
When a wallet shows a fee of a fraction of a cent on an L2, it is quoting the sum of two separate charges that are priced by two separate markets. Treating that sum as a single “L2 gas price” is the most common reason readers are surprised when rollup costs jump on a day when the rollup itself looks quiet.
Meter one: execution on the rollup
The first component pays for running your transaction inside the rollup’s own virtual machine. It is priced in gas, like Ethereum, but the gas price is set by the rollup’s own fee rules and by whoever orders its transactions. On most rollups today that is a single sequencer run by the operating team. The sequencer’s price is not a pure cost pass-through: the operator can charge a margin above what it spends, and that margin is revenue. This is a design fact, not an accusation; it is worth knowing because it means part of what you pay is a business decision rather than a protocol outcome.
ZK rollups add one more internal cost on this side of the ledger: generating validity proofs. Ethereum’s own documentation on ZK rollups notes that computing and verifying those proofs is “substantial” and can raise fees, and estimates on-chain verification at roughly 500,000 gas per proof, shared across the batch.
Meter two: the data bill paid to Ethereum
The second component is the rollup’s cost of publishing your transaction data on Ethereum so anyone can reconstruct the L2 state. Since the Dencun upgrade (March 2024) that data mostly travels in blobs, the transaction type defined by EIP-4844. Blobs have their own base fee, separate from ordinary gas, that rises when blocks carry more than the target number of blobs and falls when they carry fewer. Blob fees are burned, and nodes are only obliged to serve blob data for about 18 days.
The capacity of that blob lane has been raised twice since launch. EIP-4844 started at a target of 3 and a maximum of 6 blobs per block. EIP-7691, shipped in Pectra on 7 May 2025, moved that to 6 and 9. After Fusaka activated on 3 December 2025, two parameter-only forks lifted it again: to a target of 10 and maximum of 15 on 9 December 2025, and to 14 and 21 on 7 January 2026, according to the mainnet announcement. More room means the blob base fee sits at its floor more often; it does not mean it can never rise.
Why the two meters diverge
The execution meter reacts to demand on one rollup. The data meter reacts to demand from every rollup posting to Ethereum at the same time. So a quiet rollup can still see higher data costs when other L2s are busy, and a busy rollup can stay cheap if blob space across the network is plentiful. That is also why rollup fees and L1 fees can move in opposite directions during the same hour: since 2024 they no longer compete for the same block space.
What cheap L2s mean for the L1 burn
Under EIP-1559 the base fee on Ethereum is burned. When activity moves to rollups, less execution gas is spent on L1 and less ETH is burned; blob fees are burned too, but at low prices they are small. The net effect on ETH supply depends on the balance between issuance and burn at any given time, and we do not model or forecast it here. Our explainer on issuance minus burn walks through how to read that balance from live data.
The Settlement Test
- What settles: the rollup’s batch data (as blobs or calldata) and, separately, its state commitments on Ethereum. Your transaction is final on L1 only once the batch carrying it is in a finalized Ethereum block and, for withdrawals, once the rollup’s proof or challenge rules are satisfied.
- What it costs: the sum of the L2 execution fee and your share of the L1 data fee. We do not bake figures into this article; the gas and fees page shows current L1 and L2 readings with their fetch time, and the L2 fee comparator does the arithmetic.
- What breaks: if the sequencer stops, the execution meter stops with it until the forced-inclusion path is used. If blob demand spikes, the data meter rises for every rollup at once.
How to read a rollup fee quote
Ask which part moved. If only the execution part rose, the cause is local to that rollup. If the data part rose, check blob demand on Ethereum. If both stay flat but the total is still high, look at the size of your transaction: data cost scales with the bytes you publish, so a complex call costs more on both meters. For the mechanics behind blobs, see our guide to blob fees and data availability and the glossary entry on the blob.
Primary sources
Educational content, human-edited and possibly AI-assisted. It describes how systems work and is not financial, legal or tax advice. Report an error via our corrections page.


