This is how Ethereum's quantum security will work according to Thomas Coratger
Ethereum splits its quantum defense into execution, consensus and data. We explain leanXMSS, leanVM and the PQ registry with clear figures and timelines.
This is how Ethereum's quantum security will work according to Thomas Coratger
Imagine your house has three different doors and each one needs a new lock before a forecast storm. Ethereum faces a similar situation in October 2026, with three layers that must become resistant to quantum computing before the risk matures. In January 2026 the Foundation created a dedicated post-quantum security team led by Thomas Coratger, with public work at pq.ethereum.org. That team coordinates more than ten client teams in weekly interoperability devnets and publishes its roadmap by forks.
The word qubit often sounds distant, but the calendar is already concrete and affects wallets and validators. In March 2026 Google Quantum AI estimated that breaking the 256-bit elliptic curve could require about 1,200 logical qubits, roughly 20 times fewer than previous estimates, as summarized by ethereum.org. Current hardware handles a few thousand noisy physical qubits, and each logical qubit requires many physical ones for error correction. The gap remains wide, though it is narrowing faster than predicted three years ago.
Ethereum's solution is not a single switch but three coordinated paths with different tools and their own pace. The execution layer bets on flexible accounts that change signature scheme without a fork per scheme. The consensus layer prepares hash-based signatures with aggregation through succinct proofs. The data layer adapts sampling and commitments for much larger and heavier signatures.
What does each Ethereum layer protect?
The execution layer is where you sign your transactions and where your accounts and contracts live. Today most use ECDSA over secp256k1, with small keys and cheap verification on every node. The future risk is that a powerful machine derives your private key from your already visible public key. The defense consists of giving each account the ability to choose a new scheme without waiting for the entire network.
The consensus layer is where validators vote on blocks with BLS signatures that are aggregated into a single proof. BLS allows combining hundreds of thousands of votes into just 96 bytes, an efficiency hard to match. The problem is that BLS relies on elliptic pairings sensitive to quantum advances. The defense requires replacing the entire scheme through forks and with very demanding performance tests.
The data layer is where the blobs and the proofs that sustain availability for layer twos travel. Its current commitments use KZG, with mathematical assumptions that must also be reviewed against the quantum horizon. The challenge is keeping bandwidth under control when signatures grow from 96 bytes to about 3,000 bytes. The defense combines efficient sampling and new commitments compatible with succinct proofs.
Execution for accounts and contracts with flexible per-user signatures
Consensus for validator votes with aggregated signatures in each slot
Data for blobs and availability with light, verifiable commitments
A clearer way to see Ethereum's future post-quantum protection
This split explains why the timeline speaks of 2029 for core infrastructure and of extra years for full migration. Changing an account is an individual decision that can be done in waves. Changing consensus requires thousands of nodes to adopt the same scheme at the same time.
Individual accounts that migrate in waves with help from their usual wallet
Collective consensus that requires coordinated adoption across thousands of nodes at once
Intermediate data that accompanies both paces without slowing the peer-to-peer network
The answer lies in native account abstraction driven by EIP 8141 and its frame model. Each account will be able to define its own signature verification logic within the protocol's rules. A user will adopt a post-quantum scheme when their wallet supports it and explains it clearly. The network does not need to halt everything to move millions of accounts on the same day.
The planned process happens in three phases that already appear in the public documentation from October 2026. First comes preparation with verification precompiles and wallet tools for controlled tests. Then gradual adoption advances, with accounts that migrate of their own accord and with help from their usual program. Finally comes protocol consolidation, when the new schemes become the normal option for new accounts.
The candidates under study include Falcon and Dilithium in a first line of work. SPHINCS appears as a hash-only alternative with other trade-offs between size and speed. Choosing too early exposes you to later weakness or to denial problems from large keys. Choosing too late leaves users exposed during the risk window. That is why agility is as valuable as the final algorithm for the mainnet.
The official guidance maintains that funds are safe today and that the wallet program will guide each step, as explained on the quantum resistance page. You do not need to move funds in October 2026 or change wallets because of headlines about artificial intelligence. You should update only from official sources and keep tested backups. The real migration will arrive with clear notices inside your own application.
Why does consensus need leanXMSS and leanVM?
Consensus cannot copy the flexible execution recipe because its signatures must be combined in each slot. BLS allows adding signatures with a simple elliptic operation and keeping the network light. No post-quantum scheme offers that native aggregation at the same cost. Replacing it outright would blow up the size of each block and the propagation time across the network.
The answer combines two complementary pieces that already have names in the roadmap. leanXMSS is a hash-only signature scheme for validators, with security backed by SHA or BLAKE. leanVM is a minimal virtual machine that generates succinct proofs that many leanXMSS signatures are valid. Together they compress the data by a factor close to 250 and recover the lost efficiency.
The figures help understand the technical leap the network takes in each slot. A BLS signature takes 96 bytes and is aggregated by direct addition. A hash signature is around 3,000 bytes and requires a proof to be compressed. Measured aggregated proofs are between 127 and 344 KiB depending on configuration, with a network target of around 256 KiB per second. Each design must meet that budget without pushing out modest operators.
Piece
What it contributes
Key figure
leanXMSS
Validator signatures with hash only
About 3,000 bytes per signature
leanVM
Aggregation with succinct proofs
Compression close to 250x
Registry in I
Post-quantum keys alongside BLS
Basis for attestations
Attestations in L
Post-quantum votes in real time
Full economic finality
EIP 8292 also defines the aggregator role to separate the heavy proof from block production. Any node with sufficient hardware can offer itself as an aggregator and prove signatures within the slot interval. Ordinary proposers and validators continue to cheaply verify a single proof. The design protects home operators and avoids centralizing production in a few machines.
The data layer sustains the availability that scaling solutions and light clients need. When signatures grow, blobs and proofs also weigh more on the peer-to-peer network. The plan introduces adapted sampling called leanDA and secure commitments for blobs in stage M. Each change seeks to keep propagation under one second with stable-size proofs.
The intermediate milestone called MV-PQ allows continued operation during Q-day with reduced guarantees. It includes a post-quantum heartbeat in consensus, leanDA sampling in data and leanSPHINCS transactions in execution. Researchers are still precisely defining which guarantees are relaxed and for how long. The full goal remains total resistance across all three layers by December 2029.
Formal verification accompanies the whole process as a cross-cutting tool and not as a final ornament. The Foundation is funding with 20 million a three-year project to verify zero-knowledge virtual machine components, as detailed by the post-quantum team. That investment covers primitives that sustain signature aggregation and proof compression. Without that mathematical check, every aggressive optimization would introduce doubts about the system's soundness.
The fork-by-fork timeline orders the pieces so each team knows what to deliver and when to test. The key registry arrives in I alongside the current BLS keys. Verification precompiles arrive in J for contracts and wallets. Attestations and real-time proofs arrive in L with leanVM. Full aggregation and secure blobs close out in M and in later stages with open proofs.
Registry in I to coexist with BLS without breaking daily operations
Precompiles in J so contracts and wallets can test new signatures
Attestations in L to activate post-quantum votes with full finality
What to watch if you stake from Europe
Staking looks the consensus layer in the eye because every future vote will use the new scheme. Validators will need to register post-quantum keys alongside the current ones when the registry in I is available. They will also need to follow the hardware requirements for aggregators if they wish to take on that role with heavy proofs. Those who delegate to providers should ask about their migration plan and their devnet tests.
Wallet and layer two users will see more gradual changes guided by their applications. Wallets will add support for new schemes and explain the migration within their usual flow. Layer twos will inherit commitments and proofs from the main chain without requiring technical actions from the end user. The official recommendation in October 2026 is not to act on your own or move funds out of fear.
The debate over dormant accounts remains open and affects the European governance of the ecosystem. No one has decided how to treat wallets whose owners do not respond when the migration arrives. Forcing a move clashes with censorship resistance and with property rights. Leaving them exposed keeps a systemic risk if a theoretical advance allows them to be drained.
The three-layer defense turns a huge problem into manageable pieces with clear owners. Execution gives users flexibility without collective rush. Consensus provides shared security with demanding proofs and efficient aggregation. Data sustains scalability without breaking peer-to-peer propagation. Until 2029, each fork will bring those pieces closer with public validation in Europe and worldwide.