EIP-8141 remains a Draft proposal. A pull request created inside the research window was merged after the Agent 2 handoff, changing the wording around a 100,000-gas value without establishing client adoption, fork inclusion, or mainnet behavior.
The proposal in one sentence and its current status
Draft EIP / merged wording proposal — no deployment established. Pull request 12301 was created on 6 September 2026 at 01:20:27 IST and proposed treating MAX_VERIFY_GAS as a shared public-mempool floor rather than a single hard per-node ceiling. The numerical value remained 100,000. The pull request subsequently merged, so the earlier “open PR” label is no longer current; EIP-8141 itself remains Draft.
A merge changes the proposal text. It does not establish that clients implemented the design, that a network upgrade included it, or that Ethereum mainnet uses the rule. This page therefore treats the merged text as an open standards proposal rather than a live operating instruction.
What EIP-8141 frame transactions are trying to define
The Draft describes frame transactions, including validation and payment relationships that can be expressed through ordered execution frames. Its public-mempool section aims to bound initial validation work and reduce denial-of-service exposure before a node accepts and propagates a transaction.
The design is broader than one constant. It describes a validation prefix, restricted execution patterns, state-access limits, signature-validation accounting, and conditions under which a payer is established. The shared-floor discussion concerns one policy boundary inside that larger draft.
What MAX_VERIFY_GAS bounds
Within the proposal, MAX_VERIFY_GAS accounts for public-mempool validation work across the validation prefix, including signature validation. The current value is 100,000. This is not the gas limit for an ordinary completed transaction, a fee quote, or a promise about how much work every local node will accept.
Public-mempool admission and block validity are also different layers. A mempool rule controls what a node accepts or propagates before inclusion. Consensus rules determine whether a block and its transactions are valid. Conflating those layers can turn a propagation-policy proposal into a false claim about protocol execution.
Ceiling versus shared-floor interpretation
The motivating observation was that validation cost can differ among clients, libraries, hardware profiles, storage modes, and partially stateless configurations. A single universal per-node maximum can therefore be too rigid. The proposed two-level model requires public-mempool nodes to accept and propagate validation prefixes within a common floor while allowing a node to admit more work as local headroom.
Local headroom would widen only that node’s acceptance policy and would not change block validity. The information gain is conceptual: the same 100,000 number can retain its value while its normative role changes from a maximum for every node to a minimum common propagation budget.
Why client cost differences motivated the change
Validation does not have identical operational cost everywhere. Signature and pairing libraries, parallelism, cache state, storage design, and proof-oriented node modes can make the same gas-accounted path consume different resources. The authors argued that a floor should be chosen around the weakest profile the network intends to keep in the shared inclusion-list set.
This rationale still requires implementation review. Peering behavior, transaction announcements, simulation, revalidation after state changes, and denial-of-service resistance all affect whether a proposed floor produces consistent propagation. A standards text cannot substitute for client measurements and adversarial testing.
Measurements and application calculations remain attributed
The related research reply supplied rationale and application-oriented calculations, including discussion of privacy applications. Those measurements were author-reported and were not independently reproduced by CryptossInsights. They should be read as inputs to review, not guarantees that a particular application will propagate or perform consistently.
No statement here promises privacy, lower fees, inclusion, or production suitability. The relevant review question is whether the proposal’s accounting and minimum guarantee can be implemented consistently across the intended client set.
Draft, review, implementation, and activation are separate stages
EIP-1 distinguishes Draft, Review, Last Call, and Final stages. Repository review and a merged wording change can improve a Draft without making it final. Fork consideration, client implementation, testing, release, network selection, and activation would each need their own evidence.
The page will change status only when those records justify it. For background, read CryptossInsights’s protocol-stage method, smart-contract risk guide, smart-contract glossary, and data method. The proposal has no identified effect on Indian VDA law or tax.