Machines don't just need to know what they're seeing. They need to know exactly where they are. That's where $GEOD becomes interesting. Standard GNSS can leave you with errors of several meters. That's fine for humans. It's a problem for a drone landing on a pad, an autonomous tractor following crop rows, or a robot navigating a construction site. @GEODNET turns ordinary rooftops into a global RTK correction network. Here's how it works. Anyone can deploy a multi-band GNSS receiver, mount it with a clear view of the sky, and connect it to the internet. The fixed station continuously monitors satellite signals and calculates positioning errors caused by atmospheric conditions, satellite orbit and clock errors, multipath, and other factors. Those corrections are streamed through GEODNET in real time. Quality scoring then evaluates factors like uptime, signal strength, multipath, and station stability. Operators earn $GEOD based on data quality and location demand, with incentives designed to push coverage into underserved areas. A rover can then pull those corrections over a low-bandwidth connection and move from meter-level positioning to roughly 1–2 cm accuracy. That rover could be: • A drone • An autonomous tractor • A robot • A vehicle • A surveying system And this is where the bigger Physical AI thesis starts to make sense. Cameras, LiDAR, and IMUs are excellent at understanding relative surroundings. They can tell a robot there's a wall 1.3 meters away. But they don't provide a permanent global coordinate that every machine can share. Without that absolute reference: Maps drift. Fleets struggle to share the same coordinate system. Precise landing and docking become harder. Long-term spatial memory becomes unreliable. GEODNET provides that missing global precision layer. A shared, real-time reference for where things actually are. The network already has 20K+ stations across 150+ countries, with a long-term target of roughly 100,000 stations. And unlike a purely speculative infrastructure project, it already generates enterprise revenue, with 80% of that revenue used for $GEOD buybacks and burns. The thesis is bigger than GPS. As Physical AI moves from screens into the real world, robots will need a reliable way to understand not just what is around them, but exactly where they are within it. Cellular networks became invisible infrastructure for smartphones. Precision RTK could become the same for the robot economy. One rooftop station at a time.