Welcome to the next crypto level
← Nexorena

Blockchain

Ethereum fee simulation compares one price with prices by resource

In an analysis published on Ethereum Research on November 12, 2026, Fei Wu measures how separate fees affect execution capacity, state data growth and block utilization. Conservative configurations based on EIP-7999 process 4% to 34% more execution work in the simulation at comparable congestion.

Nexorena Editorial Team

Ethereum fee simulation compares one price with prices by resource

Three resources share a block’s budget

An Ethereum transaction uses several resources. The network has to execute instructions, transmit the data needed to propagate the block and retain some information after the transaction, such as an account balance or values stored by a contract. These operations put different kinds of pressure on nodes. A block that mostly adds to the data they retain puts a different load on them than one that spends most of its budget on computation.

On November 12, Fei Wu published a summary on Ethereum Research that builds on a preprint dated October 24. Wu compares mechanisms that retain a shared base fee with a mechanism based on EIP-7999, where each resource studied has its own target and base fee. In both cases, demand varies from block to block. The analysis tracks the execution work performed, growth in state data and the share of blocks using at least 98% of a hard limit.

Limits constrain the gains from a single price

In the single-price benchmark, which adapts parameters including those of EIP-8372, target utilization is half the limit. The simulation yields 151.9 million to 177.9 million units of execution gas per block, depending on the assumed propagation time. State data grows at an annualized rate of 86 to 88.1 gibibytes, against a target of 120. Raising target utilization to four-fifths increases the amount of work processed, but state growth falls and the share of blocks close to a limit rises to 13.2%–20.6%.

The researcher attributes part of this result to the fee adjustment rule in EIP-1559. At very low prices, rounding can wipe out the calculated reduction for a block below its target. A rule based on EIP-4844, which accumulates deviations over time, allows fees to fall further. In the simulations with a four-fifths target, it makes room for more computation and new state data. In exchange, 38.8% to 40.8% of blocks come close to a limit.

Results depend on the demand in the model

Conservative EIP-7999 configurations process 4% to 34% more execution work than the single-price scenario with a target set at half the limit. Their share of near-limit blocks ranges from 1.9% to 4.4%, compared with 4% to 4.5% in the benchmark. Across the separate-price configurations studied, state data growth stays close to the target of 120 gibibytes per year. Each fee can respond to demand for its resource when the others are below target.

The demand model draws on the gas limit increases of February and July 2025. Some execution levels tested go well beyond observed demand. Results also vary with the assumed responsiveness to fees, and the tested single-price scenarios cover only some of the possible configurations. The results concern a proposed design change. The size of its effect on mainnet remains to be measured under the demand and parameters that would apply there.

Article sources

Written with love byOlivia ParkerMeet this editorial pen name ↗

Explorer Nexorena