There’s a Mexican standoff in Bitcoin’s Lightning Network

Everyone on a liquidity route in Bitcoin’s Lightning Network wants the same rebalance of funds to happen, but none of them wants to be the first person to pay.

The tenuous impasse is a classic Mexican standoff.

When this situation occurs, Lightning node operators can neither pay nor not pay first without harming themselves, so nobody moves, and nobody wins. It has been a recurring problem for years.

After almost a decade of tools and research chasing this problem, the network’s routing nodes remain locked in a standoff that quietly erodes routing reliability of bitcoin (BTC).

The Lightning Network, Bitcoin’s largest layer 2 network with no connection to an altcoin, has a structural bias toward channel depletion.

That is, money tends to flow in one direction along channels from senders toward structural receivers, such as merchants receiving BTC who deliver goods and services to customers.

Routing nodes are left with channels that are stuffed with BTC on one side and depleted of BTC on the other. A channel that cannot send in both directions is effectively half-broken for routing purposes. 

Lightning’s total capacity set a fresh all-time high of roughly 5,600 BTC in December 2025, but that surge arrived almost entirely through institutional deposits from Binance and OKX into existing channels. Year-to-date data tells a different story.

BTC capacity from December’s high above 5,600 has declined to 4,884 today, and payment channels have declined from more than 80,000 in mid-2023 to about 45,000 today, nearly halving as liquidity consolidated into lopsided channels on a shrinking graph.

The cheapest fix is the one nobody will start

René Pickhardt, one of the network’s most prolific routing researchers, wrote that most channels “are expected to be depleted over time, primarily due to selfish routing behavior within the current protocol design.”

By his accounting, any given payment link has roughly a coin-flip chance of avoiding long-term depletion.

Researchers have described the embarrassingly simple yet elusive solution. 

Presenting several nodes sitting on a circular payment path connected to one another by payment channels and all lopsided in the same direction, each could push BTC around the loop and finish a complete cycle with healthier channels as a result.

Everyone would benefit if everyone cooperated at the same time.

The problem, as with every Mexican standoff, is who pays first. 

Routing BTC over Lightning costs money. Whichever node kicks off the rebalance owes routing fees to every other hop on the loop. If they wait, other nodes can receive their rebalancing for free and pocket the fee.

Although every node operator on the ring would benefit most as a collective if they all pushed BTC around for a single sending plus a single receiving fee — i.e. for nearly free in net — each operator also has the rational choice to wait for someone else to send first so they can collect without sending.

Wait for someone else to move first. It is a Mexican standoff.

Read more: Why two-party Bitcoin Lightning channels keep failing

A graveyard of fixes for Lightning channel imbalances

The industry has thrown nearly a decade of engineering at channel imbalances without solving these types of standoffs.

Alex Bosworth’s submarine swaps, announced in August 2018, let operators shuffle BTC between on-chain and Lightning to reload channels.

It helped a bit, but every swap burned a real BTC transaction fee, so adoption didn’t pick up enough to solve many of the network’s recurring Mexican standoffs.

Lightning Labs packaged the idea into Loop and, later, into Lightning Pool, a non-custodial marketplace for inbound channels since at least November 2020.

Lightning Pool activity declined within roughly a year, and the product faded from the ecosystem. 

Core Lightning’s Liquidity Ads, the closest protocol-native alternative, sees fulfillment that one recent analysis described as sporadic at best.

Amboss Technologies launched Magma in April 2022 to let operators buy and sell inbound liquidity peer-to-peer. Other rebalancing scripts like C-Otto’s rebalance-lnd and Bosworth’s Balance of Satoshis let operators pay themselves through loops when fees permit.

A new proposal this week aims to encourage cooperation at a protocol level.

A perennial problem with Lightning

None of those efforts have prevented standoffs from recurring.

Protos has previously documented Pickhardt’s argument that depletion is a structural property of the two-party channel itself, not an operational hiccup. 

His January 2026 paper identified three possible mitigations: symmetric fees per direction, convex or tiered fees, and coordinated replenishment.

The first two require fee structures most routing nodes would reject outright. The third requires somebody to volunteer to coordinate.

Coordination is where Lightning keeps getting stuck. The protocol is supposed to work with nodes routing selfishly without trusting each other.

When channels become lopsided, however, fixing the network’s most persistent liquidity problem requires exactly the type of coordination the protocol was built to avoid.

Got a tip? Send us an email securely via Protos Leaks. For more informed news, follow us on X, Bluesky, and Google News, or subscribe to our YouTube channel.