Ethereum’s Biggest Overhaul Yet? Vitalik Targets Core Architecture

• Vitalik says state tree and VM cause over 80% of proving overhead.
• The proposed binary tree could reduce proof sizes by a factor of 3 to 4.
• A long-term plan may replace EVM with RISC-V for simpler execution.

Ethereum could be heading for its most radical redesign since launch.

In a detailed technical proposal, Ethereum co-founder Vitalik Buterin laid out sweeping upgrades that aim to make the network dramatically faster to verify, cheaper to use, and far more efficient for zero-knowledge technology.

At the heart of his argument is efficiency. “They are the big bottlenecks that we have to address if we want efficient proving,” Buterin wrote, referring to Ethereum’s state tree and virtual machine. Together, he estimates, they account for more than 80% of the proving overhead.

A Simpler, Faster State Tree

Buterin proposes replacing Ethereum’s complex hexary state tree with a cleaner binary design. The impact could be huge.

Merkle proofs would become three to four times smaller. Client-side verification would be significantly cheaper. Storage would be grouped into “pages,” allowing applications to load and edit data more efficiently, potentially saving over 10,000 gas per transaction in some cases.

For developers building zero-knowledge applications, this change could remove a major bottleneck. Instead of creating separate data trees, projects could interact directly with Ethereum’s own state in a prover-friendly way.

The Hash Battle: Poseidon vs BLAKE

Buterin also entered the hash function debate. While he praised BLAKE as “beautiful and good,” he warned against losing the order-of-magnitude proving efficiency offered by Poseidon.

He outlined a future where block builders generate fast state proofs locally and outsource heavier computation elsewhere. In this setup, even AI data centers could be repurposed for Ethereum proving in under a minute.

A Future Without the EVM?

The boldest idea is long-term: replacing the Ethereum Virtual Machine with RISC-V. “If the EVM is not good enough to actually meet the needs of that generality, then we should tackle the problem head-on, and make a better VM,” Buterin wrote.

Many zero-knowledge provers already run on RISC-V. Aligning Ethereum’s execution layer with that architecture could simplify the protocol and boost efficiency dramatically.

The transition would be gradual, with full backward compatibility. This is not a cosmetic update. It is a structural reset aimed at making Ethereum leaner, faster, and built for a proof-heavy future.

Related: Buterin Unveils “quantum resistance roadmap” for Ethereum