Layer 2s explained: why fees are lower on rollups

What a rollup is doing

Ethereum mainnet is expensive because every validator has to process every transaction and keep enough data for the network to verify what happened. A rollup moves most of that work elsewhere. It collects many user transactions, executes them off the main chain and then posts a compressed record of the batch back to Ethereum.

That design keeps Ethereum as the settlement layer. If the data needed to reconstruct the rollup state is available on Ethereum, anyone can in principle verify the rollup's history and challenge a bad state transition. Ethereum's layer 2 guide gives the official overview.

Optimistic and zero-knowledge rollups

An optimistic rollup posts the batch and assumes it is correct unless someone proves otherwise during a challenge window. Arbitrum One and OP Mainnet are the best-known examples. Because the system allows time for a fraud proof, withdrawals back to Ethereum usually wait about a week.

A zero-knowledge rollup, often shortened to zk rollup, posts a cryptographic validity proof showing that the batch followed the rules. That removes the long fraud-proof window for the state transition itself, so final settlement can be quicker. The trade-off is that generating those proofs is technically harder and adds its own operational complexity.

Why fees fell further after blobs

Rollups pay Ethereum not only for settlement, but also for publishing the data users need to verify them. For a long time they had to store that data in ordinary call data, competing with every other mainnet use. The Dencun upgrade in March 2024 introduced a separate, cheaper data lane called blobs in EIP-4844, aimed specifically at rollups.

Blobs are not kept forever by every node in the same way as ordinary chain history, but they remain available long enough for rollups and their watchers to verify the batches. That was enough to cut the data part of many rollup fees sharply, which is why simple transfers and swaps on major rollups often cost cents rather than dollars.

Sequencers, bridges and withdrawals

Most rollups today have a sequencer: the service that orders incoming transactions and turns them into batches. A sequencer makes the user experience fast, but it is also a central point that can delay or censor transactions for a time. Well-designed rollups provide an escape hatch that still lets users force transactions on chain if the sequencer goes down.

Moving funds between Ethereum and a rollup also means using a bridge. Deposits are usually quick because the rollup sees the mainnet transaction and credits you soon afterwards. Withdrawals are where the design difference shows most clearly: optimistic rollups make you wait out the challenge period, while zk rollups can often settle faster because the proof has already shown the batch was valid.

What risks remain

• The rollup contracts on Ethereum can contain bugs or be upgradable by an admin key or multisig. Check those controls before parking large balances there.
• The sequencer can fail, censor or simply become congested, even if users can eventually recover through the main chain.
• Bridged assets may not be native on the rollup. A token that exists only as a wrapped version carries bridge or issuer risk as well as rollup risk.
• Fees can still spike when the rollup itself is busy, or when Ethereum data costs rise and that cost is passed through to users.

When a layer 2 makes sense

For ordinary transfers, swaps and smaller positions, a major rollup often gives the best trade-off between cost and security. It is cheaper than Ethereum mainnet and usually more battle-tested than a small independent chain. Our gas tracker shows the difference in current fees across Ethereum and its layer 2s, and the transaction lookup helps you inspect what happened once you are on one.

What a layer 2 does not do is make risk disappear. It rearranges the costs and the trust assumptions. You still need to know who controls the contracts, how the bridge works and how quickly you can get back out.