The Ethereum Foundation is accelerating plans to move the network toward post-quantum cryptography, according to researcher Justin Drake, who said progress in quantum computing and AI is increasing the pressure on Ethereum’s current security assumptions.

The updated effort, described as “Mission 2028,” aims to protect Ethereum’s execution, consensus and data layers before the JAR fork, with a mainnet Glamsterdam fork targeted for January 2027 and a broader transition completed by December 2028.

AI and quantum risks are driving the timetable

Drake said Ethereum’s cryptographic stack faces growing risks from two directions: future quantum attacks and the more immediate impact of AI-assisted security research. Ethereum client developers have reported 895 potential issues since March that were found with AI assistance, and 500 of those were confirmed as real vulnerabilities.

In response, the Ethereum Foundation’s security team has increased its staffing and added researchers focused on AI. Drake also pointed to Ethereum’s client diversity as a useful defensive advantage as the network prepares for more intensive security work.

Mission 2028 covers all three core layers

The current plan is to complete the transition by December 2028 using hash-based cryptography and formal verification. Rather than limiting the work to one part of the protocol, the Foundation now wants all three layers secured before the JAR fork: post-quantum transactions for the execution layer, post-quantum attestations for the consensus layer, and post-quantum data sampling for the data layer.

The leanVM project is expected to play a central role in these changes. Drake also said there has been progress toward end-to-end formal verification, which is intended to support the safety of the redesign as multiple parts of the protocol are updated.

New cryptographic tools are being lined up

The roadmap assigns specific replacements to major components of Ethereum’s existing cryptography. LeanSphincs is planned to replace ECDSA on the execution layer, leanXMSS is intended to replace BLS on the consensus layer, and leanDA is set to replace KZG on the data layer.

A separate component, LeanPCS, is described as a post-quantum polynomial commitment scheme that would help Ethereum move from Groth16 and Plonk toward post-quantum STARK proofs. Together, these systems form the technical base of the proposed transition.

Rollout milestones and immediate preparation steps

Two milestones anchor the timetable. The first is the Glamsterdam fork on Ethereum mainnet in January 2027. The second is the completion of the post-quantum and post-AI cryptography transition by December 2028. Drake said the changes would be introduced over the 24 months after Glamsterdam through three hard forks.

The Foundation is already urging different parts of the ecosystem to prepare. Validators are being asked to review legacy 0x00 withdrawal credentials, which are expected to be deprecated, and to participate in work on the PQ Validator Registry. Application developers are being told to examine any hard dependence on ECDSA, while wallet teams are being asked to prepare for Frame Transactions, which are expected to become an important protection method at the execution layer.

Interim protections are available, but not yet cheap

The article notes that about 72% of ETH is held in accounts that have signed at least one transaction, meaning their public keys have already been exposed. As an interim measure, one option is to use a fresh account that has not signed any transactions so the address hash is shown instead of the raw public key.

Another current option is to use Frame Transactions with a smart contract that verifies leanSphincs signatures. That would provide post-quantum protection today, although the source says it remains more expensive until leanVM is integrated into the protocol. For now, the next confirmed steps include community feedback on leanSphincs parameters and testing on the Daugi network.

Source: incrypted.com