Ethereum Proposal Could Increase the Cost of Onboarding New Users
Key Takeaways
- •EIP-8037 would meter newly created accounts, storage slots, and deployed contract code in a separate state-gas dimension at a fixed rate of 1,530 gas per byte, counting a new account as 120 bytes and a storage slot as 64 bytes.
- •A transfer that creates a new account would cost a proposed 204,600 gas in total under the paired EIP-2780 restructuring, with the sender—such as an exchange, faucet, or onboarding service—absorbing the charge.
- •A wallet receiving a token for the first time would trigger 97,920 state gas for the new balance record, paid by whoever sends the tokens, and the charge could recur if a balance falls to zero and is later refilled.
- •Applications deploying smart wallets would bear approximately 350,000 gas or more per user through ERC-4337 factories and paymasters, and no major paymaster provider has yet published revised sponsorship pricing.
- •As of November 16, compatibility findings involving eight invariants that existing contracts rely on had delayed the Glamsterdam process on the Hoodi testnet, with potential changes to the proposal's constants scheduled for discussion on November 19.

EIP-8037, proposed for Ethereum’s Glamsterdam hardfork, would increase the cost of adding accounts, contract code and storage to the network’s state by metering those changes through a separate gas dimension. For applications onboarding customers, the key question is which records a user’s first interaction creates—and whether the customer or the business pays for them.
There are three common ways a person arrives at an Ethereum application: an exchange or faucet funds a brand-new address, an existing wallet receives a token it has never held, or the application deploys a smart wallet on the user’s behalf. Each path creates a different amount of state and assigns the resulting cost to a different party.
Glamsterdam would charge for new state, not new users
Under EIP-8037, every operation that permanently enlarges Ethereum’s state—including creating an account, adding a storage slot or writing a byte of deployed contract code—would be priced in “state bytes” at a fixed rate of 1,530 gas per byte. A new account is counted as 120 bytes, while a new storage slot is counted as 64 bytes. Other transaction activity would continue to be charged as execution gas under its existing meter. Gas is Ethereum’s unit for pricing transaction work.
The stated purpose is to allow the block gas limit to increase without causing the state database stored by every node to grow at the same rate. Unbounded growth in that database would raise the hardware requirements of running a node, which is the trade-off the proposal is designed to manage. The Hoodi testnet activation that was canceled in October had been paired with a planned increase in the gas limit to 200 million, from roughly 60 million at the time. Separating state gas would allow additional capacity to be used for computation while keeping state growth bounded.
Two of the resulting charges are especially relevant to onboarding. A companion proposal, EIP-8038, would reprice reads and writes involving state that already exists. That affects the baseline: repeat use would not remain fixed at today’s cost, so comparisons would need to apply both proposals under the same assumptions rather than treating activity involving existing accounts as unchanged.
Funding a fresh address
Generating a wallet costs nothing on-chain. As Ethereum’s account documentation explains, an address is derived from a public key, allowing a user to hold a valid address that the network has never encountered. The account record is created only when something touches that address, typically through the first ETH deposit.
EIP-2780 separates the former flat 21,000-gas transfer into a 12,000-gas base cost, 6,000 gas for carrying value and 3,000 gas for touching the recipient. A transfer to an account that already exists would therefore remain at 21,000 execution gas. A transfer that creates a new account would add 183,600 state gas, bringing the proposed total to 204,600 gas.
The sender pays that cost. An exchange processing a withdrawal to a newly generated address, a faucet, or an onboarding service that supplies users with gas would absorb the charge before the recipient had opened a single application. For a service funding thousands of first-time addresses, the state component would represent the bulk of each transaction rather than a small addition to the existing transfer cost.
Receiving a token for the first time
The second path is less obvious because the user may not be new to Ethereum. Someone who has held ETH for years could receive a stablecoin for the first time. The account already exists, but the user’s balance for that token does not.
ERC-20 defines the standard interface that fungible tokens, including stablecoins, follow on Ethereum. In OpenZeppelin’s standard ERC-20 implementation, balances are stored in a mapping inside the token contract. The recipient’s entry remains at zero until the first transfer writes a positive balance. Under the proposed rules, that zero-to-positive write would create a new storage slot costing 97,920 state gas, in addition to the execution gas required by the transfer. Adding funds to an already positive balance would reuse the slot and would not incur state gas.
Two consequences follow. First, “new user” is not the correct unit for measuring the charge. The chargeable event is the creation of a new record, and a long-standing wallet creates such a record each time it enters an application or uses a token it has not previously held. Second, the record belongs to the token contract rather than the user’s account. Receiving a token does not create an account at the recipient’s address; it creates a storage slot in another contract’s state.
The charge would not necessarily occur only once during a wallet’s lifetime. If a balance is reduced to zero and later refilled, the next zero-to-positive write would create another storage record under the proposed rules. The party sending the tokens—such as an exchange, payroll contract or airdrop—would pay the associated cost.
Sponsored smart wallets
Smart-wallet onboarding shifts these costs away from the user and onto the application. The application deploys a contract account for the customer, writes its initial settings into storage and executes the first action, often as part of a single bundle.
ERC-4337 supports this model through factories, which deploy the account, and paymasters, which cover the user’s gas using funds deposited with the EntryPoint contract. The customer does not need to acquire ETH. The trade-off is that onboarding becomes a line item in the provider’s budget.
Under EIP-8037, that line item would be dominated by state costs: 183,600 gas for the new account, 1,530 gas for each byte of code written to it and 97,920 gas for every storage slot filled during initialization.
The total depends on the wallet design. A minimal proxy that delegates to shared logic writes only a few dozen bytes of code, while a full implementation deployed separately for each user writes kilobytes. A consumer application that deploys a wallet for every sign-up would therefore have a different exposure from a decentralized finance protocol whose users primarily arrive with existing accounts.
None of the major paymaster providers has published revised sponsorship pricing. For now, each provider must model its exposure using the proposal’s specifications, its wallet bytecode and its initialization process.
Four onboarding scenarios
The three onboarding paths can be compared with ordinary repeat use to show where the state charge would arise and who would carry it. All figures below are gas under the proposed rules. The smart-wallet figure is an illustration based on the specification’s constants, not a measurement of a particular product.
- Fresh address receives ETH: 204,600 gas, consisting of 21,000 execution gas plus 183,600 state gas. The sender—an exchange, faucet or onboarding service—pays.
- Existing wallet receives its first token: an additional 97,920 state gas on top of the transfer’s execution gas. Whoever sends the tokens pays.
- Smart wallet deployed for a user: approximately 350,000 gas or more, consisting of 183,600 gas for the account, 68,850 gas for a 45-byte proxy, 97,920 gas for each initial storage slot and the required execution gas. The application pays through its paymaster.
- Repeat use involving existing records: zero state gas, with execution gas only, repriced separately under EIP-8038. The user pays.
The pattern is consistent: the party triggering the creation of the first record pays for it, and that party is usually not the person being onboarded. The case in which the user pays directly is repeat use, where no new state is created.
In monetary terms, the fee still depends on multiplying the gas amount by the gas price and converting the result using ETH’s market price. Neither value is set by the proposal. As an illustration only, at a gas price of 1 gwei and an assumed ETH price of $3,000, funding a fresh address would cost about $0.61 under the proposed rules, compared with roughly $0.14 for the same transfer today. At 20 gwei, those figures would become approximately $12 and $2.80.
Whether a larger capacity allows congestion to ease enough to reduce gas prices is a demand question that the specification cannot answer.
Status of the proposal
As of November 16, the changes remained proposals. The Ethereum Foundation’s November 6 notice, written by researcher Marius van der Wijden and updated on November 9, said compatibility findings had delayed the hardfork process and that Hoodi would not fork until the issues were resolved. The notice is available at blog.ethereum.org/2026/11/06/glamsterdam-8037-8038.
Developers were working through eight invariants on which existing contracts rely and that EIP-8037 and EIP-8038 could break. The November 9 update reported four submissions identifying deployed contracts that could be affected.
Potential specification revisions were being prepared for discussion on November 19. That was a discussion date, not an activation date, and the constants described above could change as a result. Each of the scenario figures earlier in this article is derived arithmetically from those constants, so any revision would flow directly into onboarding cost estimates. Coindoo’s earlier article on how the proposed gas rules could affect existing contracts addresses the compatibility issue.
For organizations managing onboarding budgets, the immediate task is to rerun the four scenarios using their own wallet bytecode, initialization process and gas-price assumptions. They also need to monitor the November 19 discussion for possible changes to the per-byte rate. The headline charges may be revised, but the central structure of the proposal remains that newly created state carries a cost and that the party sponsoring or triggering its creation generally pays it.