Ethereum's cryptography roadmap has become the subject of a public dispute, drawing competing positions from venture firm a16z, StarkWare CEO Eli Ben-Sasson, Gabriel Shapiro, and Ethereum co-founder Vitalik Buterin. The debate centers on whether the network should anchor its future security in hash or lattice-based cryptography, and on how Ethereum should support different signature systems. Buterin has put forward EIP-8288 as a way to verify cryptographic computations without placing their underlying data directly onchain. The backdrop matters: hash-based and lattice-based cryptography are both leading candidates in the industry-wide post-quantum transition, in which NIST has already published standards spanning both families, a context that helps explain why wallet compatibility and institutional security requirements figure prominently in the dispute.

a16z Backs Lattice Cryptography as Ben-Sasson Objects

a16z argued that hash-only designs depend on custom cryptographic constructions that sit outside NIST standards, including those planned for Ethereum's consensus and account signatures. The firm noted that the underlying hash has yet to be decided, a gap that could limit compatibility with individual hardware wallets and institutional security systems. In its view, relying exclusively on hash-based cryptography would produce more fragile systems, while adopting lattice cryptography would improve blockchain security and functionality.

Ben-Sasson, a co-inventor of the STARK proof system and the StarkWare chief executive, disagreed sharply. In a post on X, he argued that lattice cryptography has mathematical structures that artificial intelligence could potentially exploit. He also questioned the maturity of succinct constructions based on module SIS, which he said have existed only since 2012. His recommendation: wait another three to five years to evaluate the technology against potential AI attacks before entrusting it with significant funds.

Shapiro Defends Ethereum's Current Approach

Gabriel Shapiro said the exchange initially made him question Ethereum's roadmap, particularly alongside what he described as recent criticism from Justin Drake. After researching the issue, however, he concluded that Ethereum's approach remains the best option for its needs.

He observed that alternative layer-1 projects, including LayerZero, could emphasize lattice-based cryptography to gain a narrative advantage, and described the choice between lattice- and hash-based methods as a debatable strategic decision. He added that Ethereum could later support lattice-based cryptography for accounts and wallets through a precompile, provided researchers establish its security against AI attacks.

Buterin Proposes EIP-8288 for Offchain Verification

Weighing in on the discussion, Buterin said EIP-8288 could support different cryptographic systems without requiring additional precompiles. Under the proposal, users would generate signatures locally, prove them using STARKs, and submit proofs through recursive aggregation. He argued that this design would keep the underlying data and computation offchain while allowing Ethereum to verify the results.

Data capacity has been a persistent constraint across Ethereum's scaling efforts, PeerDAS and blob-carrying transactions among the designs introduced to expand it. Buterin also cautioned that blockchain data capacity remains limited. He estimated a practical ceiling of approximately eight megabytes per second, even with PeerDAS and blocks-in-blobs in place.

Whether researchers establish lattice constructions as secure against AI attacks, and how EIP-8288 advances through Ethereum's public improvement process, are the markers to watch as the debate continues.