What Ethereum developers decided
The Ethereum Foundation said All Core Developers agreed to delay the Glamsterdam hard-fork process and not activate it on the Hoodi testnet until issues around EIP-8037 and EIP-8038 are resolved. Work on the later Hegotá upgrade continues in parallel.
The decision followed findings from ChainSecurity and other reviews showing that some deployed projects rely on gas behaviours that the proposals change. The Foundation is engaging ChainSecurity and Dedaub, running internal audits and asking project teams to report affected code privately.
What the two proposals change
EIP-8037 separates state-creation charges from ordinary execution gas and introduces a state-gas reservoir for very large transactions. EIP-8038 reprices access to accounts and storage. Together they change both how expensive certain operations are and how much gas may remain after a failed or reverted call.
The goal is not to create a new token or change ether’s supply. The risk comes from applications that treat today’s gas behaviour as a security invariant—for example, code that assumes a failed sub-call leaves too little gas to write state.
The eight assumptions under review
The Ethereum-maintained audit repository lists eight broken invariants. The most common findings so far involve fixed gas budgets and gas reserves that were sized for today’s costs.
- A failed call may leave enough gas for the outer contract to record a permanent result.
- Clearing state can refill gas immediately instead of only creating a later refund.
- A reverting sub-call can return state gas that callers assumed was spent.
- Fixed limits for storage writes, account creation, relayers or signed messages may become too small.
- Subtracting two
gasleft()readings may no longer measure the true cost. - Wallets and off-chain systems may wrongly treat a mined EIP-7702 delegation as successfully applied.
These are design risks, not evidence that every Ethereum contract is vulnerable. A project is affected only if its code or off-chain services depend on one of the changed assumptions.
What teams should do now
- Search contracts and off-chain services for hard-coded gas limits, post-call
gasleft()checks and fixed relayer reserves. - Test failure paths under the proposed Glamsterdam rules, including proxies and nested calls.
- Check wallets and services that use EIP-7702 delegations or infer success from transaction inclusion alone.
- Preserve retry paths rather than permanently marking work complete after an ambiguous failed call.
- Report affected projects to the Ethereum Foundation using the private contact in the official notice.
The relevant question is not “does my contract use gas?” It is “does safety depend on gas behaving exactly as it does today?”
What remains unknown
There is no new Hoodi activation date. Developers may keep the proposals, modify them or adjust the wider fork plan after the review. The number of affected production projects is also not public, partly because responsible disclosure is being handled privately.
Ordinary users do not need to move ether, replace a wallet or sign an “upgrade” transaction. Any such request is inconsistent with the Foundation’s developer-focused notice.
Sources you can open
01This is a developer review. There is no public user claim page, wallet migration or token distribution associated with the Glamsterdam delay.