Vitalik Buterin published a consolidated execution layer reform proposal in late February 2026, calling for Ethereum to replace its current state storage structure with a more efficient binary tree design and, further down the road, to swap out the Ethereum Virtual Machine (EVM) itself in favor of RISC-V, an open-source processor instruction set already used by most zero-knowledge (ZK) proof systems. The proposals arrive alongside the Ethereum Foundation's February 2026 Protocol Priorities Update, which formally lists binary trees and statelessness as long-term execution layer targets.

---

Two Proposals, Two Timelines

The first part of the plan, formalized as EIP-7864, addresses how Ethereum stores its global state. Today, the network uses a hexary Merkle Patricia Tree, a branching data structure verified using the keccak hash function. Both the tree format and the hash function were not designed for ZK proof generation, and the combination now accounts for over 80 percent of current ZK proving bottlenecks, according to discussion on Ethereum Magicians.

The proposed binary tree reduces branch sizes by roughly 75 percent compared to the current structure. Replacing keccak with Blake3 could yield a roughly 3x speed improvement in hash computation; replacing it with Poseidon, a hash function designed specifically for ZK circuits, could deliver up to 100x improvement. To make that figure concrete, Poseidon achieves approximately 2 million hashes per second compared to roughly 15,000 for keccak, though Poseidon remains under security review. For applications that read storage frequently, such as remittance contracts or DeFi protocols, the change could translate to savings of more than 10,000 gas per transaction on initial storage slot reads. The binary tree proposal achieves this in part by grouping storage slots into pages of 64 to 256 slots, which lowers the cost of accessing multiple values within the same slot range and is the structural mechanism behind the gas savings figure.

Buterin noted in his writing on binary state trees that "shorter branches make client-side verification cheaper."

The second part, replacing the EVM with RISC-V, is explicitly speculative. Buterin first raised the concept on the Ethereum Magicians forum in April 2025. The core argument is that ZK-EVM provers already internally translate EVM bytecode into RISC-V instructions to generate proofs, adding an unnecessary translation step. Exposing RISC-V natively would remove that layer. The proposed rollout is designed to be backward-compatible across three stages: first replacing approximately 80 percent of existing precompiles (specialized built-in functions) with RISC-V code, then allowing developers to deploy RISC-V contracts directly, and finally wrapping the EVM itself as a smart contract inside the new virtual machine so that existing deployed code continues to function, except for some gas cost shifts that would accompany the transition. For developers concerned about having to learn new low-level languages, the expectation is that most would continue using Solidity or Vyper with RISC-V compiler backends rather than writing raw Rust or assembly.

Buterin addressed one of the most common technical objections directly. "Most of the usage of the EVM is not u256 based," he wrote on Ethereum Magicians, referring to the 256-bit integer arithmetic that the EVM handles natively and that RISC-V does not. "The great majority of values that the stack deals with are u32 or u64."

Not everyone is persuaded. Ben Adams, an Ethereum contributor, raised concerns about performance for operations that do require 256-bit arithmetic: "If that was decomposed to RISC-V 64-bit or worse 32-bit instructions, that then becomes an extremely hard problem to recognise the patterns and then recompose back to 256-bit operations." Adam Cochran, a crypto analyst and developer, questioned whether the work was the right priority at all: "Does L1 execution improvement justify the technical lift when L2 enablement offers broader benefits?" Levs57, an Ethereum Magicians contributor, challenged the foundational premise more directly, arguing that RISC-V zkVMs still require precompiles that limit the promised simplicity gains, which calls into question whether the transition actually reduces prover stack complexity in the way the proposal assumes.

---

Where Ethereum Stands Today

The proposals land against a backdrop of record network activity. As of early March 2026, approximately 37.1 million ETH is staked on the network, and Ethereum accounts for 57 percent of total value locked across all DeFi protocols, with more than $159 billion in stablecoins (USDC and USDT combined) settled on the chain. ETH was trading near $1,973 as of March 1, 2026. Monthly decentralized exchange volume on Ethereum reached $86 billion in the fourth quarter of 2025. That volume also illustrates why execution layer efficiency matters at this scale: proving costs and storage overhead compound across every transaction the network processes, and the gap between the current architecture and a ZK-optimized one grows more consequential as activity increases. The coordination challenge of changing that infrastructure at this size helps explain why the RISC-V proposal carries no scheduled implementation date.

---

Regional Implications

For developers outside North America and Europe, both proposals carry practical stakes. In Nigeria, which now accounts for roughly 3 percent of global blockchain developers and half of all Web3 contributors across Africa, gas cost reductions on storage-heavy contracts are directly relevant to remittance and microloan applications where transaction fees cut into already thin margins. Nigerian blockchain startups raised $20 million in 2024, with communities such as ETHNigeria, Web3Bridge, and Web3Ladies training developers primarily on EVM-compatible toolchains. Looking further ahead, the RISC-V transition may prove less disruptive for African developer communities than it initially appears: training pipelines across the continent already include Rust, which is common in adjacent Solana and Polkadot ecosystems and is the primary language for RISC-V smart contract development, positioning those developers ahead of the curve. The faster ZK proving speeds that both proposals are designed to enable also carry a specific benefit for African fintech: on-chain privacy for salary payments, microloans, and remittances is a growing demand in the region, and improving proof generation efficiency is a prerequisite for making that privacy practical at scale.

In South Asia, India's Web3 developer job market registered 1,719 open positions in 2025, but structural factors including a 30 percent tax on virtual digital asset gains and a 1 percent tax deducted at source on transactions push many developers toward building for international users. For that community, the binary tree changes require near-term attention: gas cost profile shifts will require dApp audits and redeployment planning regardless of whether teams agree with the underlying design direction. The proposals also carry a forward-looking angle that is particularly relevant to India's regulatory environment. Better client-side proving and privacy tooling, enabled by faster ZK proof generation, could help Indian developers build compliant but privacy-preserving financial applications, offering a path to serve domestic users within the existing tax and reporting framework rather than routing around it.

---

What Comes Next

The near-term Ethereum upgrade calendar focuses on other priorities. The Glamsterdam upgrade, targeted for the first half of 2026, addresses parallel execution and higher gas limits. The Hegota upgrade, expected in the second half of the year, includes enshrined proposer-builder separation via EIP-7732. The execution layer overhaul sits beyond both windows. EIP-7864 is the closer of the two components, with active specification work underway. The RISC-V transition, by Buterin's own framing, is a directional signal for where the protocol could go rather than a change with a delivery date attached to it. In Ethereum's development culture, directional signals from the protocol's founders tend to shape tooling investments, client roadmaps, and research priorities well before any formal upgrade process begins, which means developers building ZK-adjacent infrastructure have reason to track both proposals now rather than waiting for a specification to stabilize.