L2
Layer 2. A blockchain that runs on top of an L1 to provide cheaper or faster transactions while inheriting the L1's security. L2s batch many transactions and post compressed proofs back to the L1.
Also known as: Layer 2, rollup
L2s exist because most L1s have a fundamental scalability ceiling. Ethereum can process about 15-30 transactions per second on its base layer. NEAR is faster but still bound by validator coordination overhead. When demand exceeds the base layer’s throughput, fees spike and the chain becomes expensive to use. L2s solve this by running their own execution environment with much higher throughput, batching transactions, and posting only periodic state updates back to the L1 for final settlement.
The two main categories are optimistic rollups (Arbitrum, Optimism, Base) and ZK rollups (zkSync, Scroll, StarkNet, Polygon zkEVM). Optimistic rollups assume transactions are valid by default and use a fraud-proof challenge period to catch invalid ones. ZK rollups generate cryptographic proofs that every transaction is valid before posting to the L1. ZK rollups are faster to settle (no challenge period) but slower to generate proofs. Both achieve the same goal of scaling throughput without sacrificing L1 security.
Base (Coinbase’s L2) deserves a special mention in DeAI context because it’s where Venice’s VVV token lives, where NEAR Intents settles many cross-chain swaps, and where most modern x402-compatible payment rails are deployed. Base is an Optimism-stack rollup with high throughput, low fees, and Coinbase distribution. The OYM Venice review and the agent commerce stack article cover its role in the broader DeAI infrastructure.
The honest framing is that L2s are a pragmatic compromise. They give users cheap transactions but require trusting a more centralised execution layer than the L1 underneath. Most L2s have a single sequencer (the entity that orders and batches transactions) which is a centralisation point. Most L2s also have privileged “admin” upgrade paths that could in theory be abused. The trust assumptions are weaker than running directly on the L1 but stronger than using a centralised exchange. The OYM Freedom Score’s Infrastructure dimension treats L2-based projects as inheriting both the strengths and weaknesses of the underlying L2 architecture.