Vitalik Buterin argues that incremental reforms cannot resolve Ethereum’s proving bottleneck, and outlined a two-part architectural transformation that goes deeper than any upgrade since The Merge.
Ethereum co-founder Vitalik Buterin on Sunday published a substantive technical argument that the network’s path to viable client-side proving runs through two deep architectural changes—a binary state tree replacing the existing hexary Merkle Patricia structure, and a long-term migration from the EVM to the RISC-V instruction set architecture. According to Buterin, both solutions are effectively non-optional if Ethereum is to fulfil its ambition as a general-purpose, verifiable computation platform.
The post arrives as Ethereum’s developer community is navigating an increasingly complex upgrade pipeline, with the Glamsterdam hard fork targeted for the first half of 2026 and the Hegota upgrade expected to follow later in the year.
The Proving Problem
To understand the urgency behind Buterin’s argument, it is necessary to situate it within Ethereum’s broader ZK roadmap. The network’s long-term scaling and trust-minimisation strategy depends on the ability to generate cryptographic proofs of execution cheaply enough that ordinary clients—not just specialized hardware operators—can verify the chain independently. This is the promise of “client-side proving.”
According to Buterin, the current state tree and virtual machine together account for more than 80% of the proving overhead. No amount of optimization at the margins can overcome two components that are architecturally mismatched with the proof systems they must feed. The choice, he argues, is between costly incrementalism that yields diminishing returns, or deep structural reform that is painful in the short term but resolves the bottleneck durably.
“They are ‘deep’ changes that many shrink away from, thinking that it is more ‘pragmatic’ to be incrementalist,” wrote Buterin.
Now, execution layer changes. I’ve already talked about account abstraction, multidimensional gas, BALs, and ZK-EVMs.
I’ve also talked here about a short-term EVM upgrade that I think will be super-valuable: a vectorized math precompile (basically, do 32-bit or potentially…
— vitalik.eth (@VitalikButerin) March 1, 2026
The State Tree Proposal: EIP-7864
The first pillar of Buterin’s framework centers on EIP-7864, a proposal that would replace Ethereum’s hexary Keccak Merkle Patricia Tree with a binary tree using a more ZK-friendly hash function. EIP-7864 has been n draft since January 2025 and credited for the most part to developer Guillaume Ballet
The efficiency gains are layered. Binary Merkle branches are inherently four times shorter than their hexary equivalents, meaningfully reducing the data bandwidth required for light client operations. This includes tools like Helios, which today must transmit and verify significantly larger proof structures.
Compounding this, a hash function substitution to either Blake3 or a Poseidon variant could deliver proving efficiency improvements ranging from 3x to 100x, depending on implementation and security parameters. Buterin noted that Poseidon, while offering the higher end of that range, requires additional cryptographic security review before it can be responsibly deployed at protocol level.
The proposal also represents an evolution of Ethereum’s own architectural planning. Verkle Trees—a different cryptographic structure—were previously the leading candidate for inclusion in a 2026 hard fork. Concerns around quantum computing vulnerability in the elliptic curve cryptography underpinning Verkle Trees helped reignite interest in binary alternatives from around mid-2024, a pivot that EIP-7864 now formalises as a concrete proposal.
The VM Question: EVM to RISC-V
The second and more architecturally ambitious proposal revisits Buterin’s April 2025 argument that Ethereum should ultimately replace the EVM with RISC-V—the open-source instruction set architecture that most major ZK proof systems already use as their internal proving substrate.
The migration path Buterin outlined is deliberately staged to manage disruption. In the first phase, RISC-V would handle only precompiles—the specialized, high-demand operations that already receive performance exceptions within the EVM. The second phase would open the door to user-deployed RISC-V contracts, operating in parallel with existing EVM contracts. The final stage—politically the most significant—would involve retiring the EVM itself by converting it into a RISC-V smart contract that emulates EVM semantics, preserving backward compatibility without perpetuating the underlying inefficiency.
The three-stage architecture reflects a lesson from Ethereum’s own history: the network has executed major in-flight changes—most notably The Merge, which transitioned consensus from proof-of-work to proof-of-stake without a chain break—and Buterin has explicitly invoked that precedent to argue that the developer community should not treat architectural conservatism as a proxy for maturity.
“This is still more speculative and non-consensus. Ethereum would certainly be *fine* if all we do is EVM + GPU. But a better VM can make Ethereum beautiful and great,” argues Buterin.
Dissent and Alternatives
The RISC-V proposal has indeed not achieved consensus. In November 2025, researchers from Offchain Labs—the core development team behind the Arbitrum Layer 2 network—published a structured counterargument contending that WebAssembly (WASM) represents a superior long-term choice for Ethereum’s smart contract execution format.
The Offchain Labs position rests on a conceptual distinction between the “delivery ISA”—the instruction set in which contracts are authored and deployed—and the “proving ISA”—the format into which execution is translated for ZK proof generation. Their argument is that these need not be identical, and that WASM offers developer tooling, ecosystem maturity, and security audit coverage that RISC-V currently lacks in a smart contract context.
The debate is not merely technical as it touches on Ethereum’s relationship with its existing developer base, the backward compatibility obligations of a $300 billion+ settlement layer, and the governance question of who has legitimate authority to propose foundational architectural changes. Buterin’s outsized influence in Ethereum’s informal governance structure means his framing carries weight that extends beyond the technical merits—a dynamic that periodically generates friction with the broader development community.
Implications for the Upgrade Roadmap
For investors and developers tracking Ethereum’s competitive position, the significance of these proposals is both technical and strategic. The network’s medium-term differentiation thesis rests on its ability to serve as a credibly neutral, high-integrity execution environment for a ZK-centric future—one in which client-side verification and proof aggregation replace the current reliance on social trust in block producers and infrastructure operators.
If the state tree and VM changes are successfully implemented, the resulting efficiency gains would disproportionately benefit applications that depend on cheap, frequent proof generation: rollup settlement, identity verification, private computation, and cross-chain bridging. The competitive moat of early adoption in these categories is non-trivial; protocols and infrastructure teams that build against a more proving-efficient Ethereum base layer will not need to unwind and rebuild when the architecture matures.
The near-term risk is execution. Deep architectural changes to a live network with tens of billions of dollars in locked value carry non-trivial technical and coordination risk. Client diversity—Ethereum’s deliberate strategy of maintaining multiple independent implementations—complicates the synchronization of major state tree changes across teams operating on different codebases. The Glamsterdam and Hegota upgrade timelines will be stress-tested by the ambition of what is now on the table.
What Buterin’s Sunday post signals, above all, is that Ethereum’s core contributors are prepared to accept the discomfort of deep structural change rather than defer it indefinitely. Whether that posture translates into cohesive execution or fragments into the kind of governance friction that has delayed previous upgrades remains the central unanswered question for the network heading into 2026.
Vitalik Buterin is set to speak at Hong Kong Web3 Festival 2026, the 4th edition of one of Asia’s largest and most influential crypto events, taking place from April 20 to 23 at the Hong Kong Convention and Exhibition Center (HKCEC).