How did the topology of the Bitcoin Lightning Network change?

Who runs the nodes and who dominates the network?

A historic analysis of the topology of the Lightning Network

The Bitcoin Lightning Network consists of nodes and channels. Each channel has a capacity in sats which equals the amount of bitcoin locked. Nodes publish channel_announcement messages to the Lightning Network. Other nodes recieve those messages and construct their own topology of the network. When facilitating a payment a node needs to calculate the path of the payment. This step requires the node to know the topology of the network.

The Bitcoin Lightning network is a distributed network meaning that there is not a single source of trust, but every node has their own subjective view The protocol has introduced three types of so called gossip messages that nodes can publish to let other peers know of updates of the topology.

The ln-history platform has collected those gossip messages and persists them in an index database. Utilizing the platform historic snapshots of the Lightning Network can be optained very quickly.

Node Count Over Time

The node count has grown rapidly during 2021 and 2022. In the following years many node operators have decided to shut down their nodes. In 2025 we can see that the node count stays above 10000.

Network Capacity Over Time

The number of locked bitcoin on the other hand has not seen a big drop as the node count. When comparing the locked bitcoin count with the equivalent value in USD the total capacity of the Lightning Network is at an all time high. Taking into consideration that prices are denominated in USD, the Lightning Network has never held more liquidity (denominated in USD).

Network Diameter Over Time

When trying to see how the Lightning evolved we can take a look at the diameter of the network. The diameter of a graph is the longest shortest path. We can examine that this value has remained constant at 9 nodes. This means that the longest possible payment would need to take 9 hops.

Median Payment Fee Over Time

Nodes set up a fee policy for each outward channel direction. This information gets propagated using the channel_update message. The fee policy of a channel consists of the fee_base_msat and the fee_proportional_millionths. We simulate 10000 payments of 100000 sats and calculate the median fee of the payments. The fee is stable below 7 sats.

Altruistic nodes

The graph shows that the share of channels with a fee_base_msat set to 0 is at slightly lower than 45 %. If we now check how many channels have set fee_proportional_millionths set to 0, we get a value of around 8.5 % of channels.

Combining both settings we can can see the following.

This graph looks almost identical to the previous graph, indicating that if nodes set the fee_proportional_millionthsto 0, the fee_base_msat is also set to 0.

We can examine that the share of altruistic nodes is around 8.5 %.

Underlying Networks used

The Lightning Network can use both the clearnet with IP based routing as well as the more private Tor network with onion based routing.

We can see that by just looking at the number of nodes, much more nodes run behind the Tor network. But this graph might be misleading as the next one shows.

Most of the liquidity is managed by nodes using clearnet. The reason for that is that information speed is much higher in cleanet compared to Tor network.

Top 10 nodes by channel count

Lastly we want to examine which nodes dominate the Lightning Network. A dominant node is a node with many channels. We can see that enterprise nodes such as ACINQ, 1ML, WalletOfSatoshi or Kraken lead the list.

Final words

Looking at the analyzed data it seems that the Lightning Network is maturing and professional. There is still a fair amount of node operators running and supporting the network altruistically. Nonetheless enterprises have the highest stake in the network.