LGND wants to make ChatGPT for the Earth

The planet is inundated with data, particularly in the form of imagery captured by satellites, which can amount to around 100 terabytes daily. However, deciphering this wealth of information poses significant challenges. For instance, a critical question for California's economy is: how many fire breaks exist to halt wildfires, and how have these changed since the last fire season? Traditionally, answering such questions required manual analysis of images, a process limited by human capacity. Nathaniel Manning, co-founder and CEO of LGND, highlighted that while advancements in neural networks have improved the situation, the costs associated with creating necessary datasets can skyrocket into the hundreds of thousands of dollars. LGND's goal is to drastically reduce these costs and enhance efficiency. Bruno Sánchez-Andrade Nuño, LGND’s co-founder and chief scientist, emphasized that their mission isn't to replace human effort but to amplify it significantly. The startup recently secured a $9 million seed funding round led by Javelin Venture Partners, with participation from various investors including AENU, Coalition Operators, and several angel investors. At the heart of LGND’s innovation is the use of vector embeddings for geographic data. While current geographic information typically exists in pixels or traditional vector formats, embeddings provide a more efficient method for summarizing spatial data. Nuño explained that these embeddings streamline the computational process, offering users a way to uncover relationships between various geographic points without excessive resource consumption. For example, fire breaks may manifest as roads or bodies of water, each represented differently on a map yet sharing common characteristics. Embeddings simplify the identification of these features by highlighting areas devoid of vegetation and adhering to specific width requirements based on surrounding plant life. LGND has developed an enterprise application designed to assist companies in querying spatial data, alongside an API for users with specialized needs. Manning envisions a future where businesses can utilize LGND’s embeddings to explore geospatial data in innovative ways. For instance, an AI travel assistant could help users find rentals based on multiple specific criteria, such as proximity to snorkeling spots or avoiding construction noise. If LGND can successfully launch tools that make geospatial data more accessible, it stands to capture a significant share of a market estimated at nearly $400 billion. As Manning puts it, LGND aims to be the 'Standard Oil for this data'.