Vitalik Buterin's Lean Ethereum roadmap commits the network to a three-to-four-year rebuild spanning seven hard forks — replacing the EVM, elevating quantum resistance to an immediate requirement, and building shielded transfers into the base layer rather than the application layer.
Vitalik Buterin published an updated Ethereum roadmap on 4 July that commits the network to replacing almost every layer of its technical stack over the next three to four years. The document, titled Lean Ethereum, was posted to ethresear.ch on Saturday and immediately reframed the debates that had been running through the client teams since March. Quantum resistance and native privacy — for years described as long-horizon goals — are now immediate engineering requirements.
The scope of what Buterin is proposing has no clean precedent. He describes it as the biggest rebuild since the Merge, and the technical shopping list makes clear he isn't overstating things. STARKs will replace SNARKs in every place the protocol currently uses zero-knowledge proofs. The execution layer that today runs on the Ethereum Virtual Machine will be rewritten around a RISC-V-based architecture. State growth — the technical debt that has quietly turned Ethereum nodes into 2TB liabilities — gets addressed through statelessness and Verkle-tree replacements. And every cryptographic primitive still relying on elliptic-curve assumptions goes.
Quantum resistance is the frame that ties it all together. Post-quantum signature schemes will replace ECDSA at the account-abstraction layer; hash-based commitment schemes replace the constructions currently underpinning bridges and rollups. Buterin's argument is not that quantum computers will crack Ethereum next year — he doesn't claim that anywhere in the strawmap. His argument is that a network being written for the next decade cannot ship in 2026 with signature schemes that a Shor's-algorithm implementation would defeat.
The seven-fork schedule laid out in the document is where the ambition meets reality. Client teams — Geth, Nethermind, Reth, Besu, Erigon — will need to ship coordinated upgrades roughly every six months to hit the four-year deadline, and each fork touches multiple subsystems at once. Glamsterdam, currently in its final devnet phase and targeted for the end of 2026, is only the second of seven. Anyone who watched the Merge slip repeatedly during 2020 and 2021 knows what happens to hard-fork timelines when subsystems collide.
Privacy is the piece that got the strongest reaction from developers over the weekend. Buterin's phrasing — that privacy is no longer an afterthought — commits the protocol to shielded-transfer primitives at the base layer, not the application layer. That distinction matters. Every previous Ethereum privacy story has ended with Tornado Cash and its sanctions blast radius. Building shielded transfers into the protocol rather than shipping them through an application means the OFAC playbook that killed Tornado becomes much harder to use again.
There's a candid piece in the strawmap about what got Buterin here. He describes years of watching Ethereum research fragment across a growing list of EIPs and semi-independent client teams, each optimising for its own timelines. Lean Ethereum is an attempt to unify all of it under one program. Whether the client teams — some of whom have publicly disagreed with the Foundation for the past six months — accept that unification is the question the roadmap can't answer on its own.
The market response was muted. Ether closed Sunday at $1,760, roughly flat on the week, and the derivatives markets showed no unusual positioning around the announcement. Institutional buyers who track protocol roadmaps have had the general direction telegraphed for months through ethresear.ch and the client-team calls; the strawmap consolidates rather than surprises. Retail attention, judging by the CT reaction, was more focused on Strategy's Bitcoin sale over the same weekend.
What the strawmap doesn't do is answer the question that has dogged Ethereum since Aave and the Layer-2 stack started pulling activity off mainnet: does a rebuild this deep still matter if the applications aren't there? Layer-2 fees are already at a few cents. Base-layer congestion has stopped being the pressure point. A four-year rebuild aimed at quantum safety and privacy is defensible on its technical merits; it also risks arriving at a network that has been substantially hollowed out by the L2s Ethereum's own upgrade path enabled.
Buterin's answer, implicit in the document, is that the network has to be worth having when the L2 economy consolidates. If a rollup with $50 billion in TVL ends up settling to a base layer whose signature scheme a quantum attacker can break, the L2 dies with the base layer. Rebuilding now — before the L2 stack is fully mature — is the version of the argument that makes rebuilding possible at all. Do it after the L2s have crystallised, and the coordination cost multiplies.
The strawmap ends without a shipping date for the seventh fork. Buterin has consistently refused to commit to timelines his client teams haven't signed off on, and there's no sign that changes here. The Ethereum Foundation, having spent the first half of 2026 losing eight senior staff and watching EthLabs launch as a rival institution backed by $11.3 billion in ETH, is in no position to promise anything the client teams might not deliver. What the strawmap does deliver is a target — and a target is more than the Foundation has offered publicly in a year.
