The Ethereum Foundation is dropping the Poseidon hash function from its base layer, ending an eight-year research effort. Researcher Justin Drake announced the decision on August 13, with the network now expected to lean on traditional hashes like SHA2 or BLAKE2s instead.
Why Poseidon is out
The shift comes down to SNARK design. New proof systems built over binary fields — work that traces back to Binius research by Benjamin Diamond and Jim Posen in 2023, plus a newer effort called Flock — let conventional hashes match Poseidon's performance inside SNARKs. Drake said the new designs can prove roughly 1 million traditional hash calls per second on a laptop. That's about 100x slower than native CPU compute, but it's fast enough to make the tradeoff worth it.
The case for plain hashes
Drake's argument is about assumptions. Poseidon and similar constructions lean on algebraic structures that carry more risk than a plain hash. Plain hashes rest on minimal assumptions, and that's the appeal. He pointed to recent blows against lattice-based HAWK and isogeny-based SQIsign signature schemes as evidence that exotic structures can crack under scrutiny.
What the timeline looks like
The rollout is gradual. A production-grade leanVM is penciled in for 2027, with consensus, data, and execution layer deployments following in 2028. Nothing is happening overnight.
What doesn't change
The move doesn't make Poseidon obsolete, and it doesn't force existing projects to migrate. This is a base-layer decision, not a mandate. Projects already running on Poseidon can keep doing what they're doing.
The 2027 leanVM target is the next concrete milestone. Whether the binary-field proof systems hold up at scale between now and then is the open question.
