What Is Layer 2 in Crypto?

A layer 2 (L2) is a blockchain that settles on a base-layer blockchain (L1) rather than maintaining its own independent security set. On Ethereum, L2s batch hundreds of transactions into a single L1 call, paying a fraction of the gas per user. Arbitrum, Optimism, Base, zkSync, Starknet, and Polygon zkEVM are the major Ethereum L2s — collectively handling more daily transactions than Ethereum itself.

Also known as: L2, layer-2 scaling, layer 2 blockchain

Ask Stingray anything about Layer 2

Why L2s exist

Ethereum L1 throughput is capped at ~15 TPS. With DeFi, NFTs, and gaming competing for space, gas fees during peak activity hit $50-$200 per swap in 2021. L2s solve this by processing transactions off the L1 but posting proofs or data back to L1 for settlement — gaining 10-100x throughput while inheriting L1’s security.

The basic pattern:

  1. User sends a transaction to the L2 (from their wallet, bridged USDC, or an exchange that supports deposits).
  2. The L2 processes the transaction, updates L2 state, issues a near-instant confirmation.
  3. Periodically (seconds to minutes), the L2 batches transactions and posts a summary to L1.
  4. The L1 verifies the batch (via a fraud proof, a validity proof, or a simple re-execution).
  5. Once finalized on L1, L2 state is settled.

Major Ethereum L2 families

  • Optimistic rollups (Arbitrum, Optimism, Base) — assume batches are valid; allow a 7-day window for fraud proofs. Cheap, simple to build, but withdrawals back to L1 take the full challenge window.
  • ZK rollups (zkSync, Starknet, Polygon zkEVM, Linea) — post a cryptographic validity proof with each batch. Withdrawals are near-instant once the proof is verified, but proof generation is expensive.
  • Validiums and volitions — variant designs where data availability is handled off-chain (cheaper but different trust assumptions).

Collectively, L2s hold $30-40B TVL across 2024-2025 and handle the majority of retail-scale DeFi and NFT activity.

Risks and considerations

L2s inherit L1 security for transaction finality but introduce their own specific risks:

  • Sequencer centralization — most L2s run a single sequencer (the entity that orders transactions). If the sequencer censors or goes down, users can’t transact until L1 fallback kicks in.
  • Upgradability — L2 contracts are typically upgradable by multisigs. A compromised multisig can drain the bridge.
  • Bridge risk — the canonical L1↔L2 bridge carries significant TVL. Exploits have hit multiple third-party bridges, though canonical rollup bridges have been safer.
  • Data availability — if the L2 can’t publish its batch data, users can’t verify state. Dedicated DA layers (Celestia, EigenDA) are gaining adoption for this reason.

L2Beat publishes a canonical tracker of each L2’s security properties, upgrade authority, and current stage (Stage 0 = trusted, Stage 2 = trust-minimized). Check the stage and upgradability before deploying large capital to a specific L2.

For users, the practical choice among L2s usually comes down to: which has the apps you want, which has the lowest fees today, and which has native USDC (Circle has issued native on most major L2s). Arbitrum and Base dominate Ethereum-alignement DeFi activity in 2024-2025; zkSync and Starknet lead the ZK camp.

Related terms

Rollup A rollup is a specific type of layer-2 blockchain that executes transactions off-chain but posts transaction data to an L1 for data availability and security. ZK Rollup A ZK rollup (zero-knowledge rollup) is a layer-2 chain that posts a cryptographic proof of correct execution to L1 with each batch of transactions. Optimistic Rollup An optimistic rollup posts transaction batches to L1 assuming they're valid, with a challenge window (typically 7 days) during which anyone can submit a frau… Bridge A bridge is a protocol that moves assets or messages between two blockchains.