Ethereum co-founder Vitalik Buterin published a new set of takeaways on “Lean Ethereum,” the multi-year plan to rebuild most of the network’s protocol, in a post on X on July 4.

Buterin said the update follows a meeting of Ethereum researchers in Berlin two weeks earlier, which continued discussions held with client teams in Svalbard in April. The revised roadmap, known as the strawmap, was published alongside the post.

Buterin called Lean Ethereum “the third major iteration of Ethereum in the same way that the Merge was the second,” adding that “almost every major piece of the protocol will be replaced” over three to four years. He said the rollout is designed, like the 2022 Merge, to minimize disruption to existing applications.

Lean Ethereum

According to Buterin’s post, the plan replaces direct transaction re-execution with verification through recursive STARKs, a cryptographic proof system he said would become “an enshrined first-class core component of the protocol.”

Other listed changes include swapping quantum-vulnerable cryptography for quantum-safe alternatives, decoupling the available chain from finality to enable one- or two-round finality, introducing multidimensional gas, altering what types of state the network supports, and changing client architecture.

Buterin also said Hegotá — referred to in the post by its internal codename H-star — is “probably Ethereum’s last thematically ‘pre-Lean’ fork.” Starting with the following fork, internally named I-star, most future upgrades will carry what he called “a very strong ‘Lean’ feel.”

Quantum Safety and Privacy

Buterin wrote that quantum safety “has shifted up a LOT in priority,” making a quantum-safe design for blobs — the temporary data storage layer-2 networks rely on — “urgent.” He said that work has been ongoing for months. The Defiant has previously reported on Ethereum researchers’ efforts toward this goal, including a post-quantum key registry and the Ethereum Foundation’s post-quantum research hub.

Privacy, Buterin said, “is no longer an afterthought, it is a first class goal.” He said that when designing new elements such as Frames, the mempool, or additions to the state tree, researchers now ask how “quantum-safe, intermediary-free privacy protocol transactions” would pass through them, and at what overhead.

Buterin also tied the plan to formal verification, saying it would allow the protocol to become more comfortable with canonicalization — defining protocol pieces directly as bytecode in a specific language. He pointed to evm-asm, which he said is being written in part to serve as a canonical proof system for the EVM.

State Redesign

Buterin described changes to Ethereum’s state, the running record of account balances and contract data, as “probably the single most disruptive part of the plan.” He said there is growing consensus around keeping today’s flexible “dynamic” state largely unchanged while scaling it only moderately, and adding a new, more restrictive type of state designed to scale much further without requiring block builders to sync or store all of it.

As an example, Buterin described a possible Ethereum in 2030 holding 2 terabytes of present-day-style dynamic state alongside 100 terabytes of the new, more scalable state type. He said the new format would suit ERC-20 tokens, NFTs and many DeFi applications, but not “highly central” objects such as Uniswap contracts or onchain order books.

No application would be required to migrate, Buterin said, but doing so could be “very cost-effective”: rewriting an ERC-20 token to use a new UTXO-based storage design currently under exploration could cut its transaction fees by more than 10x. He listed keyed nonces, ring buffers, UTXOs, statically accessible state and temporary state as current ideas for the new state types, and said the design will need extensive feedback from application developers, including those building privacy-focused applications.

Buterin separately flagged the incentive structure behind storing a much larger total state size as a “first-class research area,” noting that simply requiring each node to store a fixed percentage of data doesn’t explain why they would be willing to serve it.

On execution, Buterin said Ethereum will eventually need a virtual machine beyond the EVM — at minimum something like leanISA to support recursive STARKs — with leanISA and RISC-V as the leading contenders. His stated preference is for the EVM to become a high-level compiler feature while the protocol interacts directly with RISC-V or leanISA, though he said that shift “is still far away.”

Buterin closed the post by saying gas limit increases, blob increases and slot-time decreases will recur over roughly the next five years, with a large gas limit increase expected alongside the upcoming Glamsterdam upgrade.

Mixed Reactions

Reaction on X centered on the roadmap’s ambition and its timeline. Dankrad Feist, an Ethereum Foundation researcher whose work on data sharding gave danksharding its name, wrote that the strawmap “has lots of REALLY COOL features” and that “fully proven STF and scaling to Gigagas with finality in seconds gets me excited,” but argued the three-to-four-year timeline “is very slow.” He said the Foundation “should be ambitious and get it done in ~1 year,” calling that “realistically possible now with LLMs.”

Matt Liston pushed back on compressing the timeline publicly, writing that while a two-year delivery “seems between possible and likely,” it “would be irresponsible” for Buterin or the Foundation to communicate a one-to-two-year expectation, adding that “underpromising” is the safer approach.

DeFi analyst Ignas framed the plan as “bullish for $ETH… if only the EF shipped on time,” pointing to the Merge, which he said was “‘six months away’ for about four years.” He wrote that the roadmap “addresses all (except one) key feedback” from the market — L1 reclaiming execution from layer-2 networks, privacy, quantum resistance and faster finality — but said Ethereum’s tokenomics remain unaddressed, calling it a “non-issue if reduced fees attract more txs/users.” Ignas also said the most significant parts of the plan arrive in 2028 and beyond, with finality targeted for 2029, and warned that delays help competitors such as Tempo and Canton compete for institutional and real-world-asset adoption.

Seven Forks

The strawmap was first introduced in February by Ethereum Foundation researcher Justin Drake, outlining seven forks through 2029 organized around five “north star” goals: a fast L1, a “gigagas” L1 capable of roughly 10,000 transactions per second, a “teragas” L2, a post-quantum L1, and a private L1. The Defiant reported on the original strawmap at the time.

Buterin’s update lands roughly a week and a half after the Ethereum Foundation cut its annual budget by about 40% and eliminated 54 roles, or 20% of its staff, as part of a restructuring meant to turn the nonprofit into a leaner, endowment-style organization. Buterin has said the Foundation’s technical ambitions for the strawmap remain intact despite the reduced budget.