Bitcoin Layer 2 Comparison

Bitcoin's base layer processes roughly 7 transactions per second. Layer 2 protocols solve this by moving most activity off-chain while still anchoring to Bitcoin's security. The Lightning Network is the dominant L2 and the foundation that most other protocols connect through. This article covers Lightning and the newer alternatives — Liquid, Spark, Ark, and Ecash — how each works, what tradeoffs it makes, and how they all interoperate.

Lightning Network

What It Is

The Lightning Network is Bitcoin's primary Layer 2 payment protocol. It enables instant, high-volume transactions between nodes using a network of bidirectional payment channels. Lightning has been in production since 2018 and is by far the most widely adopted Bitcoin L2, with thousands of nodes, tens of thousands of channels, and growing integration across wallets, exchanges, and merchants.

How It Works

Payment channels: Two parties open a channel by locking bitcoin into a 2-of-2 multisig on-chain. Once open, they can send payments back and forth instantly by updating the channel state — each update is a new commitment transaction that both parties sign but don't broadcast. Only the final state hits the blockchain when the channel closes.

Multi-hop routing: You don't need a direct channel with every person you want to pay. Lightning uses onion-routed payments — your payment hops through intermediary nodes, each of which forwards the payment using HTLCs (Hash Time-Locked Contracts). Each hop only knows the previous and next hop, not the full payment path.

HTLCs and atomicity: A payment either completes fully or fails entirely. The sender creates a payment hash; the recipient reveals the preimage to claim funds. Each intermediary is protected by timelocks — if the next hop doesn't settle in time, the HTLC expires and funds return to the sender. This makes multi-hop payments trustless and atomic.

Channel capacity and liquidity: Each channel has a fixed capacity set when it's opened. The balance can shift between the two parties as payments flow, but the total capacity doesn't change. This creates a liquidity management challenge — you need sufficient inbound liquidity to receive payments and outbound liquidity to send them.

Why Lightning Is the Foundation

Lightning is not just one of several L2 options — it is the interoperability layer that connects them all. Nearly every other Bitcoin L2 (Liquid, Spark, Ark, Cashu, Fedimint) uses Lightning as its bridge to the broader Bitcoin economy. When a Cashu user pays a merchant, or an Ark user sends to a Lightning wallet, the payment routes through Lightning. This makes Lightning nodes critical infrastructure for the entire L2 ecosystem.

Key characteristics:

Fully non-custodial — you hold your own keys and can close channels unilaterally
Trustless — enforced by Bitcoin Script, not reputation or federations
Instant — payments settle in seconds
Low fees — typically under 1 sat for small payments
Largest network effect — most wallets, exchanges, and merchants support Lightning
Mature — battle-tested since 2018 with multiple independent implementations (LND, Core Lightning, Eclair, LDK)

Tradeoffs:

Requires managing channel liquidity (inbound and outbound capacity)
Must be online to receive payments
Channel opens/closes require on-chain transactions (and fees)
Running a routing node requires technical knowledge and ongoing maintenance

Taproot Assets: Stablecoins Arrive on Lightning

Lightning has historically been Bitcoin-only, unlike Liquid, which has long supported stablecoins (see below). That changed in March 2026, when Tether confirmed that USDT is live on Bitcoin and Lightning through Lightning Labs' Taproot Assets protocol, following a roughly fourteen-month integration effort. Taproot Assets issues other assets on top of Bitcoin using Taproot outputs and routes their transfers through existing Lightning channels — so a Lightning node that upgrades to support Taproot Assets can carry stablecoin payments over the same infrastructure it already uses for sats, without a separate sidechain or federation. For node runners, this means channel liquidity and routing decisions may increasingly need to account for asset-aware payments, not just BTC.

Liquid Network

What It Is

Liquid is a federated sidechain operated by the Liquid Network — a consortium of exchanges, brokers, and infrastructure companies (currently 65+ members, called functionaries). It has been in production since 2018 and is the oldest of the four systems covered here.

Liquid uses a two-way peg: you lock BTC on the Bitcoin mainchain and receive L-BTC on the Liquid sidechain at a 1:1 ratio. L-BTC can be transacted on Liquid and redeemed back for BTC.

How It Works

The Federation: Liquid relies on an 11-of-15 multisig held by functionaries running HSM (Hardware Security Module) hardware. Peg-ins and peg-outs require this quorum. The HSMs are designed to enforce federation rules and cannot sign arbitrary transactions, limiting insider risk — but the trust model still requires you to trust the federation not to collude or be coerced.

Confidential Transactions: All Liquid transactions use Confidential Transactions (CT), a cryptographic technique that hides transaction amounts and asset types from public observers while still allowing validators to confirm no new coins are created. This is the strongest on-chain privacy feature of any production Bitcoin sidechain.

Block time and finality: Liquid produces blocks every minute (vs. Bitcoin's ~10 minutes) and achieves strong finality after 2 blocks (~2 minutes). There is no probabilistic finality — the federation signs blocks, so a block signed by a quorum is final.

Assets: Liquid supports native assets beyond L-BTC, including stablecoins (USDt on Liquid), security tokens, and other issued assets. This is a feature for exchanges that need to move multiple assets on the same rails.

Peg mechanics:

Peg-in: Send BTC to a federation-controlled address. After 102 Bitcoin confirmations (~17 hours), L-BTC is credited on Liquid.
Peg-out: Burn L-BTC, request peg-out. Federation signs the Bitcoin withdrawal transaction within a few hours.

The long peg-in confirmation requirement is a deliberate security measure to prevent reorg attacks on the peg.

Spark

What It Is

Spark is a Layer 2 protocol developed by Lightspark that uses Statechains combined with a Spark Service Provider (SSP) to enable off-chain bitcoin transfers. It launched mainnet in early 2025. Unlike Liquid, Spark does not use a sidechain — all funds ultimately settle on Bitcoin L1.

How It Works

Spark builds on the Statechain concept: instead of transferring ownership of a UTXO by spending it on-chain, you transfer the signing key for that UTXO off-chain, with the cooperation of an operator.

Key components:

Leaves: UTXOs locked in a 2-of-2 multisig between the user and the SSP operator. The operator holds one key; the user holds the other.
Key transfer: When Alice sends to Bob, Alice gives Bob her signing key share and the SSP rotates its own key to invalidate Alice's old share. Bob now co-controls the UTXO with the SSP using a fresh key combination.
Pre-signed exit transactions: At each state, a time-locked exit transaction is pre-signed so the user can unilaterally withdraw to L1 without SSP cooperation.
Lightning interoperability: Spark includes a Lightning gateway that allows converting Spark leaves to Lightning payments and vice versa, enabling interoperability with the broader Lightning Network.

Trust model: You must trust the SSP not to collude with a prior owner to double-spend. The SSP signs an attestation that it has deleted old keys, but this is not cryptographically enforced — it is enforced by the SSP's reputation and legal agreements. If the SSP disappears, you can always exit to L1 via the pre-signed transactions.

No scripting limitations: Because Spark leaves are standard Bitcoin UTXOs, they can carry any Bitcoin Script, including Taproot outputs.

Unilateral Exit in Practice

The pre-signed exit transactions are the whole basis for calling Spark self-custodial, so it's worth knowing what invoking them actually looks like. A July 2026 case study, published in Blink's spark-unilateral-exit GitHub repository, walked a real mainnet wallet holding 100,000 sats across 22 leaves through a full unilateral exit, with every transaction publicly verifiable on-chain.

The headline numbers: exiting required 253 separate on-chain transaction packages — one confirmation per level of each leaf's exit chain, since exit transactions use v3 (TRUC) transaction relay rules that cap a chain to one unconfirmed parent and child at a time, so each level must confirm before the next can broadcast. Exit transactions are pre-signed at zero fee and rely on CPFP fee-bumping from a separate funding UTXO; bumping every leaf in this wallet would have cost roughly 78,500 sats — about 79% of the wallet's balance — because most of the 22 leaves were small change from ordinary wallet activity ("dust") not worth the fee to exit individually.

Triaging by economics rather than exiting everything changed the outcome substantially: of the 22 leaves, only 4 (holding 90% of the wallet's value) were worth exiting at all. Doing so cost about 8,400 sats in fees, and after the mandatory ~2-week CSV timelock on the refund transactions, roughly 89,700 of the original 100,000 sats — about 90% — reached the destination address.

The people who ran this exit summarized it well: unilateral exit is "a fire escape, not a door." It works, and the funds are genuinely recoverable without operator cooperation, but it is deliberately slow — weeks, not minutes — and can be expensive if you wait until you need it rather than keeping leaves consolidated and exit tooling ready beforehand.

Ark

What It Is

Ark is a protocol proposed by Burak Keceli in 2023 and now being developed by Ark Labs. It aims to give users a Lightning-like experience — instant sends, low fees — without requiring users to manage inbound liquidity or be online to receive payments.

Ark is still in active development, and the ecosystem has grown quickly through 2026: Bark, an alternative client implementation, is now live on Bitcoin mainnet alongside Arkade and hArk (a hash-lock-based variant), and the V-PACK standard for stateless vTXO verification has been published. Second (second.tech) is a production wallet built on Ark. The protocol continues to evolve rapidly, and breaking changes remain possible.

How It Works

Ark introduces a new primitive: the vTXO (virtual UTXO). vTXOs exist off-chain and are periodically anchored to the Bitcoin blockchain through a process called a round.

The Ark Service Provider (ASP): An ASP is a always-on server that coordinates rounds. The ASP is not custodial in the traditional sense — users hold vTXOs with pre-signed exit paths — but the ASP must cooperate for normal operations.

Rounds: Every ~5 seconds (configurable), the ASP publishes a round transaction on Bitcoin. This transaction:

Spends a pool UTXO from the previous round
Creates a new pool UTXO containing all the fresh vTXOs from this round's participants
Includes connector outputs that allow participants to exit

Each vTXO in the round is a leaf in a sparse Merkle tree committed to in the round transaction. Users receive a Merkle proof for their vTXO.

Receiving: To receive an Ark payment, you do not need to be online. You simply give the sender your Ark address. Your vTXO will appear in the next round.

Sending: Sending within Ark means swapping your old vTXO for a new vTXO in the next round. The ASP atomically exchanges vTXOs using a VTXO swap — the sender and receiver's vTXOs are both included in the same round, with an atomic link ensuring the sender's old vTXO is only consumed if the receiver's new vTXO is created.

Exit: If the ASP stops operating, users can exit their vTXOs to L1 by broadcasting the pre-signed exit transactions. There is a timelock (e.g., 4 weeks) on vTXOs — after expiry they must be refreshed by participating in a new round, or they can be unilaterally claimed on-chain.

Lightning interoperability: Ark can interoperate with Lightning through the ASP acting as a Lightning gateway, similar to how Spark works.

Ecash (Cashu and Fedimint)

Ecash is not a single protocol but a family of approaches that implement blind signature schemes to provide strong privacy. The two main implementations are Cashu and Fedimint.

Cashu

What it is: Cashu is a simple, custodial ecash protocol with a single mint operator. It is explicitly custodial — you trust the mint completely. The tradeoff is near-perfect privacy.

How it works: Cashu uses Chaumian blind signatures (invented by David Chaum in the 1980s). The flow:

Alice deposits sats at the Cashu mint via Lightning invoice.
Alice generates a random secret, blinds it, and sends the blinded value to the mint.
The mint signs the blinded value without seeing the underlying secret, returns the blind signature.
Alice unblinds the signature to produce an ecash token — a valid (secret, signature) pair.
Alice sends the token to Bob (over any channel — even a plain text file).
Bob redeems the token at the mint. The mint verifies the signature and marks the secret as spent, preventing double-spending.

The mint never sees the link between step 2 (receiving the blinded token) and step 6 (seeing the redeemed token). This is cryptographic unlinkability — even a fully logging mint cannot connect deposits to withdrawals.

Denominations: Cashu uses a UTXO-like denomination model (1, 2, 4, 8, ... sat tokens), allowing offline transactions and splitting/combining of tokens.

Trust: Cashu is fully custodial. The mint operator can rug users. This is appropriate for small amounts and communities that trust their mint operator.

Fedimint

What it is: Fedimint is a federated ecash protocol that distributes mint trust across a group of guardians using threshold signatures. It is designed for communities (families, villages, companies) that want the privacy benefits of ecash with reduced single-point-of-failure risk.

How it works: Fedimint uses the same Chaumian blind signature scheme as Cashu, but replaces the single mint key with a threshold multisig (e.g., 2-of-3 or 3-of-5 guardians). Creating ecash tokens requires a quorum of guardians to cooperate.

The federation holds the Bitcoin reserves in a multisig on-chain. Users interact with a Fedimint client that communicates with all guardians.

Modules: Fedimint is modular. The Lightning module allows the federation to operate a Lightning gateway, so users can send and receive Lightning payments using their federated ecash balance.

Trust: You trust the guardian quorum not to collude. Below the threshold, guardians cannot steal or deanonymize users. A 2-of-3 federation is harder to compromise than a single Cashu mint operator, but still requires trusting a known group of people.

At a Glance

Lightning — Non-custodial, trustless (Bitcoin Script). Seconds finality. Must be online to receive. Requires liquidity management. Production since 2018. Best for: payments, routing, infrastructure.

Liquid — Non-custodial (federated peg), 11-of-15 federation trust. ~2 min finality. Confidential Transactions for privacy. ~17 hour peg-in. Can receive offline. Production since 2018. Best for: exchanges, institutions.

Spark — Non-custodial (self-exit), SSP trust (reputational). Instant finality. Lightning gateway built in. Must be online to receive. Production since 2025. Best for: Lightning power users.

Ark — Non-custodial (self-exit), ASP trust (reputational). ~5 second rounds. Can receive offline. No liquidity management. Production (2025–2026), multiple implementations (Bark, Arkade). Best for: users who don't want to manage channels.

Cashu — Fully custodial, single mint trust. Instant finality. Cryptographic unlinkability (strongest privacy). Bearer tokens work offline. Production since 2023. Best for: communities, small amounts, privacy.

Fedimint — Federated custodial, guardian quorum trust. Instant finality. Same privacy as Cashu but distributed across a federation. Production since 2024. Best for: communities, resilience.

Tradeoffs Summary

Lightning Network

Pros:

Fully trustless — enforced by Bitcoin Script, not federations or operators
Largest network effect and ecosystem of any Bitcoin L2
Instant payments with sub-satoshi fees
Multiple independent implementations (LND, Core Lightning, Eclair, LDK)
Foundation layer that all other L2s connect through

Cons:

Requires active liquidity management (inbound/outbound)
Must be online to receive payments
Channel opens and closes cost on-chain fees
Running a routing node requires ongoing maintenance and technical knowledge

Liquid

Pros:

Longest production track record
Confidential Transactions provide real on-chain privacy
Supports multiple assets on the same rails
Fast finality for institutional use cases

Cons:

Long peg-in time (~17 hours) creates friction
Federation is a trust and regulatory attack surface
Requires running a Liquid node or trusting an explorer
Small ecosystem compared to Lightning

Spark

Pros:

No inbound liquidity management
Native Lightning interoperability
Strong self-exit guarantees via pre-signed transactions
Uses standard Bitcoin UTXOs (no new scripting)

Cons:

Trust in SSP reputationally (not cryptographically enforced)
Key transfer model is novel — long-term security assumptions less tested
Centralized SSP creates a potential single point of failure
Protocol still maturing
Unilateral exit is real but slow and can be costly if leaves aren't consolidated (see "Unilateral Exit in Practice," above)

Ark

Pros:

Receive without being online (unlike Lightning)
No liquidity management required
Strong cryptographic self-exit path
Efficient: many users share one on-chain round transaction

Cons:

vTXOs expire and must be refreshed (adds complexity)
ASP must be online and cooperative for normal operation
Protocol still evolving; breaking changes possible
Round-based model introduces latency (5-second rounds)

Cashu

Pros:

Strongest privacy of any option (cryptographic unlinkability)
Extremely simple to deploy and use
Works over any transport (even QR codes, NFC, text)
Instant transactions

Cons:

Fully custodial — mint operator can steal or disappear
No self-exit path; funds are lost if the mint goes offline
Not suitable for large balances
Trust is entirely social, not cryptographic

Fedimint

Pros:

Strong privacy (same as Cashu)
Distributed trust across a guardian quorum
Modular design supports Lightning, on-chain, and future modules
Designed for community trust models

Cons:

Still requires trusting a known group of guardians
More complex to set up than Cashu
Federation coordination adds operational overhead
No self-sovereign exit — if all guardians disappear, funds are at risk

Interoperability: How These Protocols Connect

The most important thing to understand about these L2s is that they don't exist in isolation. Lightning is the common language that connects them all.

Lightning as the Interoperability Layer

When a Cashu user pays a Lightning invoice, the mint makes the Lightning payment on their behalf. When a Fedimint user receives a payment, their federation's Lightning gateway handles the routing. Spark and Ark both include native Lightning gateways that convert between their internal state formats and Lightning payments.

This means Lightning node runners are not just serving Lightning users — they are the backbone infrastructure for the entire Bitcoin L2 ecosystem. Every gateway, mint, and ASP needs well-connected Lightning nodes to function.

Boltz: Trustless Swaps Between Layers

Boltz is a non-custodial swap service that bridges between Bitcoin on-chain, Lightning, and Liquid. It uses submarine swaps and reverse submarine swaps — atomic swap constructions that use HTLCs to ensure neither party can steal funds.

How submarine swaps work:

Lightning → On-chain (Reverse swap): You want to move funds from Lightning to an on-chain address. Boltz creates an on-chain HTLC. You pay a Lightning invoice. When Boltz receives the Lightning payment and obtains the preimage, they can claim their on-chain HTLC — but only if they've already locked funds on-chain for you. The preimage simultaneously proves payment and unlocks your on-chain funds.

On-chain → Lightning (Submarine swap): You want to move on-chain BTC to Lightning. You lock BTC in an on-chain HTLC. Boltz pays the Lightning invoice you specify. When the Lightning payment succeeds, Boltz obtains the preimage and can claim the on-chain HTLC.

Lightning ↔ Liquid: The same swap mechanics work between Lightning and Liquid's L-BTC, allowing trustless movement between the two networks.

Why Boltz matters for node runners:

Move liquidity between on-chain and Lightning without closing channels
Rebalance channel capacity by swapping in or out
Bridge to Liquid for confidential transactions or to access Liquid-based exchanges
All swaps are non-custodial — Boltz never holds your funds, the HTLC construction ensures atomicity

Boltz operates as open-source software at boltz.exchange, and the swap protocol can be self-hosted.

Which One Should You Use?

There is no universal answer — the right tool depends on your threat model and use case:

You want the most trustless, battle-tested payment network: Lightning. It's the foundation everything else builds on, and the only L2 enforced entirely by Bitcoin Script without trusting any third party.
You need to move value between exchanges with strong privacy: Liquid is the most mature option for institutional use.
You want Lightning-like speed without managing channels: Spark or Ark both remove the inbound liquidity problem; Spark is more production-ready today, Ark is more decentralized in design.
You want maximum privacy for small amounts in a trusted community: Cashu is the simplest and most private option.
You want Cashu-style privacy with reduced custodial risk: Fedimint distributes trust across a federation of guardians.
You need to move funds between layers without trusting anyone: Boltz swaps provide trustless bridges between Lightning, on-chain, and Liquid.

For Lightning node runners, these protocols represent both infrastructure you depend on and opportunity. Running gateways for Ark or Fedimint, providing Boltz swap liquidity, or simply maintaining well-connected routing capacity all contribute to the broader Bitcoin L2 ecosystem — and can generate routing revenue.