PangeAI

PangeAI is an early stage AI agent platform for Earth data, built to make geospatial intelligence as easy as asking a question. Its agent, Geoffrey, sits as an agentic layer over satellite imagery, vector geometries, terrain and coordinate systems, and lets an analyst pose a physical world question in plain language, for example which of 400 sites flooded last month or where the optimal site for a 50 megawatt solar farm in Nevada is while avoiding protected species. The Geospatial Analyst Agent ingests messy, unstructured Earth observation data, automatically selects the right datasets, and returns an answer in minutes rather than the weeks a GIS team would take, with dynamic layering that generates interactive simulations so users can toggle variables like climate risk or supply chain routes and see instant impact assessments. PangeAI is focusing first on verticals where spatial data is business critical and hard to work with: insurance, with next generation hazard scoring for flood, wildfire, wind and earthquake down to the individual property in minutes; energy, with agents that select renewable sites and monitor transmission corridors; and natural capital, with agents that detect land use change and forest degradation for asset managers. Founded in 2025 by Stanford trained experts Johanna von der Leyen, chief executive, and Marek Miltner, chief technology officer, with offices in San Francisco, Berlin and Prague, the company emerged from stealth in December 2025 with roughly $1 million in early funding led by MITON, alongside RTP Global, Tensor Ventures and Plug and Play and a research grant from Stanford, and is running pilots including with one of the largest Japanese insurance groups. The product is in early access.