Vitalik Buterin proposed a longer-term Ethereum transaction model that separates the "actions" and "dependencies" within transactions. The actions are responsible for modifying on-chain state, while the dependencies verify that the transaction has already met signature, state proof, and validity requirements prior to execution. Under this approach, some verification work can be completed before transactions are included in blocks, enabling greater potential for parallel processing.
Transaction verification and execution will be handled separately.
The current Ethereum transaction process typically combines authorization, fee payment, and contract execution into a single processing pipeline. Nodes must simultaneously verify the validity of the signature, confirm that the sender can afford the fees, and determine whether the transaction execution succeeds.
Buterin believes that some of these checks do not depend on final state changes and could theoretically be handled separately. For example, digital signatures, zero-knowledge proofs, or certain validity proofs that do not rely on on-chain state changes can be categorized as "dependent."
If transactions clearly declare which states they will access, the mempool can more easily determine which conditions are affected by prior transactions and which checks can be completed in advance. This means that more predictable transactions may achieve higher verification efficiency.
EIP-8141 is still in draft stage.
Corresponding to this approach is a draft proposal called EIP-8141. This proposal introduces a new transaction type called Frame Transaction, which splits a transaction into multiple call frames to separately handle authorization verification, fee payment, and user operation execution.
According to the draft design, transaction validity and fee payment are no longer entirely dependent on outer-layer standard signatures. Account code can define its own authorization methods and payment rules. The verification frame confirms whether conditions are met, while the sending frame is responsible for actually modifying the state.
This structure is also seen as helping to align the underlying transaction format across different EVM networks. However, EIP-8141 remains a core draft and has not yet been included in any Ethereum mainnet upgrade, nor is there a defined deployment timeline.
During the developer discussion, several technical issues were raised, including denial-of-service risks, transaction replacement rules, wallet and block builder compatibility, and limits on the number of pending transactions from the same sender in the public mempool. These issues remain under further discussion.
Recursive STARK or reducing redundant verification
Buterin’s long-term vision extends beyond EIP-8141. He also envisions that for “pure dependencies” that do not require access to on-chain state, a preliminary check could be performed at the mempool layer, rather than being repeatedly executed by each validator.
Under this model, the network can compress multiple completed verification tasks into a recursive STARK proof, which is then verified collectively by the verifier. The goal is to reduce redundant computations and alleviate part of the verification burden.
He also noted that this approach could potentially help Ethereum adapt to post-quantum cryptographic schemes in the future, as quantum-resistant signatures are typically larger and more expensive to verify. If accounts can customize their authorization methods and are combined with recursive proof aggregation, the associated verification costs may be reduced.
However, this content remains at the research stage and is not part of the current EIP-8141 specification. To be fully implemented, challenges such as proof generation, mempool coordination, data availability, and protection against error aggregation must still be addressed.