Does Running a Lightning Node Still Pay in 2026?

The Lightning Network entered the second half of 2026 looking both healthier than it has ever been and thinner than it has been in years. Public payment volume set records, big exchanges wired themselves in, and for the first time dollars began moving over the same channels that carry Bitcoin. At the same time, the number of machines actually running the network kept sliding, a slow drift that started in 2022 and has not reversed.

That contradiction is the story, and it reopens a question that used to have an easy answer. For most of Lightning’s life, spinning up a node was something a curious Bitcoiner did on a Raspberry Pi over a weekend, part to move sats cheaply and part to plant a flag for a decentralized payment layer. In 2026 the math is colder. A node that routes strangers’ payments is a small business, with revenue, costs, and a break-even line that most operators never cross.

This is a field guide to that business. Where does the money come from, where does it leak out, who is still bothering, and what does a shrinking operator base mean for the run-your-own-node promise that Bitcoiners hold close to sacred? The tension will feel familiar to anyone who has read our guide to how to read a Bitcoin miner’s margins: on both the mining and routing sides of Bitcoin, headline growth and operator economics have been pulling in opposite directions.

The Numbers Do Not Agree With Each Other

Start with the raw figures, and the first surprise is that nobody quite agrees on them. Lightning is a network of payment channels, and a large share of those channels are private, invisible to the public gossip layer that trackers scrape. Estimates of hidden capacity run to twice the visible total or more, according to research from the Lightning analytics shop Spark. So every number below is a floor, not a full count, and different trackers draw the floor in different places.

On the visible layer, public channel capacity topped 5,400 BTC in late 2023, slid to roughly 4,200 BTC by August 2025, then rebounded to a record near 5,600 BTC in December 2025 as exchanges added liquidity, per data compiled by CryptoSlate and Spark. By May 2026 it had settled back near 4,900 BTC across about 41,000 public channels. Convert that at recent prices and the visible network secures a few hundred million dollars, with the true figure likely well over a billion once private channels are counted.

Node counts tell the opposite story. The public node total peaked around 20,700 in 2022 and has drifted down to roughly 17,400 by mid-2026, a steady multi-year slide that Spark’s data pins on consolidation rather than collapse. Even that figure is generous: some node explorers that only count peers actively gossiping public channels show barely a third as many. The gap is not a rounding error. It is a sign of how hard the network is to measure once you leave the marketing dashboards behind.

Volume, meanwhile, is unambiguous. Publicly measured Lightning payments hit about $1.17 billion in a single month in November 2025, up a reported 266% year over year, Spark’s research found. More capacity concentrated in fewer nodes, moving more money: that is the shape of the paradox, and everything else in this guide is an attempt to explain it.

Routing Node vs. Wallet: Know Which One You Are Running

Before the economics make sense, one distinction has to be clear, because most people who say they use Lightning are not running anything that earns a satoshi.

A Lightning wallet spends and receives. It might run a light node under the hood, as self-custodial apps do, or it might lean entirely on someone else’s infrastructure, as custodial apps do. Either way, its job is to let you pay and get paid. It does not, as a rule, sit in the middle of other people’s payments.

A routing node is the opposite. It opens channels, funds them with real Bitcoin locked in a two-of-two multisignature address on the base chain, advertises those channels to the public gossip network, and then forwards payments that hop through it on their way somewhere else. For each forward it collects a small fee. That is the business. To do it well, a routing node needs three things at once: capital sitting idle in channels, liquidity balanced on both sides so payments can flow in either direction, and enough uptime and attention to stay useful. Miss any one and the node earns nothing while its capital does nothing.

Everything that follows is about routing nodes, because they are the part of Lightning that actually carries the network, and they are the part that is disappearing. When commentators say node counts are falling, they mean these machines, not the millions of wallets in people’s pockets. The wallets are multiplying. The routers are consolidating. Keeping the two ideas separate is the first step to understanding why.

Where the Money Comes From

Routing revenue has exactly one source: forwarding fees. When a payment passes through your channel, you charge for the privilege, and the charge has two parts. A base fee is a flat amount per payment, quoted in satoshis and often set to a fraction of one. A fee rate is proportional, quoted in parts per million (ppm) of the amount forwarded. Route a 1,000,000 satoshi payment through a channel priced at 200 ppm and you earn 200 satoshis, plus whatever the base fee adds.

Those numbers are small on purpose. Lightning’s whole pitch is that moving value should cost a fraction of a cent, so the median routing fee across the public network sits at a fraction of a satoshi in base fee plus a proportional rate in the low hundreds of parts per million. You do not get rich on any single forward. You get paid on volume, and volume only comes to nodes that are well positioned: connected to peers that send and receive a lot, with liquidity pointing in the direction payments actually want to go.

That positioning is the hard part, and it is why revenue is lumpy. A channel can sit dead for weeks, then route heavily for a day when a big merchant or exchange rebalances. Operators who treat fees as a set-and-forget setting tend to leave money on the table or, worse, price themselves out of the routes that matter. The ones who earn adjust fees constantly, watching which channels drain and which fill, nudging prices up on the popular directions and down on the ones they want to attract flow toward. It is closer to running a currency desk than to collecting rent.

The revenue ceiling is set by the network’s own efficiency. Every improvement that lets payments find cheaper routes, from better pathfinding to multi-part payments that split a transfer across several channels, is good for users and quietly deflationary for routers. Lightning is engineered to drive its own routing fees toward zero. That is a feature for anyone paying, and a problem for anyone hoping to be paid.

Where the Money Goes

If revenue is simple, costs are where the business gets punishing, because most of them are paid in the one currency Lightning was built to escape: on-chain Bitcoin fees.

Every channel begins and ends life on the base chain. Opening one is a Bitcoin transaction; closing it is another. Lightning infrastructure firm Voltage uses a clean example: assume roughly 1,500 satoshis to open and 1,500 to close a channel. On a 5,000,000 satoshi channel, that 3,000 satoshi round trip means the channel has to route enough volume to clear about 600 ppm just to break even on its own existence, before it earns a cent of profit. When the mempool is busy and on-chain fees spike, that break-even line jumps, and channels that looked profitable in a calm week turn into losses in a congested one.

Then there is rebalancing. A channel that only ever sends eventually runs dry on one side and can no longer forward. To fix it, an operator pays other nodes to route a circular payment that refills the depleted side, which means paying routing fees to earn routing fees. There is inbound liquidity to buy, capital to lock up for months at a time, server hosting, monitoring tools, and the operator’s own hours. None of it shows up in the headline yield, and all of it eats the margin.

Does It Actually Pay? The Break-Even Math

Put revenue and costs together and the answer, for most operators, is no.

Consider a mid-size node with 10 BTC spread across its channels. By one 2026 analysis from Lightning tooling project Clams, such a node might forward around 2 BTC of payments on a busy day yet earn only about 30,000 satoshis, and once server hosting and on-chain channel management are deducted the operation often lands near break-even. Data from the liquidity marketplace Amboss backs up the ceiling: yields from routing and channel leasing have run between about 1% and 4% a year over four years, according to a breakdown of the Magma market, and the higher end usually reflects small channels whose fat percentage is mostly eaten by the on-chain cost of opening them.

Voltage is blunt about the threshold. Most nodes, it notes, are not run in a profitable way, and the ones that are usually hold well over 10 Bitcoin in capacity. That is upwards of a million dollars of Bitcoin committed to earning a low-single-digit percentage, in a business that demands constant attention. Line that return up against on-chain alternatives and it looks thin. A holder can lend the same Bitcoin, or park equivalent capital in the tokenized credit markets we mapped in the on-chain yield curve, often for comparable or better yield without babysitting channels or paying miner fees every time the network reshuffles.

Which is why the honest framing of a Lightning node has shifted. Voltage’s own advice is to treat a node as a utility first and an income source second: you will get out of it as much as you put in, in both capital and time. For a business that needs to move Bitcoin cheaply, running a node is infrastructure that happens to defray some of its own cost. As a standalone yield play, it rarely clears the bar. That single reframing explains most of the hobbyist exodus in one line.

Liquidity Became a Commodity You Buy

The deepest structural change of the past few years is that Lightning liquidity stopped being something you grew and became something you buy.

The root problem is inbound capacity. A brand-new channel you open is all outbound: you can send, but nobody can send to you, so any invoice you generate is effectively unpayable until liquidity flows the other way. Bitcoin developer Matt Corallo has pointed out that this is exactly why one-click node setups so often leave newcomers with invoices that cannot be paid, a papercut that has frustrated Lightning users since the early days. Solving it by hand means convincing well-connected nodes to open channels back to you, or looping funds out and back in, both of which cost time and on-chain fees.

So a market grew up to sell inbound liquidity directly. The largest today is Amboss Magma, a peer-to-peer marketplace where operators list channel offers, specifying capacity, price, and duration, and buyers pay to have incoming liquidity opened to them on demand. Pricing is transparent and, tellingly, sized by risk and by miner fees rather than by any grand yield. Large channels of a Bitcoin or more lease at roughly 2.6% a year, while small channels under a million satoshis can list above 4%, though much of that richer headline goes straight to covering the on-chain cost of the opening transaction, the Magma breakdown notes.

The significance is cultural as much as economic. Liquidity is now priced, leased, and traded like any other commodity, with spreads and marketplaces and professional suppliers. That is a mark of maturity, the same way deep repo markets are a mark of maturity in traditional finance. It also quietly raises the barrier to entry. A hobbyist who once bootstrapped a node with patience now faces a menu of paid liquidity products, and the operators best equipped to buy at scale are, unsurprisingly, the ones already large enough to route most of the volume.

Who Runs Lightning Now: Four Kinds of Operator

If the median hobbyist node is fading, what is taking its place? The routing layer has sorted itself into roughly four archetypes, and the trend line runs from the bottom of this list to the top.

The gravity is unmistakable. Capital and volume are concentrating into fewer, larger, better-run nodes rather than spreading across thousands of hobbyist machines, and many small operators have simply closed their channels or moved their coins into wallets that handle the plumbing for them. Splicing, a protocol upgrade that lets a channel be resized without closing and reopening it on-chain, has accelerated the shift by letting big nodes manage liquidity more efficiently with fewer, larger channels. Network topology has consolidated toward well-connected hubs, and by some measures the network carries meaningfully fewer channels per node than it did at the start of the decade while moving far more money. Efficient, yes. Also more concentrated.

The Service Providers Doing the Work Now

Sitting above the individual operators is a layer that did not meaningfully exist a few years ago: Lightning Service Providers, or LSPs, and managed infrastructure firms that run nodes so their customers do not have to.

The names recur across every large integration. Lightspark, founded by former PayPal and Meta executive David Marcus, has become a default rail for institutions, with Coinbase and Nubank among those routing Lightning traffic through it; Coinbase has said a meaningful slice, on the order of 15%, of its Bitcoin withdrawals now leave over Lightning, per Spark’s tally. Voltage runs managed nodes and monitoring for businesses that want their own infrastructure without the operational burden. IBEX offers a hand-it-to-us processor model where the provider holds the Bitcoin and takes a cut. Tether-backed Speed settles BTC, USDT, and USDC for merchants internationally.

For anyone who has followed the account-abstraction wave on other chains, the pattern will look familiar. Just as smart-account wallets let a paymaster sponsor gas and manage keys so a user never touches the machinery, LSPs open channels, supply inbound liquidity, and manage routing so a wallet can offer a clean pay-and-receive experience out of the box. It is genuinely better UX. It is also a layer of intermediaries sitting between users and the base chain, earning the fees that hobbyist routers used to chase, and concentrating routing decisions in a handful of well-capitalized firms.

That is the tradeoff nobody has fully resolved. The LSP layer is why Lightning finally feels smooth enough for a mainstream app, and it is why the network’s routing map looks less like a mesh of equals every year.

Stablecoins Changed the Customer, Not the Node Math

The single biggest change to what travels over Lightning is that, since 2025, it no longer has to be Bitcoin.

Using Lightning Labs’ Taproot Assets protocol, the network can now carry tokenized assets, and the headline arrival is Tether’s USDT. The initiative was announced from El Salvador in January 2025, with Tether chief executive Paolo Ardoino framing it as a way of ‘creating practical solutions for remittances, payments, and other financial applications that demand both speed and reliability,’ and Lightning Labs chief executive Elizabeth Stark calling it ‘a new era for stablecoins’ that pairs Bitcoin’s security with Lightning’s speed. By 2026 dollars were moving over the same channels as Bitcoin, at fees measured in fractions of a cent. The technical groundwork here, the same Taproot upgrade that Bitcoin activated years earlier, is the through-line we traced in the Taproot scorecard.

For a routing operator, this matters less than the headlines suggest, and more than it first appears. Less, because the fee mechanics are unchanged: a forward is a forward, whether it moves satoshis or synthetic dollars, and the ppm math does not care which. More, because stablecoins bring a different, stickier kind of demand. A remittance app or a merchant that wants to settle in dollars is a heavier, more predictable source of volume than a Bitcoiner tipping on Nostr, and multi-asset channels raise the complexity of liquidity management, since an operator now has to think about balances in more than one asset.

The net effect is to reinforce the professionalization already under way. Serving stablecoin flow well rewards exactly the operators who have capital, tooling, and active management, and it does little for the hobbyist whose node was a weekend project. Dollars on Lightning expand the market. They do not democratize who gets paid to route it.

The Security Tax, and the Developer Who Walked

There is a cost to running a routing node that never shows up on a profit-and-loss sheet: the standing risk that someone will try to attack it, and the vigilance required to keep that risk low.

The most public reminder came in October 2023, when Bitcoin developer Antoine Riard publicly stepped back from Lightning development after disclosing a class of vulnerabilities he called replacement cycling attacks, which under the wrong conditions could let an attacker steal funds from a channel as it closes. Riard did not soften it: he wrote that the attacks put Lightning in ‘a very perilous position, where only a sustainable fix can happen at the base-layer,’ meaning a change to Bitcoin itself. Lightning experts have since deployed mitigations that harden the network against the simpler versions, and no such theft has been reported in the wild, but the underlying tension, that some Lightning risks cannot be fully closed at the second layer, has not gone away.

Riard’s departure was the loudest, not the only one. Several prominent protocol contributors have quietly reduced their involvement, and the network still carries an unsolved channel-jamming problem, where an attacker floods a node with payments that are deliberately never completed, tying up its liquidity at little cost to themselves. None of these are everyday events. All of them add to the mental overhead of running a node, and overhead is precisely what pushes marginal operators out. Security research and disclosure are the network’s immune system, and every quiet departure thins the ranks of the people watching for the next problem.

The Regulatory Squeeze on Self-Custody

If economics explain why hobbyists are leaving, regulation explains why some of them were pushed. The pressure that reshaped who can run Lightning in the United States did not come from the Securities and Exchange Commission, and that itself is worth understanding.

Bitcoin, the asset that moves over Lightning, is treated in the United States as a commodity, which keeps it outside the SEC’s securities remit. The dollar tokens now riding the network fall under a different regime again: the GENIUS Act, the federal stablecoin law signed in July 2025, which sets bank-style rules for payment stablecoin issuers and, by its own terms, excludes those tokens from being classified as securities or commodities. In other words, neither the asset nor the stablecoins on Lightning are primarily the SEC’s problem.

The live risk for operators has been money transmission. Across 2024, that pressure pushed two of the best-known Lightning wallets out of the US market. The custodial Wallet of Satoshi had already pulled its app from US stores, and after federal prosecutors indicted the founders of Samourai Wallet on charges of running an unlicensed money transmitting business, with prosecutors alleging the service had moved more than $2 billion, ACINQ removed its self-custodial Phoenix wallet from US app stores, warning that recent announcements from US authorities had cast doubt on whether self-custodial wallet providers, Lightning service providers, or even Lightning nodes could be treated as money services businesses.

That last clause is the one that chilled hobbyist operators. If simply running a routing node could in theory land you inside a money-transmission framework, the calculus for a US resident running one for fun changes sharply. The questions are the same ones the global standard-setters have been circling for years, which we unpacked in our guide to FATF crypto guidance: when does moving someone else’s value make you a regulated intermediary, and does non-custodial software get an exemption? Until those answers are settled, regulatory ambiguity acts as one more quiet tax on running Lightning yourself.

Centralization, or Just Efficiency?

All of this lands on a single contested question: is a Lightning Network with fewer, bigger, professionally run nodes a failure of Bitcoin’s decentralization promise, or simply what a maturing payment network looks like?

Both readings have real support. The optimistic case is that consolidation is efficiency. Fewer channels per node moving more volume is a sign the network learned to use its capital better, splicing cut the on-chain churn that used to bloat the channel count, and a payment layer does not need tens of thousands of hobbyist routers any more than the internet needs every home to run a backbone router. Nodes validating Bitcoin’s consensus rules, the thing that actually keeps the money trustless, are a separate population entirely, and that number remains robust. Routing is a service layer on top, and service layers professionalize.

The pessimistic case is that routing centralization has its own costs, even if consensus stays decentralized. A network where a handful of LSPs carry the bulk of exchange flow is a network with a handful of choke points, places where surveillance, censorship, or a well-aimed regulator could bite. The migration of ordinary users into custodial wallets compounds it: a custodial Lightning user has traded away the self-sovereignty that was the entire point, holding an IOU rather than their own keys, and every metric that celebrates smooth adoption is partly measuring that trade.

The honest verdict is that both are true at once, and that the public dashboards are too crude to settle it. Node counts miss the private channels that make up most of the network. Capacity charts miss who controls the capacity. What is clear is that Lightning’s routing layer, the part that moves the money, is concentrating, and that the run-your-own-node ideal is quietly becoming the province of businesses and enthusiasts with capital to spare rather than the default any user is expected to reach for.

What to Watch Next

Several threads will decide whether the operator base stabilizes or keeps thinning.

• Whether node counts find a floor. A network can shed hobbyists and stay healthy, but a routing layer that keeps concentrating eventually trades resilience for efficiency in ways that are hard to reverse.
• The BOLT 12 rollout. The next-generation offers standard promises reusable payment codes and better privacy, but adoption has lagged behind the dominant LND implementation, and until the largest node software ships it fully, wallets cannot rely on it.
• Channelless challengers. Newer designs such as Ark and Spark aim to deliver Lightning-like payments without the channel management that makes routing a chore, and they are courting exactly the developers and users that Lightning needs to keep.
• Stablecoin volume. If dollars over Lightning grow into serious remittance and merchant flow, they could deepen the network’s economics enough to widen the circle of profitable operators, or simply entrench the incumbents best placed to serve them.
• Regulatory clarity. Whether US authorities carve a clear safe harbor for non-custodial software and node operators will determine how many hobbyists feel safe coming back.

Each of these pulls in a different direction, and none resolves the core paradox on its own. Lightning can keep getting better as plumbing while getting narrower as a network of operators, and for now that is exactly what it is doing.

The Bottom Line

The blunt answer to the question in the title is that a Lightning node rarely pays on its own in 2026, and that this is less a verdict on the network than a description of what it has become. Volume and capacity sit near records, dollars ride the rails alongside Bitcoin, and the payment experience is smoother than it has ever been. The plumbing works. Running the plumbing, though, is a professional’s game now, a thin-margin utility business that rewards capital, tooling, and constant attention, and punishes the casual operator who once defined the network.

For most people the rational move is to use a good wallet and let an LSP handle the machinery, accepting the tradeoff between convenience and self-sovereignty with eyes open. For those with real Bitcoin to commit and hours to spend, a routing node can be a break-even utility that quietly earns a low-single-digit return, useful mainly to businesses that need to move Bitcoin anyway. And for Bitcoin’s decentralization ideal, the picture is unresolved: the consensus layer stays broad, the routing layer keeps narrowing, and the promise that anyone can run the network is becoming, in practice, a promise that anyone may. Whether that distinction matters is the argument Lightning will be having for years.

Frequently Asked Questions

Is running a Lightning Network node profitable in 2026?

For most operators, no. A mid-size node with well over 10 BTC in channels can earn a low-single-digit annual return at best, and often lands near break-even once on-chain fees, rebalancing, hosting, and the opportunity cost of locked capital are counted. Profitable routing is possible, but it requires significant capital and active daily management, which is why most nodes are best treated as useful infrastructure rather than an income source.

Why is the Lightning Network node count declining?

Public node counts have fallen from roughly 20,700 in 2022 to around 17,400 by mid-2026, mainly because routing profitably now demands real capital and effort that hobbyist operators cannot justify. Many have closed their channels or moved to wallets that manage liquidity for them. Protocol upgrades like splicing also let larger nodes do more with fewer channels, so capacity has concentrated in fewer, better-run machines even as the raw count drops.

How much Bitcoin do you need to run a profitable Lightning routing node?

Voltage estimates that nodes run profitably usually hold well over 10 BTC in channel capacity, which at recent prices is more than a million dollars committed to earning a low-single-digit percentage. Below that level, on-chain fees and the cost of buying inbound liquidity tend to consume any routing revenue, which is why small hobbyist nodes rarely clear break-even.

What is a Lightning Service Provider (LSP)?

An LSP is a company that runs Lightning infrastructure on behalf of wallets and apps, opening channels, supplying inbound liquidity, and managing routing so end users get a smooth pay-and-receive experience without operating a node themselves. Providers such as Lightspark, Voltage, IBEX, and Speed power much of the network’s institutional and merchant traffic, which improves usability but concentrates routing in a handful of firms.

Can you still self-custody Bitcoin on Lightning in the US?

Yes, but the options narrowed after 2024, when regulatory pressure tied to money-transmission law led wallets including the self-custodial Phoenix and the custodial Wallet of Satoshi to withdraw from the US market. Self-custodial Lightning software remains available through other apps and by running your own node, though ongoing uncertainty about whether non-custodial operators count as money services businesses continues to shape which products serve US users.

By the HOGE Wire desk, covering Bitcoin’s layers and the economics that run them. This report is journalism, not investment advice.