Lightning Fee Revenue: Can Routing Nodes Build a Sustainable Business?

Lightning Network fee revenue remains one of the most debated topics in Bitcoin infrastructure. The network processes over $1 billion in monthly transaction volume as of early 2026, yet the operators who make this possible: routing nodes forwarding payments across the network: earn fees measured in fractions of a cent per transaction. This creates a fundamental tension. The network depends on liquidity providers locking up capital in payment channels, but the economic returns for doing so barely cover operating costs for most participants.

Whether Lightning's fee economics can sustain a healthy network of independent routing nodes, or whether the model inevitably concentrates around a few well-capitalized hubs, has implications far beyond Lightning itself. It shapes how we think about the long-term viability of payment-channel networks and whether alternative Bitcoin Layer 2 architectures might offer more sustainable economic foundations.

How Lightning Routing Fees Work

Every Lightning payment that traverses more than one hop incurs routing fees set independently by each forwarding node. The fee has two components: a flat base fee charged per forwarded HTLC regardless of amount, and a proportional fee rate that scales with the payment size.

Fee calculation

The formula is straightforward: fee = base_fee_msat + (amount_msat × fee_rate_ppm / 1,000,000). The base fee is denominated in millisatoshis (thousandths of a satoshi), while the proportional rate is expressed in parts per million (ppm). A node charging 1,000 msat base fee and 200 ppm would collect 1 sat plus 0.02% of the routed amount.

The network median base fee sits at approximately 1 satoshi (999 msat), while median proportional fee rates cluster around a few ppm. In practice, a typical payment routed across two or three hops costs the sender roughly 0.01% to 0.05% of the payment amount: often less than a single satoshi for micropayments.

Zero base fee movement: A growing number of routing nodes have adopted zero base fees to improve compatibility with multipath payments, where large transactions split across multiple routes. Fixed per-HTLC charges penalize splitting, so removing them improves payment success rates at the cost of per-transaction revenue for node operators.

What Routing Nodes Actually Earn

Public data on routing revenue is sparse because much of Lightning's traffic flows through private (unannounced) channels invisible to network observers. The data that does exist paints a mixed picture.

Block Inc. (Cash App)

Block reported at Bitcoin 2025 that its Lightning nodes earn approximately 9.7% annual returns on 184 BTC of committed public capacity. This headline figure demands context: Block's nodes charge outgoing fees of 1,053 ppm and inbound fees of 2,955 ppm, with some channels set to the maximum 32-bit integer value (2,147,483,647 ppm). These rates are orders of magnitude above network averages. As researcher Riccardo Masutti noted, the double-digit yields are closely tied to an "exceptionally aggressive fee structure" that works only because Cash App controls captive traffic from its millions of users.

BitMEX Research experiment

In a widely cited 2019 experiment, BitMEX Research operated a routing node for roughly three months. Their highest annualized return reached 2.75%, with daily earnings fluctuating wildly from zero to over 3,000 satoshis. They found the revenue-maximizing fee rate sat at approximately 1 ppm (0.1 basis points): above this threshold, daily fee income declined as payments routed around the more expensive node.

Their conclusion was blunt: the system architecture appears "rigged towards users and low fees, rather than liquidity providers." While the network has grown substantially since 2019, the competitive dynamics that suppress fee rates have only intensified.

Amboss MAGMA marketplace data

The Amboss MAGMA marketplace, which facilitates explicit pricing for inbound liquidity, reports historical yields of 1% to 4% APY across four years of operation. The March 2025 LINER (Lightning Network Rate) showed liquidity service providers earning approximately 3.86% APY, while nodes purchasing liquidity paid roughly 5.44% APR. These figures represent the best-case scenario: nodes actively participating in a structured marketplace, not passive operators hoping traffic finds them.

The Cost Structure of Running a Routing Node

Fee revenue tells only half the story. To evaluate whether routing is a sustainable business, you need to account for the full cost stack that operators bear.

The break-even math is unforgiving. A node opening a 5,000,000 sat channel that costs 1,500 sats to open and an estimated 1,500 sats to eventually close needs to charge at least 600 ppm just to recover on-chain costs: before accounting for server expenses, rebalancing, or the time value of locked capital. Yet competing nodes offering 50 to 200 ppm will route most of the traffic.

Rebalancing: the hidden margin killer

Circular rebalancing is often the largest ongoing expense for routing nodes. When channels drain in one direction (common with popular merchants receiving many payments), the node must pay fees to push liquidity back to the depleted side. This creates a paradox: the more traffic a node routes, the more it spends on rebalancing, potentially consuming all the fees earned.

Negative inbound fees, now supported in Core Lightning and LND 0.18+, offer a partial solution by letting nodes offer discounts for routing through depleted channels. But this mechanism is still in early adoption, and its impact on overall routing economics remains to be seen.

Lightning Fees vs. Card Network Interchange

Lightning's fee levels become remarkable when compared to traditional payment networks. The gap between what card networks charge merchants and what Lightning routing nodes earn highlights both the network's value proposition for payers and its sustainability challenge for operators.

Lightning fees are roughly 30x to 350x cheaper than interchange fees. This is a massive win for users and merchants, but it means the entire Lightning routing infrastructure generates a fraction of the revenue that traditional payment networks extract. Visa processed $14.8 trillion in 2024 and earned tens of billions in fees. Lightning, processing an estimated $12 billion to $14 billion annually, generates perhaps a few million dollars in total routing fees across all operators.

The sustainability paradox: Lightning's low fees are its strongest competitive advantage against traditional payment rails, but they make it nearly impossible for independent routing operators to build sustainable businesses purely from routing revenue. The network's greatest feature for users is its greatest challenge for the infrastructure providers who keep it running.

The Free Rider Problem and Network Centralization

Lightning's fee economics produce a predictable outcome: capacity concentrates around a small number of well-capitalized nodes. The top 10 Lightning nodes control approximately 85% of public capacity. Research measuring the Gini coefficient for capacity distribution found it reached approximately 0.97 in 2025: near-perfect inequality, comparable to wealth distribution in the most unequal economies on Earth.

This concentration is not accidental. It is the rational result of Lightning's economic structure.

Why small operators leave

• Routing revenue scales with channel capacity and network position. Nodes with less than 1 BTC in channels rarely route enough traffic to cover on-chain costs.
• Rebalancing costs hit small operators disproportionately hard, as they have fewer channels to absorb directional flow imbalances.
• Inactive counterparties lock up capital indefinitely without generating returns. A channel with a peer who stops using Lightning becomes dead weight until force-closed.
• The opportunity cost of capital rises as alternative Bitcoin yield products emerge. Why lock 0.5 BTC in a Lightning channel earning 1% when staking protocols offer higher returns with less operational overhead?

As small operators exit, the network becomes more dependent on large hubs: exchanges like Bitfinex (~510 BTC in capacity), infrastructure providers like ACINQ (~446 BTC), and fintech companies like Block. These entities can subsidize routing losses because Lightning payments serve their core businesses (exchange deposits, wallet ecosystems, payment apps), not because routing fees generate standalone profit.

Channel Jamming: An Asymmetric Threat to Fee Revenue

Channel jamming attacks compound the fee sustainability problem by allowing adversaries to disable routing capacity at negligible cost. An attacker sends HTLCs that are never resolved, tying up a node's channel slots and capacity without completing payments.

The economics are starkly asymmetric. Research from the Lightning jamming research group estimates that jamming a single routing node requires opening just one channel (approximately 156 sats in on-chain costs) with ongoing costs of roughly 200 sats per month. Reducing network-wide payment success by 20% (jamming approximately 14,000 channels) would cost between $28 and $282 per month at current prices.

For a routing node earning a few thousand satoshis per day, even a brief jamming attack wipes out days or weeks of revenue while the locked capital sits idle. Proposed mitigations include token-based systems and up-front fees, but neither has been widely deployed. For a deeper analysis of jamming defenses, see our coverage of channel jamming mitigation strategies.

Can Volume Growth Alone Make Routing Profitable?

One common argument is that Lightning simply needs more volume: as adoption grows, the same channels will route more payments, increasing revenue without proportional cost increases. This thesis has some merit but faces structural limits.

What scales

Server costs are essentially fixed. A node that routes 10 payments per day and one that routes 10,000 pay roughly the same hosting fees. The marginal cost of forwarding an additional HTLC is near zero in computational terms. If payment volume grows 100x while fees remain constant, routing revenue grows 100x against a largely fixed cost base.

What does not scale

Capital requirements scale directly with volume. Outbound liquidity constrains maximum payment sizes, and total throughput depends on how quickly channels can be replenished. A node routing $10,000 per day needs far less capital than one routing $1 million per day, and the rebalancing costs grow proportionally.

More fundamentally, volume growth attracts competition. New nodes enter when they observe profitable routes, adding capacity and driving down fees along those paths. This is healthy for users but prevents existing operators from capturing the value of network growth. The equilibrium naturally trends toward a fee floor set by marginal cost rather than a margin that rewards capital providers for their risk and opportunity cost.

Average payment size matters

River Financial reported that average Lightning transaction sizes grew from $11.84 in mid-2023 to over $200 by late 2025, driven by institutional adoption and exchange settlement. Because proportional fees scale with amount, larger payments generate more revenue per forwarded HTLC. This trend helps routing economics, but it also reflects a shift away from Lightning's original micropayment use case toward a model where profitability depends on high-value transfers.

Emerging Solutions for Routing Economics

Recognizing these challenges, the Lightning ecosystem has developed several approaches to improve the economic sustainability of liquidity provision.

Liquidity marketplaces

Platforms like Amboss MAGMA and the protocol-level liquidity ads specification allow nodes to explicitly price and sell channel capacity. Instead of hoping that traffic will flow through their channels, operators can charge upfront for the service of providing inbound liquidity. Amboss launched its Rails product in May 2025, enabling Bitcoin holders to earn yield as self-custodial liquidity providers.

Lightning Service Providers

The LSP model bundles liquidity provision with wallet services. ACINQ (Phoenix), Breez, and others open channels on behalf of end users, recovering costs through service fees rather than routing fees alone. This shifts the economic model from "charge per routed payment" to "charge for access to the network," similar to how ISPs monetize internet infrastructure. As explored in our analysis of Lightning Service Providers, LSPs are becoming the primary economic engine for Lightning infrastructure.

Dual-funded channels and fee innovation

Dual-funded channels let both parties contribute capital when opening a channel, sharing the upfront on-chain cost. Negative inbound fees allow nodes to offer discounts on depleted channels, reducing the need for expensive circular rebalancing. These protocol-level improvements chip away at specific cost drivers but do not fundamentally change the fee revenue equation.

Alternative Economic Models: Beyond Routing Fees

Lightning's fee sustainability challenges highlight a broader question: is the payment-channel model, where infrastructure revenue depends on per-transaction routing fees, the right economic foundation for a Bitcoin Layer 2?

Several alternative approaches have emerged. Federated models like Fedimint pool custody and liquidity, spreading costs across a community. Rollup-based systems amortize settlement costs across batches of transactions rather than individual channel operations. And statechain-based architectures like Spark eliminate the channel model entirely: transfers happen by rotating key shares rather than routing payments through a network of funded channels.

The economic difference is structural. In Lightning, every payment path requires pre-funded channels, and every routing node must maintain balanced liquidity across all its channels. In Spark, there are no channels to fund, no liquidity to balance, and no routing fees to collect or pay. The cost model shifts from ongoing capital allocation by a network of independent operators to a simpler operational model where operators facilitate transfers without locking capital in bilateral channels.

This does not make Lightning's model wrong: its fully trustless architecture remains a powerful property that statechain-based systems trade away in favor of 1-of-n operator trust assumptions. But it does mean that fee sustainability is not a universal Layer 2 challenge: it is specific to the payment-channel-network model and the capital-intensive infrastructure it requires.

What Comes Next for Lightning Fee Economics

The most likely evolution is continued bifurcation. Large institutional nodes (exchanges, payment companies, LSPs) will subsidize routing because it serves their business model, while independent operators will increasingly rely on liquidity marketplaces and service fees rather than pure routing revenue.

Industry analysts project 40% to 60% fee decreases over the next two years as enterprise adoption scales and competition intensifies. This is excellent news for users: Lightning payments will become cheaper and more reliable. It is a more complicated story for infrastructure providers who need to find revenue models that go beyond per-transaction fees.

The parallel is informative: email servers are largely unprofitable to run independently, yet email works because large companies operate servers to support their broader businesses. Lightning may follow a similar path, with routing as a utility subsidized by those who benefit most from its existence, rather than a standalone profit center.

For developers building on Bitcoin's Layer 2 ecosystem, the choice of infrastructure model matters. Spark's developer documentation covers how its channelless architecture avoids the liquidity management overhead that drives routing costs, while our State of Lightning in 2026 analysis provides additional context on the network's trajectory.

This article is for educational purposes only. It does not constitute financial or investment advice. Bitcoin and Layer 2 protocols involve technical and financial risk. Always do your own research and understand the tradeoffs before using any protocol.