Rollups Explained: How Layer 2 Scales Ethereum
Optimistic vs zero-knowledge rollups, why they slash fees, and what "inheriting L1 security" actually means.
Why Ethereum needs help
Ethereum's base layer processes only a handful of transactions per second, and when demand spikes, fees follow. Rollups fix this by executing transactions off-chain in bulk, then posting a compressed summary back to Ethereum. You get cheaper, faster transactions while still settling on the most secure smart-contract chain.
The core trick: batch and prove
A rollup takes hundreds of transactions, executes them on its own layer, and publishes the resulting data plus a proof to Ethereum. Because the data lives on L1, anyone can reconstruct the rollup's state and challenge fraud — that is what "inheriting L1 security" means.
L2 batch -> compress -> post to Ethereum L1
txs_in_batch: 843
l1_calldata: ~92 KB
cost_per_tx: ~$0.01 (vs ~$1.80 on L1)Optimistic rollups
Optimistic rollups assume every batch is valid and only run a computation if someone disputes it. That dispute window — usually about seven days — is why withdrawals back to L1 can be slow unless you use a liquidity bridge. Arbitrum and Optimism take this approach.
Zero-knowledge rollups
ZK rollups post a cryptographic validity proof with every batch, mathematically guaranteeing the transactions were executed correctly. No fraud-proof window means faster finality, at the cost of heavier proving work. zkSync, Starknet, and Scroll live here.
Which one wins?
Both. Optimistic rollups are mature and EVM-compatible today; ZK rollups offer stronger guarantees and are catching up fast on developer experience. For users the practical difference is withdrawal speed and which apps have deployed — the security model, in both cases, ultimately leans on Ethereum.
Written by the newsroom
Independent crypto and web3 coverage, verified against on-chain data. Replace this bio with your own — a line about who you are and the beats you cover goes a long way in earning a reader's trust.