Ethereum Foundation abandons Poseidon hash for layer 1

On August 13, Ethereum Foundation researcher Justin Drake announced that the network's future layer-1 design will abandon Poseidon, a hash function built specifically for zero-knowledge proofs, in favour of the standard SHA-2 and BLAKE2s functions. The reversal follows eight years and, by Drake's own account, eight figures of research into proof-friendly cryptography.

The shift reflects progress in proof systems rather than any flaw in Poseidon: newer binary-field SNARKs, developed through research projects including Binius and Flock, now handle conventional hashes efficiently enough that a specialised hash no longer offers a decisive advantage, while carrying a smaller and better-studied attack surface. No existing Ethereum application uses Poseidon on layer 1, and no migration has been ordered; the change concerns Ethereum's planned leanVM, targeted for production readiness in 2027, with deployment to follow in 2028.

Choosing well-studied cryptography over a purpose-built alternative, this late in an eight-year research effort, is the kind of reversal that looks obvious only once the tools around it have changed.