Ethereum co-founder Vitalik Buterin has officially proposed the network's most significant execution layer overhaul since The Merge, outlining a mandatory transition to Binary State Trees (EIP-7864) and a long-term retirement of the Ethereum Virtual Machine (EVM) in favor of the RISC-V architecture. In a detailed roadmap released on March 1, 2026, Buterin argued that these structural reforms are no longer optional, identifying the current state data structure and virtual machine as responsible for "over 80% of the network's technical bottlenecks." This bold strategic pivot aims to future-proof Ethereum against quantum threats while exponentially increasing zero-knowledge (ZK) proof efficiency.

The End of Merkle Patricia Trees: Introducing EIP-7864

At the core of the new roadmap is EIP-7864, a proposal to replace Ethereum's decade-old Hexary Merkle Patricia Tree (MPT) with a more streamlined Binary State Tree. For years, the Hexary MPT has served as the backbone of Ethereum's data storage, but its complexity has become a hindrance to scalability.

According to Buterin, switching to a binary structure will shorten Merkle branches by a factor of four. This reduction is critical for light clients like Helios, allowing them to verify the chain's state with significantly less bandwidth. "These are 'deep' changes that many forgo, considering the incremental approach more 'pragmatic'," Buterin wrote, emphasizing that the binary shift is practically mandatory for client scenarios that rely on efficient proofs.

The upgrade also introduces "paged" storage slots, grouping adjacent data into efficient blocks. For DeFi developers, this is a game-changer: applications that access contiguous storage could see gas cost reductions of over 10,000 gas per transaction. Additionally, the proposal suggests replacing the Keccak256 hash function with faster alternatives like Blake3 or Poseidon, which could boost proof generation speeds by 3x to 100x.

RISC-V vs. Ethereum Virtual Machine: A Paradigm Shift

Perhaps the most controversial aspect of the 2026 roadmap is the planned phase-out of the Ethereum Virtual Machine (EVM). While the EVM has been the industry standard for smart contracts, it was not originally designed for the era of Zero-Knowledge rollups (ZK-rollups).

Buterin proposes adopting RISC-V, an open-standard instruction set architecture (ISA) widely used in hardware and ZK provers. The transition is aimed at solving the "proving overhead" that currently limits Ethereum's L1 scalability. A RISC-V-based execution layer would allow for native client-side ZK proof generation, making the network faster and more verifiable on consumer hardware.

The Three-Stage Transition Plan

To avoid disrupting the billions of dollars in value locked in current smart contracts, the transition to RISC-V will occur in three calculated stages:

  • Stage 1: Precompiles. The network will introduce RISC-V engines solely for specific, complex precompiled contracts, testing the architecture's stability in a sandbox environment.
  • Stage 2: Parallel Deployment. Developers will be allowed to deploy new smart contracts directly in RISC-V bytecode alongside existing EVM contracts.
  • Stage 3: EVM as Middleware. In the final stage, the EVM itself will be deprecated as a native engine and re-implemented as a smart contract running on top of the RISC-V architecture. This ensures full backward compatibility while moving the base layer to the new standard.

Roadmap 2026: The Path to 'Glamsterdam' and 'Hegota'

These upgrades are slated to begin integration as early as the Glamsterdam hard fork, expected in the first half of 2026. While Glamsterdam may focus on the initial groundwork for EIP-7864, the full implementation of the RISC-V transition is a multi-year project likely to mature alongside the Hegota upgrade later this year.

The industry's reaction has been largely positive, though cautious. Developers acknowledge that while the learning curve for RISC-V might be steeper, the benefits for ZK-EVMs are undeniable. By aligning Ethereum with a standard hardware architecture, the network effectively outsources optimization to the broader tech industry, rather than relying solely on niche blockchain engineering.

Addressing the 80% Bottleneck

Vitalik's analysis is stark: the state tree and the VM are the "primary bottlenecks" holding back Ethereum's next leap in performance. By addressing these foundational layers now, Ethereum is positioning itself not just as a settlement layer, but as a hyper-efficient computation engine capable of handling the global financial throughput of the next decade.

As the community digests these proposals, the focus now turns to the core developer calls in the coming weeks, where the exact specifications of EIP-7864 will be finalized. For investors and developers alike, the message is clear: the future of Ethereum is binary, and it runs on RISC-V.