Vitalik Buterin Proposes New Approach to Cryptography Processing in Ethereum

Ethereum co-founder Vitalik Buterin has endorsed EIP-8288 – a method to verify complex cryptography without adding new verification functions to the network. In his opinion, heavy cryptographic work should occur outside the main execution process.

EIP-8288, "Signature and Proof Aggregation in the Transaction Queue," was created by Buterin with Ethereum researcher Thomas Korajac on June 3, 2026. The idea is to combine signatures and STARK proofs before validators verify their collective validity. This will allow for quantum-resistant signatures without separate built-in functions for each type. It is currently a draft: implementation and network consensus are needed.

In publications on October 19, Buterin noted that the Ethereum virtual machine and specialized cryptographic functions are not suitable for every new system. Users will be able to generate proofs locally and submit them through aggregation: a transaction declares which statements to verify without including all the original proof in the block. The network verifies the aggregated result, covering dependencies. Buterin described this as part of a "cryptographic world computer," where the blockchain processes only what requires joint execution and verification.

He estimated a practical limit of about 8 MB/s even with PeerDAS and "blobs in blocks." Hash signatures may require 2–3 KB and 150–200 thousand gas for direct verification, while STARKs exceed 128 KB, in some configurations up to ~512 KB. Hence the sense of aggregation: not publishing large signatures and proofs as regular data.

Aggregation starts in the transaction queue: participants verify proofs, combine them, and pass the package further; the interval is 1000 ms. Block builders collect transactions and aggregated proofs, and the final recursive STARK covers the dependencies of selected transactions. Current parameters are up to 16 signatures and one STARK dependency per transaction. Lean Ethereum tools are used; the specification considers compatibility with FOCIL and relies on EIP-8141 and EIP-2718.

EIP-8288 is a draft of the main category (standard track). A reference implementation is expected, some parameters are not finalized, and there is no activation date on the main network. Nodes without updates will reject new format transactions and blocks with recursive headers – a coordinated network update is needed. Regular transactions will retain their previous behavior.

Separately, Buterin spoke about lattice signatures for accounts and wallets when confirming their security and about cryptography resistant to AI attacks – this differs from standard quantum resistance.