The growth of the Lightning Network

10 Oct 2021

The growth of the Lightning Network

In our previous article, we illustrated a stylized Lightning Network with 5 participants conducting 20 transactions to visualize the inner workings of Lightning. In reality, the Lightning Network consists of near 100,000 users (per August 2021).

This article covers Lightning's growth and tendencies since the launch of the Lightning Network mainnet in 2018.

This blog post is an extract from our latest report: “The State of Lightning”. Powered by OpenNode - Supported by Lightning Labs

Network capacity

In its nearly three-year history, the Lightning Network has grown to hold a bitcoin capacity of almost 3,000 BTC, with about 17,000 nodes operating on the network and 73,700 unique channels connecting nodes from across the globe. As of late September 2021, the public bitcoin capacity on lightning equals more than $120 million, with users across the world finding value in the option to conduct near-instant small payments at very negligible fees.Lightning Network: BTC Capacity

Preview

The Lightning Network has seen two distinct phases of growth. We will cover these periods in detail in the next paragraphs. The first growth phase began shortly after the launch of Lightning's mainnet and can be categorized as a very experimental phase. A slow period in terms of user adoption followed in the latter half of 2019 and 2020. However, while the adoption rate was slow, developers worked full throttle towards improving the network and creating tools and services, setting the stage for a new adoption phase. In 2021, several forces created the perfect storm to ignite Lightning’s second major adoption phase.Let’s start with the first growth phase. We characterize the first adoption phase of the Lightning Network from June 2018 to May 2019 as Lightning’s experimental age. The period saw the emergence of some massive nodes, drama, and a mythical community-driven effort to spread awareness of the capabilities of Lightning, sparking the Lightning Network to grow from a BTC capacity of 20 BTC to a peak of 1104 BTC in 11 eventful months. The first phase is illustrated below.Lightning Network: BTC Capacity, May 2018-June 2019

Lightning Network BTC Capacity, May 2018-June 2019

Preview

The first significant growth of Lightnings BTC capacity occurred in the summer of 2018. The increase was alarming to many since it was driven by one massive node entering the network accounting for more than 50% of the capacity on the network. The infamous node was named LN.SHITCOIN.COM and was owned by Andreas Brekken, who created the node motivated by experimenting with Lightning and writing down his experiences as he went. Brekken wrote four medium posts about his experiences. We recommend giving these posts a read, as they paint a picture of the state of lightning in June 2018. The second stage of growth in 2018 was relatively similar to what we saw with Brekken in the summer. In November, LNBIG entered the market, and he went in deep. However, contrary to Brekken, LNBIG did not garner as much attention, and he has also remained active in the network ever since. His nodes are, to this day, acting as important hubs on Lightning.The third growth phase of these 11 months occurred in January 2019. It can be summarized as a community-driven effort leading the Lightning Network to gain massive attention both within and outside of the bitcoin community. In January 2019, Twitter user Hodlonaut started an experiment where he sent 100,000 sats to the first person he trusted who responded to his tweet. The transaction was under the condition that the recipient added 10,000 sats and then sent it to a new user. This chain of transactions was renamed “The Lightning Torch” and quickly gained attention on Twitter. The torch ended up being passed around by 282 unique participants and was in the hands of many influential people, shedding more attention to this nascent network. The torch was in the hands of Jack Dorsey (Twitter/Square), Reid Hoffman (Former: PayPal COO and LinkedIn co-founder), Fidelity, plus many, many important bitcoin and crypto personalities. Additionally, the torch was sent to an Iranian bitcoin user, a country sanctioned by the West. In total, the torch was passed 282 times, ending with 0.0429 BTC donated from Torkel Rogstad to Bitcoin Venezuela. Concurrently, the charity also received a 0.41 BTC donation from the bitcoin community. The Lightning Torch ended up as a very successful experiment, leading to a substantial donation and increased awareness of Lightning, likely contributing to pushing the BTC capacity up towards its peak of 1104 BTC.Throughout 2021, The Lightning Network has once again gained public attention at scale. The 2021 growth is a story of the stars aligning, as seen below. Lightning Network: BTC Capacity per September 30th, 2021

Lightning Network BTC Capacity per September 30th, 2021

Preview

On June 5th, El Salvador announced that they would implement bitcoin as legal tender in the country. The Lightning Network is a crucial ingredient in making this transition economically feasible for Salvadorans. From May 2021, the overall interest in trading bitcoin declined as the bullish momentum from Q1 disappeared. Additionally, more services implemented SegWit transactions. This led on-chain transactions to decline substantially, in turn leading the mempool to clear and on-chain transaction fees to plummet, creating the perfect moment to open new channels on the Lightning Network. Low on-chain fees and high awareness due to the El Salvador news created the perfect storm for Lightning, evident by the blistering growth rate of the BTC capacity. From January 1st, 2021, to September 30th, 2021, the BTC capacity on the Lightning Network grew by 181%, from 1058 BTC to 2968 BTC.Until now, we’ve exclusively focused on the bitcoin capacity of the Lightning Network to highlight the growth and history of the network. The increased bitcoin capacity of Lightning has developed in tandem with a massive influx in the public channels on the Lightning Network. From May 1st till September 30th, the number of public channels grew by 80% from 39,281 channels to 70,583 channels, amid a period of unusually low on-chain fees. This does not reflect the complete picture of the number of channels on the Lightning Network, as many channels are private and not broadcast to the entire network. Thus, the actual channel count is likely far higher than what’s reflected by the public data. BitMEX Research estimated that 27.8% of all Lightning channels were private in January 2020. To our understanding from talks with industry-leading experts, the share of private channels on Lightning could be even higher now. Thus, the BTC capacity, channel count, and node count from public data is most likely a significant underestimation of the current size of the Lightning Network.On-chain Fees vs. Public Lightning Channels

On-chain Fees vs. Public Lightning Channels

Preview

The chart above illustrates the relationship between the on-chain fees of bitcoin and the opening of Lightning channels. The chart demonstrates that the opening of new Lightning channels has coincided with lower transaction fees. Interestingly, the number of public channels on the Lightning Network has tended to flatten in periods with rising on-chain fees, such as in April, in the middle of May, and the start of July. Since the middle of July, on-chain fees have been remarkably low, while the growth of public Lightning channels has accelerated.The Lightning Network is an alternative to on-chain transactions, and usage should be in higher demand in periods with high on-chain fees on the Bitcoin blockchain. However, at the time being, we are in the latter stage of Lightning’s proof of concept phase. In this period, entities with a long-term positive outlook on Lightning will seek to scale into the Lightning Network and open channels when on-chain fees are unusually low, such as the fee environment seen in bitcoin since June. By opening channels during times of low on-chain fees, these entities are positioned to interact with the Lightning Network in routing transactions and deliver services in periods where on-chain fees climb again. The Lightning Network will likely see increasing adoption in a scenario where BTC grows further, more merchants accept bitcoin, and the on-chain fees reach a sustained high level. In such a scenario, entities will likely find it economically reasonable to open new Lightning channels, so they can conduct indefinite amounts of low fee transactions without having to worry about the cumbersome on-chain fees. In this scenario, routing fees might also provide an additional appealing revenue stream, further enhancing the fruitfulness of being a well-connected node. This further incentivizes the opening of more channels in order to route payments from users seeking to pay for groceries, trade on exchanges, or make payments to friends in an efficient manner. All metrics suggest that the adoption of the Lightning Network is growing with nodes, channels, and capacity, all trending up. Considering that the Lightning Network is a second-layer solution aiming to scale bitcoin, it makes sense to contrast the growth of the Lightning Network with how it helps to scale bitcoin. Below, we attempt to contextualize how the activity on Lightning contributes to reducing the mempool and saving blockspace.

Lightning scales Bitcoin

The Lightning Network is a powerful second-layer scaling solution for bitcoin, as illustrated below, depicting how Lightning transactions free up block space on the Bitcoin blockchain. The chart is based on estimated transaction data from commonly used wallets on the Lightning Network. We count all transfers to and from wallets in our estimate of the number of Lightning transactions and do not count routing transactions. We are confident that our estimate is close to the actual transaction numbers from these wallets but emphasize that our estimates are, indeed, estimates. Further, the chart assumes that an average Bitcoin block consists of 2000 transactions.Estimate of Blocks / Blocks Days Saved per Month Through Lightning Transactions

Estimate of Blocks - Blocks Days Saved per Month Through Lightning Transactions

Preview

Per our Lightning Network transaction estimates, we find that the Lightning Network processed 663,000 transactions into and from commonly used wallets in September 2021. This number of transactions would require 332 full on-chain bitcoin blocks, equaling 2.30 days of bitcoin blocks. Thus, through transactions on the Lightning Network, the on-chain mempool was spared from transactions that would have absorbed a lot of block space. In essence, the bitcoin network gained 2.30 days worth of transactions in September, caused by the growing adoption of the Lightning Network. This helps scale Bitcoin and contributes to lowering the on-chain fees.These estimations should, of course, be interpreted with a grain of salt. First and foremost, a majority of the transactions occurring on the Lightning Network are micropayments. These transactions would not be economically feasible as on-chain transactions and would thus not be a part of the mempool in the first place. Microtransactions are one of the magical features enabled by the Lightning Network. A whole new subset of opportunities emerges as micropayments with bitcoin are enabled. Thus, while Lightning contributes to offsetting some of the on-chain pressure for transactions, the 332 blocks represent an emerging new form of spending bitcoin. This new form of spending bitcoin has led to the creations of hundreds of projects, all working within the Lighting Network utilizing the power of microtransactions to create useful, immersive, and creative tools for users to enjoy.

A centralized network?

One of the key ingredients in the Lightning Network is the payment channels between nodes, allowing users to transfer funds through a series of nodes by using channels, ending up at the desired end-point destination. Lighting’s elegant routing feature does not come without a cost. The network structurally rewards the most well-connected nodes (hereby hubs), where hubs are an essential part of the network. This architecture makes Lightning transactions structurally more centralized than on-chain bitcoin transactions. Nevertheless, this structure is a feature of Lightning, and without this feature, the network would lose its finesse. While the design, to a certain extent, leads to centralization, users are free to open channels to avoid the central hubs. Hubs are merely a very convenient feature of the Lightning Network, making sure the network is connected. Let’s take a deeper dive into the importance of certain nodes and channels on the Lightning Network. Below, we illustrate the number of cut nodes and channels on the Lightning Network. A cut channel is a channel between two nodes that connects different components of the network. This channel's removal would prevent other nodes from having a path. A cut node is the same idea, except it's a crucial node instead of a channel. In the Lightning Network, cut nodes and channels can be seen as weak points of the network as their removal would entirely disconnect two parts of the network. Therefore, they should ideally be minimized. This is especially true for cut nodes with many channels or cut channels of a large capacity, as the network heavily relies on them to function successfully. Lightning Network Cut Nodes & Cut Channels

Lightning Network Cut Nodes & Cut Channels

Preview

In the chart above, we see that the percentage of cut nodes has trended downwards since August 2020, currently sitting at approximately 8%. This illustrates that there are some critical nodes in terms of routing on Lightning at the moment. Still, the entrance of new hubs has contributed to reducing the vulnerability of specific nodes being cut. The percentage of cut channels has also been trending down since May 2021, currently sitting at 10%. Lightning Network Clustering

Preview

The chart above illustrates the average clustering of the Lightning Network. A value of 1 indicates that nodes form cliques. In layman’s terms, a clique is a closed group of nodes, not reaching the broad network. A value of 0 means that the average node is a hub, with none of its peers being connected. The average clustering coefficient has trended down since February 2019, aligning with the growth of cut channels in the same period. The declining clustering coefficient indicates that peers on the Lightning Network make increasingly more rational decisions when opening new channels. In general, channels now tend to be opened with nodes that are connected to a subset of nodes, which the channel creator has not opened a channel with. The opening of this channel gives the channel creator a route to this subset of nodes and the ability to route transactions to these nodes. Given these nodes also tend to be connected to another subset of nodes, the channel creator gets the ability to route a transaction more efficiently on the broader network. While this increases the importance of hubs and increases the percentage share of cut channels, it also contributes to making the Lightning Network a more efficient and well-connected payment network.In sum, the Lightning Network is more centralized than the Bitcoin blockchain when accounting for hubs and payment paths. However, this is a desired feature of the Lightning Network, enabling users to find a cost-efficient path to the recipient of a transaction. Users always have the option to open their own channels, both private and public, without relying on the large hubs. Thus while the hubs in some sense lead to increased centralization of the Lightning Network, the network itself is not inherently centralized, it’s merely a property of the network that users may opt to take advantage of, or avoid utilizing, and an interested and motivated party can also always launch a new hub node.

This blog post is an extract from our latest report: “The State of Lightning”. Click the link below to download the full report.