT for the Earth.”
- Every day, satellites collect 100 terabytes of imagery about the planet, but transforming that raw data into useful answers remains a complex and expensive challenge.
- For questions as vital as “How many fire breaks does California have, and how have they changed since last year’s wildfires?” traditional analysis still relies heavily on time-consuming manual review or costly, task-specific AI models.
Making Spatial Analysis 10–100x More Efficient
LGND’s founders—CEO Nathaniel Manning and chief scientist Bruno Sánchez-Andrade Nuño—want to break this bottleneck, making geospatial data queries faster, cheaper, and vastly more efficient for everyone from disaster response teams to real estate companies.
- “We are not looking to replace people,” Nuño says. “We’re looking to make them 10 times more efficient, 100 times more efficient.”
- The company’s approach centers on vector embeddings: smart, compressed data summaries that can quickly reveal relationships and patterns across spatial datasets, reducing the need for deep technical expertise or massive computing power.
How Embeddings Transform Geospatial Search
Traditionally, geospatial data exists as pixels or vectors (points, lines, polygons), which are flexible but require significant expertise to interpret.
- LGND’s embeddings distill spatial data into compact representations, capturing the essence of features like roads, rivers, or fire breaks—making it far easier for AI to find relevant answers.
- For example, fire breaks may be rivers, roads, or clearings with no vegetation, and have specific minimum widths. LGND’s technology makes it possible to identify all matching features with a single, simple query.
Practical Applications—and Massive Market Potential
LGND’s technology supports an enterprise app for businesses and an API for developers with advanced geospatial needs.
- Manning envisions a future where companies (or consumers) can ask complex, natural-language questions about the physical world:
- “Find me a three-bedroom rental near good snorkeling, on a white sand beach, with little seaweed in February, and no construction within a kilometer.”
- Building traditional geospatial models for such queries would be expensive and slow. LGND’s embeddings could answer in seconds, not weeks.
Recent Funding and Vision for Scale
LGND just raised a $9 million seed round led by Javelin Venture Partners, with backing from AENU, Clocktower Ventures, Coalition Operators, MCJ, Overture, Ridgeline, Space Capital, and notable angels like Keyhole’s John Hanke and Ramp’s Karim Atiyeh.
- The startup is targeting a slice of the $400 billion global geospatial market, aiming to become the “Standard Oil for this data”—making complex Earth insights as simple to access as chatting with an AI.
What’s Next for Geospatial AI?
With LGND, the goal is to democratize powerful spatial analysis, so more people and organizations can ask—and answer—questions about the Earth with the ease of a Google search or a ChatGPT conversation.
- If successful, LGND could change how industries as diverse as agriculture, disaster response, tourism, and logistics leverage the flood of satellite and sensor data available today.
