In a post-quantum move, Ethereum L1 moves Poseidon to the side.

Justin Drake, a researcher at the Ethereum Foundation, has stated that layer one of the Ethereum network will abandon Poseidon following an eight-year research project costing eight figures and instead use established hash functions like SHA-2s or BLAKE2s.

Poseidon turns its back on Ethereum L1

Justin Drake, in a post on Aug. 13, X, said that Ethereum Foundation has abandoned Poseidon and is moving towards more traditional options.

“Goodbye, Poseidon!” Drake wrote that the decision was the result of an investigation. “8-year, 8-figure rabbit hole” In post-quantum Cryptography.

Poseidon, a hashing function developed for proofs of zero knowledge systems, was released in 2019. Because of its structure, it was less expensive to process within Succinct Non-Interactive Arguments of Knowledge, also known as SNARKs.

In 2018, as part of the work it is doing on zero-knowledge, the Ethereum Foundation invested in special hashes. Poseidon was later a popular choice among zk-rollups (zkVMs) and systems that protect billions in cryptoassets.

Drake’s announcement was more about Ethereum’s future L1 Architecture than a removal of Poseidon immediately from applications. The roadmap has decided that it is not necessary to remove the functionality from existing projects, such as virtual machines and rollups.

Poseidon’s initial advantage has now been eroded by advances in the proof-system design. Drake claims that SNARKs, which are based on binary computations, can be compared to the performance of SHA-2 or BLAKE2 functions.

“In hindsight the key was not SNARK-friendly hashes, but hash-friendly SNARKs.”

The binary-field SNARKs makes traditional hashes more convenient

Binary fields enable proof systems more easily to handle the Boolean logical used in standard hash functions. Earlier SNARK design often relied upon large prime fields where it was costly to represent bit-based operations like XOR.

Binary-field computing systems can align their computations with zeros and one in traditional computing by using the two smallest prime numbers. Drake stated that the resultant designs could prove approximately 1 million hash calls in a second using a laptop with an overhead roughly 100 times greater than native CPU execution.

According to the article, research projects such as Binius and Flock have contributed to performance improvements. Binius uses binary-field arithmetic for zero-knowledgeproofs. Flock focuses mainly on large-scale Boolean computations including SHA-256 and Keccak.

Drake also mentioned SNARK.fast. This is an open, automated project using artificial intelligence (AI) to enhance proving codes. Its best result achieved 1.8 million BLAKE3 compressed per second. That’s a 255% gain over its benchmark.

By using established hashes, Ethereum could be less dependent on cryptographic functions requiring years of analysis. SHA, BLAKE, and other zero-knowledge cryptographic systems have been extensively studied, though their integration into Ethereum still requires research, audits, testing, and more.

It also alters the relationship of Ethereum’s proof infrastructure and hash functions. Researchers could build SNARKs around existing hash functions with a long track record of security, rather than designing a brand new hash to fit the limits of a SNARK.

Hash-based hash signatures are preferred by Ethereum’s post-quantum works

Drake tied the decision back to Ethereum’s Post-Quantum Security Program, which develops the protocol needed for computers that can break elliptic curve cryptography. Ethereum uses such systems as a basis for its user accounts as well parts of data and consensus infrastructure.

The crypto.news reported in AugustVitalik Buterin, Ethereum co-founder has pushed quantum security up the updated roadmap. Plan also includes native privacy, formal validation, post-quantum scalability, and potential replacements for certain parts of Ethereum Virtual Machine.

Drake stated that recent advancements in AI-assisted cryptoanalysis has created setbacks to more complex systems. He cited HAWK (a signature design based on lattice) and SQIsign (which relies isogeny based cryptography).

In his opinion, these problems support the use of hash-based digital signatures. The designs employed are relatively straightforward and well-researched, even though the individual signatures may be too big to use directly at Ethereum’s scale.

SNARK aggregation can help solve the size issue. Proof systems can compress and verify multiple hash-based digital signatures into a single proof, reducing the size of the proof. This is more efficient than checking each signature individually.

This method can also be used to support multisignature agreements and threshold signatures k-of n, where a particular number of parties must approve a given transaction. Drake explained that the flexibility is due to the SNARK, which can be used to verify the rules of authorization without putting all the underlying data onto the blockchain.

The wallet is already in testing. An Ethereum researcher published his findings in June. demonstrated account protection A SPHINCS based signature verifyor was employed, where one optimised version required approximately 127,000 gas while carrying a 3,704 byte signature. The cost to verify a single account was estimated at $0.07 at that time by the researcher.

U.S. standards increase pressure on early preparation

The roadmap is a response to a concern that U.S. businesses and investors using Ethereum have been addressing as part of a migration plan. Federal standards agencies are already tackling this issue. In August 2024, the National Institute of Standards and Technology released its three first post-quantum encryption standards and encouraged administrators to start integrating them.

NIST standards are not used to guide Ethereum’s choice of protocol, nor do the algorithms selected by the NIST agency make it quantum-safe. However, its migration work shows that American organizations are prepared before cryptographically important quantum computers become readily available.

The independent Coinbase advisory board came to a similar conclusion in a fifty-page document published in April. Drake was one of its members, along with Stanford cryptograph Dan Boneh and EigenLayer’s Sreeram Kanan. Coinbase cryptography chief Yehuda Lindell, as well as distributed systems researcher Dahlia malkhi, were also on the board.

Board members concluded that blockchains were secure from attacks by quantum computers, but cautioned that the process of replacing signatures that are vulnerable across exchanges, networks and wallets would take several years. According to the advisory paperSome quantum-resistant alternative solutions could raise the blockchain data requirement by up to 38 times.

Ethereum is using proof aggregation in order to reduce costs and replace cryptography that’s vulnerable. Drake stated that hash-based SNARKs can compress any number of post quantum signatures to a compact proof, suitable for block inclusion.

LeanVM targets production deployment in 2027

The Ethereum Post-Quantum team has developed binary-field technology as part leanVM. This is a zero-knowledge, minimal virtual machine that can verify cryptographic evidence and aggregate it.

Drake confirmed that current plans call for the production of a leanVM by 2027. Ethereum’s data, execution, and consensus layers will be deployed in 2028. However, each change to the protocol still requires testing and approval from Ethereum’s development teams.

This schedule is part of Ethereum’s Strawmap. It is a document for technical coordination that extends through 2029, not a calendar. The earlier Strawmap coverage There are seven forks that have been proposed, including faster slots and shorter finality.

Ethereum’s Post-Quantum Team is currently working on Binius, Flock and binary-field-related systems, while also developing leanVM Benchmarks. Drake explained that 2028’s work will apply a proof-of-concept technology to the three different layers of the blockchain: data, execution, and consensus.