TLDR
- Vitalik Buterin says AI “vibe coding” helped prototype Ethereum’s full 2030 roadmap in just a few weeks
- He warned the AI-generated code likely has critical bugs and stub implementations
- Buterin proposed taking half the AI speed gains and redirecting them into security improvements
- He laid out a plan to replace Ethereum’s state tree and eventually swap out the EVM for RISC-V
- Ethereum’s next two upgrades, Glamsterdam and Hegota, are both planned for 2026
Vitalik Buterin Says AI Could Speed Up Ethereum’s (ETH) Roadmap, But Warns of Security Risks
Ethereum co-founder Vitalik Buterin said artificial intelligence is accelerating development on the Ethereum network faster than most people expected.
A developer took up a February bet with Buterin and used AI to prototype Ethereum’s full roadmap through 2030 in just a few weeks. Buterin called it “quite an impressive experiment” on X over the weekend.
This is quite an impressive experiment. Vibe-coding the entire 2030 roadmap within weeks.
Obviously such a thing built in two weeks without even having the EIPs has massive caveats: almost certainly lots of critical bugs, and probably in some cases "stub" versions of a thing… https://t.co/ZlTg0r2hvI
— vitalik.eth (@VitalikButerin) February 28, 2026
Buterin said AI is “massively accelerating coding” and that people “should be open to the possibility that the Ethereum roadmap will finish much faster than people expect.”
He also said it could be completed “at a much higher standard of security than people expect.”
However, Buterin was clear that the AI-generated code almost certainly contains critical bugs. He said some parts may be “stub” versions where the AI did not even attempt a full implementation.
“But six months ago, even this was far outside the realm of possibility,” Buterin added.
He recommended developers take only half the time savings from AI and use the other half on security. That means generating more test cases, formally verifying code, and building multiple implementations of each component.
Buterin said he is personally excited about the idea that bug-free code, “long considered an idealistic delusion,” could become a real expectation.
Ethereum’s State Tree and EVM Overhaul
On Sunday, Buterin also published a detailed breakdown of two major architectural changes he sees as critical to Ethereum’s future.
The first is switching from the current hexary Keccak Merkle Patricia Tree to a binary state tree under EIP-7864. This proposal has been in draft since January 2025.
The binary tree would produce Merkle branches four times shorter than the current structure. Changing the hash function could also improve proving efficiency by 3x to 100x.
Verkle Trees were previously considered for a 2026 hard fork, but quantum computing concerns helped shift interest back toward binary trees around mid-2024.
The second change involves replacing the EVM with RISC-V, the open-source instruction set most ZK provers already use. Buterin first proposed this in April 2025.
Pushback and Next Steps
Researchers from Offchain Labs, the team behind Arbitrum, published a rebuttal in November 2025 arguing that WebAssembly is a better long-term choice than RISC-V for Ethereum’s smart contract format.
Buterin said these two changes together account for over 80% of Ethereum’s proving bottleneck, making both “basically mandatory.”
Ethereum’s Glamsterdam upgrade is planned for the first half of 2026, with Hegota to follow later that year. Developers have not yet finalized the headline EIP for either fork.
