Bitcoin scaling explained: the Lightning Network and beyond

Bitcoin scaling is the challenge of handling far more payments than the base blockchain was ever designed to carry. The chain adds a new block only about every ten minutes, and each block holds a strictly limited amount of data, so the base layer clears only a handful of transactions per second. That is fine for large, final settlement, but impractical for buying a coffee. The fix most of the network has converged on is the Lightning Network, a second layer that handles everyday payments off-chain and touches the main chain only when it needs to. I spent years writing explainers like this at Binance Academy, and the clearest way in is to follow a single channel between two people from open to close.

What is the Bitcoin scaling problem?

Bitcoin's base layer settles only around seven to ten transactions per second, because blocks are capped in size and arrive roughly every ten minutes. A card network such as Visa can process tens of thousands per second. That throughput gap, together with a firm settlement wait of up to an hour, is the Bitcoin scaling problem in a nutshell.

Two design limits set the ceiling. Blocks arrive on average every ten minutes, and each one carries a bounded amount of data (about 1 MB of base data, or close to 4 MB of block weight since the SegWit upgrade). Multiply that out and you get the well-known estimate of roughly seven to ten transactions per second for the entire world. How Bitcoin works covers the block and confirmation mechanics in detail.

There is a second issue: finality. A payment is not treated as firmly settled the instant it lands in a block; most recipients wait until it is several blocks deep, which pushes real settlement out to something like an hour. People often compare this unfavourably with Visa, but that is not a fair comparison. Visa's instant approval at the till is not final settlement; the networks and banks reconcile and settle the money later, often batched once a day. Bitcoin's confirmation, by contrast, is final the moment it is buried under enough blocks. So the base layer is slower to hand you a receipt, but that receipt is far stronger. Either way, waiting an hour to buy a coffee is a non-starter. Satoshi's whitepaper framed bitcoin as peer-to-peer electronic cash, yet on its own the base layer is poorly suited to cash-like spending, and that tension is what the scaling debate is really about.

Underneath all of this sits the so-called blockchain trilemma: a network can be decentralised, secure and scalable, but pushing hard on all three at once is extremely difficult. Bitcoin deliberately prioritises decentralisation and security, and refuses to buy raw throughput by sacrificing either. That single choice is why scaling has to happen in layers built on top, rather than by simply speeding up the base chain.

On-chain versus off-chain scaling: what is the difference?

On-chain scaling means making the base blockchain itself carry more, usually by enlarging blocks. Off-chain scaling keeps the base layer small and moves everyday activity onto systems built above it, settling back to the chain only occasionally. Bitcoin firmly took the off-chain route, and almost every serious project reflects that decision.

The on-chain approach is tempting because it is simple: bigger blocks mean more transactions per block. The catch is that a bigger chain is heavier to store and validate, so fewer people can afford to run a full node, and the network drifts toward centralisation. That exact trade-off triggered the 2017 block size wars and the splits that followed, which I cover in Bitcoin's fork history.

Off-chain scaling borrows a trick from ordinary life. Think of a bar tab: you do not settle up after every drink, you run a tab and pay the net total once at the end of the night. Layer-2 systems do the same with bitcoin. They record many payments among themselves, keep a running balance, and write only the final result back to the blockchain. The heavy traffic never touches the base layer, which stays lean and secure.

How does the Lightning Network work?

The Lightning Network is a layer-2 network of two-party payment channels, first proposed in 2015 and rolled out over the years that followed. Two people lock bitcoin into a shared channel, then pay each other simply by re-signing how that pot is divided, without broadcasting anything. Only opening and closing a channel ever touch the blockchain.

A channel is a 2-of-2 multisig arrangement: a shared wallet that requires both signatures to spend. Suppose Alice and Bob open one. Each funds it with 1 bitcoin, so the channel holds 2 (this funding transaction is the single on-chain step at the start). At the same time they both sign a settlement transaction that would pay 1 bitcoin back to each of them, but crucially they do not broadcast it. Either party could publish it at any moment to reclaim their share.

Now Alice wants to pay Bob 0.5 bitcoin. Instead of touching the chain, they jointly sign a new settlement transaction, this one paying 1.5 to Bob and 0.5 to Alice, and they invalidate the previous version. Value has moved from Alice to Bob, instantly, with nothing broadcast. They can repeat this as often as they like, re-signing an updated split each time, for as long as the channel stays open. When either of them wants out, they broadcast the latest signed settlement, the multisig pays out according to the current balance, and the channel closes. That closing broadcast is the second and final on-chain step.

The clever part is the word "network". A channel only lets two specific people pay each other, which would be useless on its own. But channels connect. Say Bob also has a channel with Charlie. Alice has no direct channel to Charlie, yet Lightning can route her payment through Bob: Alice pays Bob, Bob pays Charlie, each hop re-signing its own channel. Bob ends up net flat, taking only a tiny routing fee, and the money reaches Charlie. Payments can hop across several intermediaries, with the network finding an efficient path automatically, so nobody needs a direct channel with everybody. The channel and routing mechanics were first set out in the Lightning Network whitepaper by Joseph Poon and Thaddeus Dryja.

Because all of this is just signing and not broadcasting, a Lightning payment clears in a fraction of a second, for fees that are usually a rounding error, and total capacity across the network runs orders of magnitude beyond the base chain. That makes it well suited to small, everyday spending, which is exactly how it is used. El Salvador, which made headlines for adopting bitcoin, leans on Lightning for real merchant payments in shops and restaurants. Upgrades to the base layer help here too: Taproot, activated in November 2021, lets channel opens and closes blend in with ordinary transactions and makes the underlying scripts cheaper and more private.

What are the advantages and limits of Lightning?

Lightning's strengths line up neatly with Bitcoin's ethos: it moves real bitcoin, you keep control of your own keys, there is no trusted middleman, and the network is decentralised with no single point of failure. Its limits are that it is younger and less battle-tested than the base chain, and that channel capacity caps how much you can move.

On the plus side, Lightning is self-custodial by design, so you are not handing your coins to anyone. It is fast, cheap and, because payments are not written to a public ledger, more private than an on-chain transaction. Those same private payments make it genuinely hard to measure how widely Lightning is used, which is worth remembering whenever you see a confident adoption figure.

The limits are real but narrow. The technology is newer, so it has a shorter track record and is less hardened against attack than Bitcoin itself. More practically, a channel can only move as much as it holds, so you cannot easily push a very large payment through a web of small channels. Channels also need their liquidity balanced: send a lot in one direction and you use up your capacity to keep sending that way until funds flow back. In our example, once Bob has forwarded 0.25 to Charlie, he can send less toward Charlie until the channel is rebalanced. This can be automated, and modern wallets increasingly hide it, but it is a genuine constraint on the network's flexibility. For the everyday, low-value payments Lightning is built for, none of this gets in the way.

What other layer-2 and sidechain options exist?

Beyond Lightning, bitcoin can move off-chain through sidechains such as Liquid, ownership-transfer schemes such as statechains, wrapped bitcoin tokens issued on other blockchains, and plain custodial transfers inside an exchange. Each one buys speed or extra features by accepting a different trust trade-off, and usually a larger one than Lightning asks of you.

Liquid is a sidechain run by a federation of businesses (it was built by Blockstream). You lock bitcoin on the main chain and receive an equivalent token, L-BTC, on Liquid, redeemable one for one. Blocks come every minute or two, transaction amounts can be kept confidential, and other assets can be issued on it, which is why exchanges and traders like it for quick settlement. The trade-off is that you are trusting the federation to honour the peg rather than the open mining network, so it is more centralised than the base layer. You can read more at Blockstream's Liquid pages.

Statechains take a different angle: they let you transfer ownership of an entire coin off-chain by handing over control of its key, with a statechain operator co-signing. The design is meant to ensure the operator cannot spend on its own and previous owners cannot claw the coin back. It is handy for passing whole coins around cheaply, but it leans on the operator behaving honestly, in particular deleting old keys, and it remains newer and more experimental than the options above.

Wrapped bitcoin works by having a custodian hold real bitcoin and issue a matching token on another chain, such as WBTC on Ethereum, with similar tokens on Solana, Polygon and others. That token can move at the speed of the host chain and, more importantly, can be used inside that chain's financial applications, for example as loan collateral. The token is effectively a receipt for bitcoin held elsewhere, much as old paper notes were receipts for gold in a vault. The catch is significant: you give your bitcoin to a custodian and take on that chain's risks too. If trust breaks, the token can de-peg from bitcoin's price, as a Solana wrapped bitcoin did around the collapse of the FTX exchange. Unless you specifically need bitcoin inside another chain's tools, wrapping it mainly adds counterparty risk.

Finally, the simplest off-chain method is an internal transfer inside a centralised exchange. If you and the recipient both use the same platform, it just edits two balances in its database: instant and free. It is also custodial, meaning you do not hold the keys, and if the exchange fails you can lose the lot, as many did with FTX. It is a convenient side effect of how exchanges work, not a safe place to keep meaningful sums. If you are weighing this up, our guides on buying and selling bitcoin and choosing a wallet are the place to start.

Why does Bitcoin keep its base-layer blocks small on purpose?

Small blocks keep the full ledger cheap to store and fast to verify, so ordinary people, not only data centres, can run a node and independently enforce the rules. That wide base of independent verifiers is exactly what makes Bitcoin hard to capture or quietly change. Scaling is pushed upward to protect it.

If blocks were much larger, the chain would grow so quickly that keeping a full copy would become the preserve of a few well-funded operators. Verification would centralise, and a network that only a handful of parties can check is a network that is easier to pressure or coerce. The ten-minute block interval serves the same goal: it gives every node on Earth, including those on slow connections, enough time to receive and validate each block before the next one arrives.

None of this means Bitcoin refuses to improve. SegWit in 2017 restructured transactions to fit more into each block, and Taproot in 2021 made complex spends, including the multisig that Lightning depends on, cheaper and more efficient. Both added real capacity without abandoning small blocks. This is the settlement the block size wars ultimately reached: keep the base layer lean and decentralised, and let higher layers such as Lightning carry the volume.

Frequently asked questions

Is the Lightning Network safe to use?

For the small, everyday amounts it is designed for, Lightning is broadly considered safe, and you keep custody of your own keys throughout. It is younger and less battle-tested than Bitcoin's base layer, so most people treat it like a physical wallet: convenient for spending money, not the place to store their life savings.

Do I need to run a node to use Lightning?

No. Plenty of Lightning wallets manage channels, routing and liquidity for you behind the scenes, so sending a payment feels like any other app. Running your own node gives you more control and privacy, but it is optional. Most everyday users rely on wallet software that handles the technical work automatically.

What is the difference between Lightning and wrapped bitcoin?

Lightning keeps you in control of real bitcoin and has no custodian: you settle back to the main chain whenever you choose. Wrapped bitcoin hands your coins to a custodian who issues a token on another chain. Lightning is built for payments; wrapped bitcoin exists mainly to use bitcoin inside other chains' applications.

Will Bitcoin just increase its block size to scale?

Almost certainly not, and that debate has already been fought. Larger blocks would raise throughput but make the chain costlier to store and verify, weakening the decentralisation that secures it. The network chose to keep blocks small and scale through layers such as Lightning instead, a decision at the heart of the 2017 forks.