Ethereum co-founder Vitalik Buterin published a long-term protocol plan on July 4, 2026, titled "Lean Ethereum," outlining a 3-to-4-year rebuild of the network's core infrastructure. The plan follows a high-level research summit in Berlin and earlier client-team discussions in Svalbard. It descri...
"If the Merge was the second evolution, Lean Ethereum is the next step. Nearly every core part of the protocol will change." — Vitalik Buterin, Ethereum Co-Founder
Ethereum co-founder Vitalik Buterin published a long-term protocol plan on July 4, 2026, titled "Lean Ethereum," outlining a 3-to-4-year rebuild of the network's core infrastructure. The plan follows a high-level research summit in Berlin and earlier client-team discussions in Svalbard. It describes the network's third major iteration after launch and the Merge, targeting five base-layer goals: faster L1 finality, gigagas L1 throughput (~10,000 transactions/second), teragas L2 scaling (~1 GB/s rollup data availability), post-quantum cryptographic security, and private L1 transfers.
The roadmap does not arrive through a single hard fork. It is structured as a series of protocol upgrades beginning with the Glamsterdam fork (expected Q3 2026) and extending through approximately 2029, when core post-quantum infrastructure is targeted for completion. ETH traded at approximately $1,792 on July 5, up over 3% in 24 hours following the announcement, on a market capitalization of roughly $213 billion.
The most structurally significant change in the roadmap is the shift from direct transaction re-execution to recursive STARK-based verification. Under the current model, every Ethereum node re-executes every transaction to verify state transitions. Under Lean Ethereum, one prover performs the computation, and all other nodes verify a compact cryptographic proof.
This architecture enables several downstream improvements. It reduces the computational burden on validators, allows for zkEVM proof aggregation across Layer 2 rollups, and creates a path toward stateless clients that can verify the chain without storing its full state. According to Buterin, the current tree structure and virtual machine together account for more than 80% of the bottleneck in efficient proof generation. Replacing both could yield efficiency gains exceeding 100x on L1.
Recursive STARKs — Scalable Transparent Arguments of Knowledge — are cryptographic proofs that require no trusted setup and are quantum-resistant by construction, since they rely on hash functions rather than elliptic-curve assumptions. Their integration as a native verification component represents a fundamental shift in how Ethereum processes and validates transactions.
Buterin stated that "quantum safety has shifted up a LOT in priority," citing urgency around quantum-safe approaches for data blobs. The quantum resistance strategy addresses four vulnerable protocol surfaces:
Consensus layer. BLS signatures, currently used for validator attestations, would be replaced with hash-based schemes. The roadmap specifically references Winternitz-family signatures, with STARKs handling aggregation and proofs.
Data availability. KZG polynomial commitments, used to verify blob data on the beacon chain, would migrate to STARK-based commitments. KZG relies on elliptic-curve pairings vulnerable to Shor's algorithm on a sufficiently powerful quantum computer.
Account-level signatures. EIP-8141, under consideration for the Hegotá hard fork (planned for late 2026 or 2027), introduces native account abstraction that allows individual accounts to select their own signature verification schemes. This enables opt-in migration to quantum-safe signatures without requiring a protocol-wide flag day.
Timeline. Core post-quantum infrastructure targets completion by approximately 2029. The approach is incremental: Glamsterdam (Q3 2026) addresses execution-layer changes; Hegotá follows with consensus and account-level upgrades. The Ethereum Foundation confirmed the structured fork milestone approach.
Google Quantum AI published research in March 2026 estimating that breaking current cryptographic signatures could require fewer than 500,000 qubits and run in minutes, adding urgency to the timeline. Current quantum computers operate with hundreds to low thousands of qubits, but the trajectory is accelerating.
The roadmap envisions a three-stage transition away from the Ethereum Virtual Machine:
Stage 1. RISC-V is introduced for precompiled contracts only, replacing the current set of hardcoded cryptographic operations with a general-purpose instruction set.
Stage 2. RISC-V (or an Ethereum-specific variant called "leanISA") opens to user-deployed contracts, running alongside the EVM.
Stage 3. The EVM is retired as a native execution environment and reimplemented as a smart contract written in the new VM. Full backward compatibility is preserved for existing contracts, with gas cost adjustments that Buterin characterized as minor relative to ongoing scalability work.
The rationale is performance and provability. RISC-V's regular instruction set is substantially more efficient to generate zero-knowledge proofs over, compared to the EVM's stack-based architecture with its irregular opcode set. This directly feeds the STARK verification architecture: a simpler VM produces proofs faster and cheaper.
Existing Solidity and Vyper contracts would not require rewrites. The EVM would continue to function as a compiler target, meaning developers write in the same high-level languages, but the underlying execution environment changes.
Buterin outlined a vision for Ethereum's state model by 2030 that separates storage into two tiers:
Dynamic state (~2 TB). The existing state tree, containing account balances, contract storage, and nonces.
Scalable state (~100 TB). A new storage layer optimized for high-volume, lower-cost data. Designed for ERC-20 token balances, NFT ownership records, and DeFi positions that currently consume expensive storage slots.
The scalable state layer would use a different tree structure — likely a binary Verkle or STARK-friendly Merkle variant — that is cheaper to prove and more efficient to access. The economic incentive structure is designed so that migration is financially attractive rather than mandatory: applications storing data in the new format would pay lower gas costs, potentially reducing certain token-transfer costs by more than tenfold, according to the roadmap.
This is an economic design choice. Rather than forcing migration through deprecation, the protocol creates a cost differential that makes the new architecture the rational default for new deployments.
Finality. The current consensus mechanism requires multiple epochs (approximately 12.8 minutes) to reach finality. The roadmap targets one- or two-round finality, reducing confirmation times to seconds. This directly impacts institutional use cases where settlement speed determines protocol selection.
Multidimensional gas pricing. The current single-dimensional gas model prices all operations on one axis. The roadmap introduces separate pricing for compute, data, and state access. This is structurally similar to EIP-4844's introduction of blob gas for data availability, but extended across the entire execution environment.
Gas limit. Glamsterdam targets a 200-million gas limit floor, up from the current approximately 60 million. This represents a 3.3x increase in per-block capacity, enabled by ePBS (Enshrined Proposer-Builder Separation, EIP-7732), which separates transaction validation from consensus and expands the data propagation window from 2 seconds to roughly 9 seconds.
"Privacy is no longer an afterthought; it is a first-class goal," Buterin wrote. The roadmap positions private L1 transfers as one of five base-layer objectives, alongside scalability, finality, quantum safety, and throughput.
The approach integrates privacy into protocol design evaluation. According to the roadmap, developers will evaluate new protocol components by asking how they can support intermediary-free, quantum-safe privacy protocols while keeping overhead low. This is a design constraint, not a feature addition.
The RISC-V transition directly supports this goal. Privacy-preserving computation requires efficient proof generation, and the simpler instruction set reduces the cost of generating zero-knowledge proofs for private transactions.
The Ethereum Foundation's restructured priorities now explicitly list privacy alongside scaling, security, censorship resistance, and post-quantum work.
Glamsterdam reached its final devnet stage in mid-June 2026, with multi-client devnets running the full slate of planned EIPs. Key components:
The upgrade is expected in Q3 2026, having slipped from an original June target. No confirmed mainnet date exists; the meta specification (EIP-7773) remains in Draft status.
Glamsterdam represents the first concrete milestone in the Lean Ethereum sequence. Hegotá follows, targeting account abstraction (EIP-8141) and initial quantum-resistance features.
The Lean Ethereum roadmap has direct economic consequences for the protocol's value flows:
Validator economics. The shift from re-execution to proof verification changes validator hardware requirements. Proving is computationally intensive but can be performed by specialized actors, while verification becomes lightweight. This may reduce the capital expenditure threshold for solo validators while creating a new market for proving services.
Fee structure. Multidimensional gas pricing will redistribute fee revenue across different resource types. Applications that are compute-heavy but data-light (and vice versa) will see different cost profiles. The 100 TB scalable state, with its lower gas costs for token transfers, compresses fee revenue from high-volume, low-complexity operations.
L2 economics. Teragas L2 scaling at 1 GB/s data availability substantially reduces rollup posting costs. This benefits L2 operators and end users but further compresses L1 fee revenue from data availability, extending the dynamic already visible post-EIP-4844.
Competitive positioning. The roadmap's 10,000 TPS L1 target and seconds-to-finality put Ethereum's performance specifications closer to newer L1 chains like Solana (currently approximately 4,000 TPS with 400ms finality targets under Alpenglow). The difference: Ethereum's approach routes through proof-based verification rather than hardware-intensive parallel execution.
The Lean Ethereum roadmap represents the most comprehensive protocol redesign since the Merge. Its scope — touching verification, execution, consensus, state storage, cryptography, and privacy — is comparable in ambition to Ethereum's proof-of-stake transition, but distributed across multiple years rather than concentrated in a single event.
The economic logic is straightforward: Ethereum's current architecture was designed for a network processing thousands of transactions per day, not millions. The protocol's fee revenue, validator economics, and competitive position all depend on whether its throughput and finality can match the demands of institutional adoption and high-frequency on-chain applications.
Whether the Ethereum development community can execute a multi-year, multi-component protocol redesign while maintaining network stability is the central risk. The Merge demonstrated this capability once. Lean Ethereum requires sustaining it across a longer timeline with more moving parts. The first test arrives with Glamsterdam in Q3 2026.