#以太坊基金会l1弃用poseidon哈希
Ethereum Foundation research fellow Justin Drake has recently publicly announced that, after 8 years of research and work at the million-level scale in post-quantum cryptography, the Ethereum Layer 1 (L1) roadmap will officially abandon Poseidon—“zero-knowledge-proof-friendly hashing” widely used since 2019—in favor of traditional hash functions such as SHA-2 or BLAKE2s.
This strategic shift is made possible by major breakthroughs in SNARK design. With the maturation of cutting-edge technologies such as binary field operations, the performance of conventional hash functions within SNARK circuits has improved significantly. At present, on ordinary laptops, it is already possible to verify roughly one million traditional hash operations per second. This signals that Ethereum’s technical focus is moving from “seeking ZK-friendly hashes” to “adapting ZK design to hashes.”
Analysis suggests that this move is intended to strengthen Ethereum’s security confidence in the face of future threats from quantum computing. According to the latest roadmap, Ethereum plans to release production-grade leanVM in 2027 and fully implement it across the consensus, execution, and data layers by 2028. However, this change is mainly targeted at the L1 underlying architecture. Existing applications such as various zk-Rollups and zkVMs may choose hash implementations as needed and are not subject to mandatory replacement.
Ethereum Foundation research fellow Justin Drake has recently publicly announced that, after 8 years of research and work at the million-level scale in post-quantum cryptography, the Ethereum Layer 1 (L1) roadmap will officially abandon Poseidon—“zero-knowledge-proof-friendly hashing” widely used since 2019—in favor of traditional hash functions such as SHA-2 or BLAKE2s.
This strategic shift is made possible by major breakthroughs in SNARK design. With the maturation of cutting-edge technologies such as binary field operations, the performance of conventional hash functions within SNARK circuits has improved significantly. At present, on ordinary laptops, it is already possible to verify roughly one million traditional hash operations per second. This signals that Ethereum’s technical focus is moving from “seeking ZK-friendly hashes” to “adapting ZK design to hashes.”
Analysis suggests that this move is intended to strengthen Ethereum’s security confidence in the face of future threats from quantum computing. According to the latest roadmap, Ethereum plans to release production-grade leanVM in 2027 and fully implement it across the consensus, execution, and data layers by 2028. However, this change is mainly targeted at the L1 underlying architecture. Existing applications such as various zk-Rollups and zkVMs may choose hash implementations as needed and are not subject to mandatory replacement.