GEODNET: The Rails For Physical AI

Jul 27, 2026 | Chainletter |

Executive Summary

GEODNET is the world’s largest RTK network, a system of ground stations that corrects ordinary satellite GPS signals down to centimetre-level accuracy, essential for the autonomous machines that cannot navigate on the metre-level errors an unaided phone or car tolerates. 21,429 community-owned stations across 11,469 cities in 170 countries deliver that accuracy at roughly one forty-fifth of incumbent prices. It began not as a token in search of a use but as a technology company in Arizona whose founders adopted a blockchain for the soundest reason: the problem, paying thousands of station hosts worldwide, could not be solved any other way. Four years on, the network has been built and financed entirely by its community. Moreover the industry incumbents – companies like Trimble, Hexagon, and Topcon – have begun renting the network rather than competing with it, the clearest signal that replication is harder than partnership.

The tokenomics are one of the cleanest we cover. Customers pay in fiat; 80% of revenue buys GEOD on the open market and burns it, permanently, verifiably, by rule. Implied revenue has grown roughly 50% in two quarters to an annualised run-rate above $10 million, and the burn now retires close to 10% of circulating supply a year, the equivalent of a 9% mandatory buyback yield against a $91.3 million market capitalisation. In June, the network’s scheduled annual halving (GEODNET calls it the “halving”) cut token issuance in half; burn and issuance now sit at parity, and with issuance halving again every June while revenue compounds, the network is on the cusp of sustained net deflation, arriving just as its demand curve steepens. The robotics and drone economy runs on exactly what this network sells, and its Western manufacturers are being armed with a GEODNET-native system-on-chip built by the founder’s adjacent equity venture, a demand engine for the token rather than a competing claim on it.

We identify a number of risks. Revenue must keep compounding: the burn that drives the token’s scarcity is funded entirely from that revenue, and if growth stalls, so does the burn. Regulatory and manufacturing timelines sit outside GEODNET’s control. Token liquidity remains thin. And mining rewards, new GEOD paid to operators for real work, continue on a schedule that shrinks every year but never fully stops. But the asymmetry is unusual and the thesis is observable in real time, on-chain.

This report sets out the story, the mechanism, the numbers, and the triggers we are watching.

GEODNET – The Rails For Physical AI