Layer 2 & Scaling

Bitcoin's base layer puts one goal first: "anyone in the world can run a full node." The block limit is defined in weight units, 4,000,000 WU, which in practice yields blocks of roughly 1-2MB with a theoretical maximum near 4MB. That limit and the 10-minute block interval follow from the same design philosophy.

The constraint caps theoretical throughput at about 7-14 transactions per second, far below what a global payment network needs.

A Layer 2 (L2) is a protocol built on top of Bitcoin's base layer. Most transactions happen on the L2, and only the final settlement is recorded on the base layer, which is where the extra capacity comes from.

L2s inherit Bitcoin's security (Proof of Work, decentralization) while each makes its own trade-offs in speed, privacy, and smart-contract capability.

Unlike Ethereum's L2s (Arbitrum, Optimism, and others), Bitcoin's L2s take widely different technical approaches, and no single standard has emerged.

One caveat belongs up front: "Bitcoin Layer 2" is not itself a technically agreed category. There is no common standard for how much of the base layer's security a system must inherit before it earns the name, and in practice the label covers chains with their own consensus and even arrangements where funds are simply handed to an operator. The term is a marketing convenience as much as anything. This article therefore judges projects not by whether they are "L2s" but by the categories set out next and by each one's trust assumptions.

What gets lumped together as "Layer 2" in fact divides into distinct categories. The name tells you less than two questions do: where do the funds sit, and whose approval do you need to get them back.

Payment channels: two parties lock funds and update balances off-chain. The defining property is that you can exit on-chain unilaterally, without the counterparty's consent (Lightning).

Sidechains: a separate chain with its own consensus. Movement to and from the base layer runs through a "peg," and whoever secures that peg, whether a federation or merge-mining, becomes the party you have to trust (Liquid, Rootstock).

Federations / threshold signatures: several parties hold shares of a key and move funds on the agreement of some number of them. Collusion above the threshold can take the funds (Liquid's functionaries, the sBTC signer set, Fedimint's guardians).

Rollups / fraud proofs: execution happens off-chain, and on-chain verification runs only when a result is disputed. On Bitcoin, the BitVM family aims in this direction, but practical deployment is still under way (Bitlayer, Citrea, BOB).

Client-side validation: an asset's validity is established not by the chain but by validation data passed between the parties, with only a minimal commitment placed on the blockchain (RGB, Taproot Assets). Nodes do not recognize the asset, so losing the validation data means losing the asset.

Bridges: mechanisms for moving assets between chains. The largest losses of funds in crypto have happened at bridges, again and again. The danger in "moving funds onto an L2" is, in most cases, concentrated in the bridge design. The record of those incidents is covered in "Major Incidents & Lessons."

The design was published in 2015, and major implementations reached mainnet in 2018. It is the most widely adopted Bitcoin L2, built on payment channels and specialized for micro-payments and instant settlement.

Public capacity, channel count, and node count all move week to week. This site keeps those figures in one place, in "Lightning Network Primer"; see that article for current statistics. Duplicating them here would guarantee that the two pages contradict each other after any update.

March 2026: Tether launched USDT on the Lightning Network through Taproot Assets, making dollar-denominated stablecoin payments on a Bitcoin L2 real.

LSPs (Lightning Service Providers) emerged so that mobile payments work without the user managing channels. Where an LSP or a custodial wallet takes over, though, what you trust is that business rather than the protocol.

Working deployments include Strike in the U.S. and tipping across Nostr. El Salvador's Chivo wallet also used Lightning, but following the January 2025 legal amendment and the country's IMF agreement, the government has stated it will end public involvement by selling or shutting Chivo down. It no longer counts as a current deployment.

See the dedicated "Lightning Network Primer" topic for a deeper treatment.

A federated sidechain launched in 2018 by Blockstream, a major Bitcoin Core development company.

It uses L-BTC (Liquid Bitcoin), a version of bitcoin pegged 1:1, with a two-way peg to mainnet Bitcoin.

Features: fast 2-minute blocks, Confidential Transactions for privacy, and Issued Assets for multiple currencies and tokens.

Governance has two tiers. The Liquid Federation had roughly 87 participating organizations as of Q1 2026, but block signing and custody of the peg are performed by the 15 nodes known as functionaries. The two-way peg with mainnet Bitcoin is secured by an 11-of-15 multisig across those 15 keys.

L-BTC's backing therefore rests on two premises: that 11 functionaries do not collude, and that fewer than five go offline at once. This 11-of-15 is the core of Liquid's trust assumption, a design that knowingly accepts less decentralization than Bitcoin itself in exchange for speed and confidential transactions.

Tether issues USDT-Liquid, used for OTC settlement between major exchanges.

Rootstock (RSK): an EVM-compatible sidechain merge-mined with Bitcoin, launched in 2018. It runs Ethereum-compatible smart contracts on Bitcoin.

Stacks: uses its own Proof of Transfer (PoX) consensus. Stacks miners send BTC to stackers (STX stakers) in exchange for STX block rewards.

Stacks 3.0 (Nakamoto Release, 2024) tightened its integration with Bitcoin finality. Stacks transactions are "anchored" into Bitcoin blocks and inherit its security.

sBTC is the implementation for handling BTC on Stacks. Deposits, and only deposits, went live on mainnet in December 2024, with withdrawals following on April 30, 2025. The peg is secured by an initial signer set of 15 institutions, with funds moved by threshold signature.

It is sometimes described as "non-custodial," but users cannot recover funds on their own: the design assumes the signer set keeps operating correctly, which places it closer to a federation in trust terms. Trustlessness and a more decentralized signer set are stated as staged goals.

Though called Bitcoin L2s, these have their own consensus and their own tokens; "affiliated chain" describes them more accurately.

October 2023: Robin Linus published a novel idea for verifying arbitrary computation on Bitcoin.

Rather than adding new Bitcoin Script operations, it uses "fraud proofs" for optimistic execution. On-chain verification runs only when results are disputed.

The mechanism resembles Ethereum's Optimistic Rollups and in theory allows Turing-complete computation on a Bitcoin L2.

Since 2023, BitVM1, BitVM2, and further improvements have followed, and several projects (Bitlayer, Citrea, BOB) are building BTC L2s on BitVM.

Practical deployment has its problems, including the many on-chain transactions a worst case may require, but the idea opened new ground in Bitcoin L2 design.

Founded in 2022, with mainnet launched in 2024. It pledges BTC as security for other PoS chains while the coins stay on the Bitcoin mainchain.

To be precise, what is time-locked is not a private key but the staking output, a UTXO. The user sends their funds to a script carrying several spending conditions: (1) the owner can withdraw after a set period, (2) the covenant committee can agree to an early unbonding, and (3) a slashing transaction can confiscate the funds if misbehavior occurs.

What makes that slashing possible is EOTS (Extractable One-Time Signature). A finality provider signs each block height exactly once, using randomness committed in advance. Signing two different blocks at the same height reuses that randomness, and from the two signatures anyone can reconstruct the provider's private key. The recovered key is then used to sign the pre-arranged slashing transactions, confiscating the BTC delegated to that provider.

So "self-custody staking" is accurate in the sense that the funds never move to a third party's address. But if the finality provider you delegated to double-signs, your stake is cut even though you did nothing wrong, and the correctness of the scripts and the behavior of the covenant committee are part of the premise too.

BTC staking spread rapidly between 2024 and 2026, and "Restaking" projects built on Babylon (Lorenzo, Pendle, and others) have appeared.

It provides a new yield source for BTC holders, while drawing criticism from those who hold that having no yield is precisely the point of Bitcoin.

Ark: a protocol aiming at instant payments through a different mechanism than Lightning. Instead of long-lived channels, it uses a shared-UTXO model of joint holdings.

Fedimint: a federated Bitcoin custody system based on Chaumian ecash. Privacy is high, and it works as a community- or family-scale Bitcoin bank. The funds are actually held by operators called guardians, however, and collusion above the threshold takes them. Structurally, it is custody.

Cashu: a similar ecash protocol, lighter and friendlier to developers than Fedimint, but the funds are held by a single mint. If the mint absconds, users have no way to recover them. It suits small amounts held with an operator you trust.

These are tools for handling small amounts, with high privacy, between parties you trust. They sidestep Lightning's liquidity and channel-management complexity by giving up self-custody. The strength and the weakness come from the same design.

Still experimental as of 2026, but gaining ground in privacy-focused communities.

Comparing these by "fast or slow" tells you less than comparing who can cause you to lose funds and whether you can exit on-chain without that party's consent. What follows is a contrast of trust assumptions, not a recommendation.

On speed: Lightning settles in seconds, Liquid on 2-minute blocks, Rootstock and Stacks in tens of seconds to minutes, and BitVM-family systems depend on the implementation. In practice this difference rarely decides anything; the second and third columns above matter far more.

Privacy has no single verdict either. Lightning keeps transactions off the public ledger, but the channel graph is public and information reaches counterparties along the route and any provider you use. Liquid's confidential transactions hide amounts and asset types, but within the federation's field of view. Fedimint and Cashu offer strong privacy against outsiders while the operator holds the funds.

Making up for the falling miner reward after each halving is a long-term problem for the layers that depend on on-chain fees. L2 development is one of the pillars holding up the wider Bitcoin economy.

Ironically, Ordinals and Runes have livened up the Bitcoin fee market and supported miner revenue.

If cross-L2 atomic swaps become practical, interoperability across Lightning ↔ Liquid ↔ Stacks improves.

Practical BitVM-family L2s could bring Ethereum-compatible DeFi to Bitcoin.

Critics argue that a proliferation of L2s sacrifices Bitcoin's "simplicity." Which L2 comes out on top will be settled by competition over the coming years.