Vitalik Buterin said he has updated Ethereum’s 2023 roadmap and overlaid the original items onto the current Strawmap. He wrote that the overlap is still high overall, though the ordering of several items has changed.
Among the changes he listed, quantum security has moved up in priority. VDF and many proposed EVM improvements have moved down. Some earlier items have also been replaced by what he described as better constructions, including a shift from Verkle to unified BT to PBT, and from state expiry to new state types.
New additions in the Strawmap
Buterin said the most striking change is the appearance of several items that were not present in the 2023 roadmap at all. In his framing, those additions reflect a real change in Ethereum’s priorities.
He singled out a top-level focus on strong privacy. That set of work includes keyed nonces and recent roots, multiple aspects of FOCIL, lean privacy pool, and wormholes.
On aggressive scaling in a post-quantum setting, he listed leanSPHINCS signatures and aggregation, along with zkzk frames. He also referenced the EIP-8288 discussion tied to zkzk frames.
Other additions include simplification of specifications to help formal verification, blob and gas futures, and native rollups. Buterin said those ideas either did not exist in 2023 or were not mature enough to be considered at the time. On native rollups, he noted that SNARKs were still far from mature enough in 2023 for that discussion.
More room to rethink the future of the EVM
Buterin said zkzk frames already imply that the protocol will expose some non-EVM ISA, or instruction set architecture, to users. He named leanISA and RISC-V as the current leading candidates.
In his view, those ISAs are simpler, more modern and more efficient than the EVM. Once they exist, he asked why they should not be exposed to all developers. He then raised a follow-on question: whether the EVM should become an IR on top of that ISA, rather than remain a fixed feature that greatly complicates the base protocol. He added that this deeper line of exploration is still too early even for the Strawmap.
New state types as a different scaling model
Buterin wrote that new state types are not only a replacement for state expiry. He described them as a fundamentally different Ethereum scaling paradigm.
He said there is a shared theme in scaling work, and that it appears in both state types and zkzk frames, which he called new ideas. Rather than trying to maximize the scaling of all Ethereum activity, the aim is to build specialized mechanisms with stricter properties that are more scaling-friendly, while supporting the heaviest loads users and applications face today, such as token transfers and swaps, and the loads they may face tomorrow, such as privacy protocols.
STARKs and AI-assisted formal verification move to the center
Another shared theme, he said, is treating STARKs and AI-accelerated formal verification as first-class elements and being willing to bet Ethereum’s technical future on them.
According to Buterin, recursive STARKs will appear across many layers of the protocol. One specific primitive, which he described as an "aggregation to union verification dependency" primitive, is expected to be used in three places across the protocol: EL, CL and DL. He said that design is only safe with formal verification, and that complete formal verification is only feasible with the help of modern AI tools.
Buterin’s closing view
Buterin said Ethereum has taken many steps forward in maturity, backed by what he described as the enormous effort of dozens of Ethereum researchers and developers working across these features.
He wrote that Ethereum will be quantum-secure, will put user privacy first, will be secure, will be censorship-resistant, and will deliver high performance and scalability while meeting those requirements. He also said Ethereum will be lean.