GeoNet Connects Over 22,000 GNSS Reference Stations for Robot and Drone Positioning Correction

GeoNet (GEODNET) connects more than 22,000 global navigation satellite system (GNSS) reference stations to provide positioning correction data for robots, drones, and other devices. The network is positioned as infrastructure supporting AI systems that move in physical environments by improving location accuracy.
The GeoNet solution page states that the global community operates over 22,000 reference stations with correction data latency under one second. While centimeter-level accuracy is claimed, independent verification of actual performance based on region and reception environment has not been confirmed.
Reference stations collect satellite navigation signals to generate correction values, which are then transmitted to devices. Devices apply these values to reduce positioning errors in satellite signals. GeoNet's technical documentation describes the technology as applying real-time kinematic (RTK) corrections to GNSS carrier phase measurements for centimeter-level accuracy.
Correction data is provided via API and NTRIP methods, targeting drones, robots, autonomous vehicles, precision agriculture equipment, and surveying devices. The positioning correction network does not guarantee safe operation or commercialization of devices; actual effectiveness varies based on device type, communication environment, and operational method.
GeoNet adopts a decentralized physical infrastructure (DePIN) structure where participants install reference stations and contribute to the network. According to project documents, reference station operators receive GEOD tokens for their network contributions. The GeoNet materials state that 80% of data revenue is used for token buybacks and burns, though the applicable revenue scope and actual execution scale by period have not been verified.
Crypto Briefing reported on October 2, 2026, that GeoNet's enterprise clients exceed 150 with Q3 revenue reaching $2.8 million (approximately 38 billion won). However, official financial data and reporting criteria supporting these figures have not been confirmed, making it difficult to verify as actual performance metrics.
Reference station distribution by country varies across GeoNet's public pages, showing over 150 countries on one page and 174 countries on another. It is unclear whether the differing numbers reflect different timeframes or reporting scopes based solely on available data.
PitchBook reported that $3.49 billion (approximately 4.7464 trillion won) flowed into robotics and physical AI sectors in the first half of 2026, a trend previously covered by this publication. However, reference station scale and industry investment trends alone cannot determine GeoNet's field deployment or recurring revenue.
Korean Source
This article is an English localization of a Korean-language crypto news report. Original headline: 2만2000개 기준국 연결한 지오드넷, 로봇·드론 위치 보정