Vitalik Pushes Ethereum Redesign Past Layer 2 Focus

• EIP-7864 would replace the hexary Merkle Patricia tree with a binary tree to cut proof size and costs.

• Buterin suggests RISC-V–style execution and vectorized precompiles to improve zero-knowledge efficiency.

• Critics warn deeper protocol changes may increase complexity and expand potential attack surfaces.

Vitalik Buterin is redirecting Ethereum’s scaling debate away from Layer 2 solutions and back to the protocol’s core. In recent remarks shared publicly, he argued Ethereum’s main long-term limits sit inside its state tree and virtual machine. He said these constraints matter now as zero-knowledge proofs become central to Ethereum’s roadmap.

State Tree Changes Take Priority

According to Buterin, the state tree and execution layer account for over 80% of current proving costs. As a result, he highlighted EIP-7864, which proposes replacing Ethereum’s hexary Merkle Patricia tree. The proposal would introduce a binary tree structure designed to shorten Merkle proofs by roughly four times.

Notably, shorter proofs would reduce bandwidth needs for verification. This change would benefit lightweight clients and privacy-focused applications. In addition, the binary tree would group storage slots into pages, improving efficiency when applications access related data.

Many decentralized applications frequently load adjacent storage slots. Because of this pattern, Buterin said some transactions could save more than 10,000 gas. He also suggested pairing the new tree with more efficient hash functions to further speed proof generation. Overall, the goal is to make Ethereum’s base layer more compatible with ZK systems.

Rethinking the Ethereum Virtual Machine

Beyond state storage, Buterin also outlined a longer-term rethink of Ethereum’s execution engine. He raised the possibility of moving beyond the EVM toward a RISC-V–based architecture. RISC-V already powers many modern proving systems, which could simplify integration.

He argued that growing reliance on precompiles signals discomfort with the EVM’s limits. As an interim step, he proposed a vectorized math precompile, often described as a “GPU for the EVM.” This could accelerate cryptographic operations significantly.

Debate Over Added Complexity

However, the proposal has drawn criticism. Analyst DBCrypto warned that repeated deep changes could increase abstraction and risk. He said each added layer may expand attack surfaces and trust assumptions.

Still, according to Buterin, Ethereum must evolve as zero-knowledge proofs shift from optional tools to core infrastructure. He stressed that future scaling gains may come from foundational changes, not additional layers.