Ethereum co-founder Vitalik Buterin has proposed two major changes aimed at cutting costs and boosting efficiency for proof generation on the Ethereum network. The proposals target the system’s biggest bottlenecks, which account for over 80% of proving costs.
Near-term change: binary state tree (EIP-7864)
- Ethereum’s current hexary keccak Merkle Patricia Tree would be replaced with a unified binary tree using BLAKE3 or, in the future, Poseidon2.
- This reduces Merkle proof sizes by roughly 75% and shortens branch lengths 3–4x, cutting data verification costs and improving bandwidth efficiency.
- Page-based storage groups 64–256 adjacent slots, letting early-slot decentralized apps save more than 10,000 gas per transaction.
- Binary trees simplify auditing, create uniform depth across contracts, and pave the way for future state expiry mechanisms.
Long-term change: RISC-V virtual machine
- Buterin suggests replacing the Ethereum Virtual Machine with a RISC-V VM, arguing it aligns better with zero-knowledge proof systems, reduces the need for precompiles, and allows local client-side proofs.
- Deployment would happen in stages:
- New VM handles precompiles.
- Contracts can be deployed directly on the new VM.
- The EVM is reimplemented as a smart contract on the new VM, maintaining backward compatibility while adjusting gas costs.
- Advantages include raw execution efficiency, alignment with existing ZK provers, support for privacy and verification features, and simplicity in implementation.
Together, the binary tree and RISC-V VM target Ethereum’s proving-heavy workload. While the VM replacement is still under discussion in the developer community, the changes are designed sequentially: start with the binary tree, then upgrade the VM once proving infrastructure is ready.
In short, these upgrades aim to drastically reduce proof costs, improve efficiency, and future-proof Ethereum for a zero-knowledge-proof-heavy environment, rather than continuing to patch the old EVM architecture.
