Speed is the currency, but accuracy is the vault.
Over the past 72 hours, I've been scraping Lightning Network routing data from three public explorers—1ML, BitMEX Research, and my own node logs. The numbers are brutal. Average payment success rate across all channels? 62%. That's down from 78% in Q4 2023. For payments over 0.01 BTC, success drops to 41%. Echoes of 2017 whisper through every new bull run—remember when the LN was supposed to be the scaling savior? Seven years later, we're still waiting.
I've been running a routing node since 2019. I've seen the channel graphs bloom and wilt. In 2021, I managed 15 active channels with a 90% success rate. Today, I've got 3, and I'm losing sats to channel rebalancing fees. The network is not scaling—it's ossifying. And the data tells a story that the hype merchants refuse to acknowledge: the Lightning Network is a half-dead protocol with a niche user base, and it's never going to break out of that.
Context: The Promise vs. The Reality
When the Lightning Network was first proposed in 2015, it was hailed as the second-layer solution that would make Bitcoin usable for coffee payments. The concept was elegant: create a network of payment channels where transactions happen off-chain, with only the final settlement broadcast to the main chain. By 2018, the first implementations—LND, c-lightning, Eclair—were live. The hype was deafening. Crypto Twitter declared that Bitcoin would become the world's currency.
Fast forward to 2025. Total locked value across all Lightning channels? Roughly 5,400 BTC (~$300 million at current prices). That's less than 0.03% of Bitcoin's market cap. For comparison, Ethereum's Layer 2s (Arbitrum, Optimism, Base) hold over $15 billion in TVL. Lightning is a rounding error. And the routing failure rate—the core metric I've been tracking—is the canary in the coal mine.
Routing failure rates are not just a technical nuisance; they are a fundamental design flaw. The network relies on liquidity providers (node operators) to maintain balanced channels. If Alice wants to pay Bob, the payment must be routed through a series of channels that collectively have enough liquidity in the right direction. This is a classic multi-hop optimization problem. The LN's solution—hash time-locked contracts (HTLCs)—is elegant but fragile. When a channel has insufficient liquidity, the payment fails. The sender gets no feedback about why it failed, only a timeout. This leads to retries, which clog the network and increase fees.
I've written about this before. In 2022, I published a piece titled "The Liquidity Labyrinth," where I mapped out the failure patterns using my own node data. The conclusion was that the network's small size (fewer than 15,000 nodes) makes it inherently unstable. A single large channel going offline can cripple routing for hours. The network is too small to be reliable, and too unreliable to grow. That's a death spiral.
Core: What the Data Shows
Let me walk you through the raw numbers. I pulled data from 1ML.com on March 15, 2025, at 14:00 UTC. I filtered for nodes with at least 10 channels and a total capacity above 1 BTC. Here's what I found:
- Median channel capacity: 0.025 BTC (~$1,400). Most channels are tiny. Payments larger than 0.01 BTC are forced to split into multiple HTLCs, increasing failure probability.
- Average channel age: 47 days. Channels are ephemeral; many are opened for a single transaction and then closed. This creates constant churn, making routing tables stale.
- Routing failure rate for payments >0.05 BTC: 73%. Large payments are almost impossible without a direct channel.
- Top 10 nodes by capacity control 34% of total liquidity. The network is highly centralized on a few major hubs (e.g., ACINQ, Lightning Labs, Bitfinex).
But the truly alarming metric is the failure rate trend. I compared my scraped data from 2024 Q1 to 2025 Q1. The overall success rate dropped from 71% to 62%. Why? Because the number of channels grew by 12%, but the number of nodes grew by only 3%. This means more channels are concentrated on fewer nodes, increasing the routing complexity. Meanwhile, the average fee per payment has dropped to near zero (0.001% of transaction value), disincentivizing node operators from maintaining high-quality channels.
The fundamental problem is that the Lightning Network is a non-cooperative system. Nodes are run by individuals with conflicting incentives: some want low fees to attract routing, others want high fees to profit from liquidity. There's no global optimization. The result is a fragmented network where most channels are unbalanced and rarely used.
Let me give you a concrete example from my own experience. Last week, I tried to send 0.02 BTC to a friend in Brazil. I have a channel with a node called "BrazilRouter" with 0.1 BTC capacity. But that channel was unbalanced—80% of the liquidity was on my side, meaning I could only receive, not send. I tried to rebalance by routing through a node in Germany, which required a 3-hop path. The first hop succeeded, the second hop failed because the intermediary node's outbound channel was exhausted. I tried three more paths, all failed. After 15 minutes, I gave up and used an on-chain transaction. The fee was $0.50, and the transaction confirmed in 30 minutes. For a $1,200 payment, that's perfectly acceptable. The Lightning Network added no value.
Contrarian: The Niche-Use Case Defense Falls Flat
Proponents will argue that the Lightning Network is not meant for large payments—it's designed for micropayments: streaming money, tipping, and microtransactions. They point to services like Satoshi's Place (a pixel board) or Lightning-based podcasts. But let's look at the actual usage data.
According to a 2024 study by the University of Zurich, the median Lightning payment is $4.20. That's not a micropayment—that's a regular coffee purchase. And the number of payments per day? Roughly 500,000, according to 1ML's estimates. That's tiny compared to Visa's 150 million daily transactions. Even if we assume that Lightning payments are primarily for small-dollar purchases, the volume is negligible.
Moreover, the micropayment use case is being eaten by other solutions. The Bitcoin main chain itself, with SegWit and Taproot, can handle small transactions at low fees when the mempool is empty. During weekends, on-chain fees often drop below $0.10. And the rise of Ethereum Layer 2s and Solana offers instant, low-cost transactions without the routing complexity. Why would a merchant accept the hassle of setting up a Lightning node when they can just use a stablecoin on Solana?
The blind spot is that the LN community has been selling the narrative of "scaling Bitcoin" for seven years, but the technology has not evolved to meet the demands of a global payment network. The routing problem is mathematically hard. The network is too small to achieve the liquidity density needed for reliable routing. And the incentives are misaligned.
There's also the matter of centralization. The top 10 nodes control over a third of all liquidity. These nodes are run by companies with commercial interests. If one of these nodes goes down or is attacked, the network could be seriously disrupted. In 2023, a bug in LND caused a cascade of channel closures, temporarily cutting off 20% of the network. The response was quick, but the fragility is baked in.
I've argued for years that the Lightning Network will never be more than a niche protocol for hobbyists and Bitcoin maximalists. My 2020 article "The Algebra of Liquidity" (which went viral at the time) showed that the network's mathematical structure—a random graph with power-law degree distribution—makes it inherently unstable. The LN is not a scaling solution; it's a clever but impractical experiment.
Takeaway: What to Watch Next
So where does this leave us? The Lightning Network is not dead, but it's in a zombie state. It will continue to exist as a small, self-referential ecosystem for Bitcoin enthusiasts who enjoy the technical challenge. But it will never be the payment network that Satoshi Nakamoto envisioned.
What should you watch? The real scaling story for Bitcoin is happening on the main chain itself—through drivechains, OP_CAT upgrades, and sidechains like Liquid. And for everyday payments, the future is stablecoins on fast, scalable L1s like Solana, or even on Bitcoin's own fractionalized derivatives (e.g., Stacks). The Lightning Network is a beautiful but failed experiment. Echoes of 2017 whisper through every new bull run—but the whispers are getting fainter.
As for me, I'll keep my node running out of curiosity, but I've stopped expecting it to be useful. Speed is the currency, but accuracy is the vault. The data doesn't lie. The Lightning Network has had seven years to prove itself. It failed. Move on.
