Where the coordination layer already runs
A production test surface most infrastructure companies never get: 350,000+ cumulative device instances across 140+ countries, grown through first-party and third-party applications on product pull alone. Every install is a mesh node, every app built on the SDK extends coverage for every other app, and the footprint exercises the protocol under real radio, power, and network conditions every day.
The network is useful because applications can act on it
The map is the footprint. These are the shipped platform capabilities that turn nearby devices into working infrastructure, each linked to its implementation context.
Phone-to-phone communication
Keep direct and group traffic moving when towers, backhaul, or power fail.
→Shared state that converges
Replicated documents merge deterministically after offline edits, with your cloud as the system of record.
→Verification with zero connectivity
Authenticate people and devices face to face with OfflineID.
→Services without a central registry
Advertise, discover, and invoke capabilities across nearby peers.
→A record that survives the outage
Opt-in telemetry the SDK uploads itself, with identifiers scrubbed by default.
→Every integration adds another node
Ship the React Native SDK on iOS and Android over the same Rust core.
→The cloud control path is the fragile part
Outages, blackouts, and shutdowns take connectivity away from more places every year, which is why coordination has to keep working locally and reconcile when the link returns. The 2025 numbers set records for the third consecutive year.
internet shutdowns across 52 countries in 2025, a record for the third straight year (Access Now)
network disruptions documented in 2025 (Cloudflare)
economic losses from shutdowns in 2025, with 4.62 billion people affected (Top10VPN)
LEO satellite shutdowns across 7 countries in 2025, up from 4 in 2024.
Where the mesh holds
Activity lights up where infrastructure fails: storms and floods, rolling blackouts, saturated towers, and sites that never had backhaul to begin with. The map above is the cumulative deployed footprint, with 11,500+ local networks showing peer activity in the trailing 30 days, the same mesh that public sector and event deployments run on.
Why the footprint matters
Every important physical system is becoming a computer, and most still depend on a distant cloud to discover one another, establish trust, exchange state, and prove what happened. That architecture fails where those systems are most valuable: in motion, in the field, behind constrained links, during outages.
The protocol was hardened on this footprint before the first enterprise conversation. Identity, delivery, recovery, and route selection have been exercised at a scale that produces real operating evidence, and that evidence is what production accounts build on.