Google Research last week introduced Geospatial Reasoning, an advanced research initiative that combines generative AI with new foundation models tailored for geospatial analysis—covering remote sensing, population dynamics, weather, infrastructure, and trajectory prediction, according to a company blog.
The initiative leverages Google’s $GOOG mapping data, including imagery from Maps, Earth, and Street View, the platform enables natural-language queries processed by Gemini to generate insights and visualizations—such as identifying flood zones, detecting solar panel clusters, or spotting impassable roads.
The tools are now accessible to such companies as Airbus, Maxar, Planet Labs, and WPP. However, such competitors as NVIDIA NVDA -0.23%↓, Niantic $, and startups such as Felt are also advancing geospatial AI—developing projects from Earth-scale simulations (Earth‑2) to AI-powered custom mapping tools for environmental monitoring.
Felt, a California-based startup, recently raised $15 million to develop AI-driven geospatial mapping. The company contends that its users can create custom maps and dashboards with AI by instructing it with prompts.
Felt’s platform, which already has been adopted by energy companies, first responders and insurers, features software that turns location data points into map visualization that is tailored to a specific need—which could include monitoring equipment and planning projects.
