Vitalik Buterin's Ethereum Roadmap Prioritizes Quantum Readiness and Privacy

Vitalik Buterin overlaid his 2023 Ethereum roadmap diagram onto the Ethereum Foundation's current strawmap on Monday and said the most striking part of the comparison is the set of items that were not on the 2023 diagram in any form: privacy, defenses against quantum computers, and an execution engine that reaches past the Ethereum Virtual Machine.

The overlay documents how far Ethereum's plans for its own base blockchain have moved in under three years. Quantum safety was pulled forward, and two projects that anchored the 2023 diagram were replaced outright. Buterin also tied Ethereum's technical direction to two dependencies that carried nowhere near this weight in 2023: a family of cryptographic proofs called STARKs, and the use of artificial intelligence to mathematically verify the protocol's code.

The strawmap — a combination of "strawman" and "roadmap" — is the Foundation's draft plan for upgrading Ethereum's base layer, published in the open and revised roughly quarterly. It is maintained by the Foundation's Architecture team of Justin Drake, Thomas Thiery, Toni Wahrstätter and Buterin, and lays out seven network upgrades by 2029 at a rough cadence of one every six months.

The document sets five long-range targets it calls "north stars." Transactions should be confirmed and settled with finality in seconds. Ethereum itself should handle 10,000 transactions per second, using zero-knowledge proofs of the Ethereum Virtual Machine, or zkEVMs, which let a computer check that a block was processed correctly without redoing the work. Layer 2 networks should reach 10 million transactions per second. The cryptography should hold for centuries against quantum computers. And the base chain should support shielded transfers that hide payment details.

"In general, a lot of overlap, but: Some things got reshuffled in order (eg. quantum safety up-prioritized). Some things deprioritized (eg. VDFs; many EVM improvements). Some things replaced with superior constructions (eg. Verkle -> unified BT -> PBT; state expiry -> new state types)," Buterin wrote.

VDFs are verifiable delay functions, cryptographic timers intended to make Ethereum's randomness harder to manipulate. Verkle trees were a redesign of how Ethereum stores account balances and contract data, meant to shrink the proofs needed to check any single piece of it. State expiry was a scheme for letting the network forget data nobody had touched in years. All three sat on the 2023 diagram.

What Wasn't There In 2023

Buterin listed seven strawmap items with no 2023 antecedent. Privacy now gets what he called "first-class attention," covering keyed nonces, recent roots, a lean privacy pool and wormholes — building blocks for hiding who is transacting with whom — plus parts of FOCIL, short for fork-choice enforced inclusion lists, a design that forces block producers to include transactions the rest of the network has already seen. FOCIL is filed as EIP-7805, one of the Ethereum Improvement Proposals through which changes to the network are specified and debated.

The second new item is scaling built for a post-quantum world. It covers leanSPHINCS, a signature scheme whose security rests on hash functions rather than the elliptic-curve math a quantum computer could break, and what Buterin calls zkzk frames. Those are the subject of EIP-8288, which he opened himself on June 5. Post-quantum signatures are far larger than today's, so the proposal lets a transaction declare its signatures and proofs as dependencies that a block builder can collapse into a single recursive proof, holding down the gas cost — the fee users pay for computation on Ethereum.

Four more items round out the list: simplifying the specification so the whole protocol can be formally verified, futures contracts on gas and on blob space — the cheap storage slots layer-2 networks buy to post their data — native rollups, and a wider design space for what comes after the Ethereum Virtual Machine. On two of them Buterin was explicit about why they are new. SNARKs — succinct non-interactive arguments of knowledge, the proof systems layer-2 networks use — "were nowhere near mature enough to even consider" native rollups in 2023, he wrote, and blob and gas futures "just didn't exist back in 2023."

Life After The EVM

The Ethereum Virtual Machine is the software engine every computer on the network runs to execute smart contracts, and it has been Ethereum's execution standard since launch. The strawmap's frame design already assumes the protocol will expose an instruction set that is not the EVM, Buterin wrote, with two candidates leading: leanISA, a design coming out of Ethereum research, and RISC-V, an open standard used in ordinary computer chips. An instruction set architecture is the basic vocabulary of commands a processor understands, and Buterin described both as "more simple, modern and efficient than the EVM."

Then he pushed the logic further. "Once they're there, why not expose them to developers everywhere? (And then, why not turn the EVM into being an IR on top of that ISA, instead of an enshrined feature massively complicating the base protocol?)" he wrote, adding that "much of the deeper exploration here is too early even for the strawmap."

An IR, or intermediate representation, is a middle layer that compiles down to the machine below it. Demoting the EVM to an IR would end its status as a hard-coded part of Ethereum and turn it into one option among several, a shift with consequences for every contract and tool built on it. Buterin has floated the RISC-V idea before; putting it in the strawmap's orbit moves it closer to the formal process.

The state-side change is framed the same way. New state types take over from state expiry, and Buterin said they are "not just a replacement for expiry, they're a fundamentally different paradigm to how Ethereum does scaling." On the data structure, Verkle gives way to the unified binary tree specified in EIP-7864 — a simpler design that stores all of Ethereum's state in one tree — and then to a further construction he abbreviates as PBT without expanding it.

Restrictive By Design

Buterin named two themes running through the new material. The first is that state types and zkzk frames both give up flexibility to gain throughput, optimizing narrow, high-volume cases instead of everything at once. The theme, he wrote, is "that instead of trying to maximally scale ALL ethereum activity, we try to create specialized mechanisms that have more restrictive properties that make them more scaling-friendly, while supporting the heaviest loads incurred by users and applications today (eg. token transfers, swaps) and tomorrow (eg. privacy protocols)."

Betting On Recursive STARKs

The second theme is a dependency. STARKs, or scalable transparent arguments of knowledge, are proofs that let one computer verify another's computation without repeating it, and recursive STARKs are proofs about other proofs, so a single small proof can stand in for a large amount of work. They appear in several layers of the protocol, and one primitive — what Buterin called "aggregate to union verified dependencies" — is expected in three places at once: the layers that execute transactions, agree on blocks, and store data.

Proofs carrying that much of the protocol have to be correct, and Buterin's answer is formal verification, the practice of mathematically proving that code does exactly what its specification says. Doing that across an entire blockchain protocol has never been practical by hand.

"This can only be safe with formal verification, which is itself only feasible with modern AI tools," he wrote, describing STARKs and AI-accelerated formal verification as objects "that we are okay betting the technical future of Ethereum on."

The strawmap's own FAQ makes a similar point about schedule, noting that the current draft "assumes human-first development, and AI-accelerated R&D could significantly compress schedules." That schedule has drawn public pushback. "But 3-4 years is very slow. I think we should be ambitious and get it done in ~1 year. I think this is realistically possible now with LLMs," or large language models, wrote Dankrad Feist, a researcher at the Stripe-backed blockchain Tempo and formerly a longtime Ethereum Foundation researcher. His July 4 post also called the strawmap's features "REALLY COOL."

Issuance, Flagged With Doubt

Nothing in the five north stars addresses how much new ETH the network pays out each year to validators, the operators who lock up ETH to help secure it. Monday's post does not mention the question either. The top reply went straight at that.

"and token economics revised ? EIP 8363 is critical first step," wrote Vanessa Grellet, managing partner at Arche Capital and a board member of the Enterprise Ethereum Alliance.

She was pointing at the tapered issuance burn, a proposal that would destroy a rising share of validator rewards to remove the incentive to stake beyond half of all ETH in existence. Core developers put it up for consideration in Hegotá, the upgrade after Glamsterdam, on Aug. 6 — the weakest stage in Ethereum's upgrade process, which commits no one to shipping it. That came after two days of opposition from Aave founder Stani Kulechov, ether.fi chief executive Mike Silagadze and other founders in decentralized finance.

The strawmap diagram itself, last revised Aug. 4, does carry a "tapered issuance burn" box. It sits in the consensus layer under a fork the strawmap pencils in for 2027 to 2028, hyperlinked to the EIP-8363 pull request and marked with the symbol the diagram's legend defines as "uncertainty about ever shipping upgrade." Justin Drake, who co-maintains the strawmap, is one of the proposal's six listed authors. Buterin has not posted on it, and none of the 181 posts in the Ethereum Magicians thread debating it are his.

Fee Base Keeps Shrinking

The throughput targets sit against a shrinking fee base. Ethereum's base layer collected $8.3 million in transaction fees in July, down from $39.6 million in August 2025, a 79% decline over 11 months, according to DefiLlama. Fees over the first 10 days of August total $2.1 million. Ethereum destroys the base portion of every transaction fee rather than paying it out, under a mechanism it adopted in 2021, so a smaller fee base means less ETH removed from circulation.

ETH trades at $1,873, down 2.6% on the day and down 55.7% over a year, with a market capitalization of $226 billion, CoinGecko data shows. Its record high of $4,946 was set on Aug. 24, 2025.

Buterin closed the post with a set of declarations.

"Ethereum will be quantum-safe. Ethereum will put users' privacy first. Ethereum will be secure. Ethereum will be censorship-resistant. Ethereum will be highly performant and scalable while satisfying the above. And Ethereum will be Lean."