BTC: Reachable Nodes by Country (Share)
Bitcoin Reachable Nodes by Country (Share): daily data through 9 Nov 2026.
The share of reachable nodes in our crawler's sample by country, from the largest countries down.
Why use it. Where nodes run says how exposed the network is to one government or one regional outage. How to read it. The bands add up to 100%. A rise or fall in one country's band after a change in its internet or its laws is the kind of shift this is for.
Calculation. Each reached address is looked up, while it is counted, in DB-IP's free IP to Country Lite and IP to ASN Lite databases (CC BY 4.0, db-ip.com), which map address ranges to the country they are used in and the network that announces them. Lookups keep only the country and network; addresses are not stored or published. Countries shown are the largest of the last 30 days. Observed by CryptoKeyhole's own crawler. Twice an hour it re-contacts every node in its sample (every Bitcoin node it reached over IPv4 in the last 24 hours), completes the version handshake and disconnects. It reads only what a node announces about itself in that handshake: its software name and version, its service flags, its protocol version, whether it wants transactions relayed and its fee filter. Addresses are used only to connect and are never stored or published. A node counts once per snapshot (by address and port); an hour or a day shows the mean of its snapshots, and the latest hour and day are partial. Reachable nodes are a sample, not every node. Nodes that do not accept incoming connections (most nodes), and nodes reachable only over Tor, I2P, CJDNS or IPv6, are not counted, because our host connects over IPv4 only. Our crawler has also not found every IPv4 node that accepts connections: it finds nodes from the addresses our own node knows and the addresses other nodes announce, and our Internet connection lets it try only a few new addresses a second, so the sample grows over its first days. Each snapshot re-contacts every node in the sample (up to three tries within minutes) and is published only if it reached at least 90% of them within 25 minutes and the sample holds at least 1,000 nodes. Software names, versions and flags are self-reported and cannot be verified. Collection began on 22 October 2026; no earlier values are reconstructed.