Ethereum heart surgery? Vitalik proposes that the Ethereum execution layer may completely replace the EVM with RISC-V.

In the future blueprint of Ethereum, a new proposal initiated by Ethereum co-founder Vitalik Buterin has sparked heated discussions in the community: replacing EVM (Ethereum Virtual Machine) with RISC-V as the virtual machine language for smart contracts. This concept has been likened to a "beam chain level upgrade" of the execution layer, not only for scalability but also to address the fundamental bottlenecks of complexity and efficiency in the current execution layer.

What is RISC-V? Why should it replace EVM?

The core of the proposal is to replace the EVM currently used by Ethereum smart contracts with an open-source, modular instruction set architecture — RISC-V. This transition will not overturn the existing developer tools and habits of Ethereum developers because:

The existing account system, cross-contract calls, storage methods, and other core abstraction layers remain unchanged.

The original Solidity and Vyper languages can be compiled with RISC-V as the backend, and the developer experience will not change significantly.

The old EVM contracts can still communicate bidirectionally with the new RISC-V contracts.

In this way, developers do not need to relearn everything, but the performance and simplicity of the Ethereum underlying layer is expected to be greatly improved.

ZK-EVM is the biggest performance bottleneck.

As multiple scalability proposals (such as EIP-4444, delayed execution, and stateless clients) are gradually launched in the future, the factors that truly limit the scalability of Ethereum L1 will focus on:

Data availability sampling and the stability of historical storage protocols

Market competition in block production

ZK-EVM proof efficiency

Currently, in the ZK-EVM proof process of a block, executing the EVM Virtual Machine logic occupies about 50% of the resources. This means that if smart contracts can run directly in a RISC-V environment, there is a chance to achieve a 50-fold or even 100-fold improvement in ZK proof performance.

Interestingly, the proof process of the current ZK-EVM actually compiles the EVM into RISC-V, which is then proven by the ZK system. Therefore, directly making RISC-V the native virtual machine of the Ethereum execution layer is not only logical but also saves the resource consumption of intermediate conversions.

Why is RISC-V fast? Comprehensive optimization from hash functions to architectural design.

The four main resource-consuming items of ZK-EVM currently are:

deserialize_inputs

initialize_witness_db

state_root_computation

block_execution

The first three can be significantly optimized by using more friendly hash functions (such as Poseidon) and binary state trees. For example, Poseidon can handle 2 million hashes per second on a laptop, far surpassing Keccak's 15,000. If these optimizations are implemented, they will greatly reduce the burden by the first 50%.

But the remaining 50% still comes from

block_execution

This part can only be fundamentally solved through more efficient VM designs, such as RISC-V.

Three implementation methods are available, ranging from conservative to aggressive.

Vitalik proposed three technical implementation paths:

– Option 1: Dual Virtual Machines coexist (minimum risk): Allows contracts to choose between using EVM or RISC-V, both of which are interoperable and share resources, balancing compatibility and innovation.

– Option 2: RISC-V packaged EVM interpreter (radical upgrade): All EVM contracts will be executed through the RISC-V built-in EVM interpreter, transitioning the overall execution layer to a unified underlying architecture.

Option 3: Protocol layer supports Virtual Machine interpreter (moderate approach): The protocol is designed with a "Virtual Machine module", which defaults to using RISC-V to implement an EVM interpreter, and allows future expansion to other languages, such as Move.

The common advantages of these paths are: they can simplify the execution layer specifications, enhance maintainability, and improve verification transparency.

Sui development company Mysten Labs co-founder: If he could do it again, he would choose Move, without considering multiple languages.

In response to this proposal, Sam Blackshear, co-founder of Sui development company Mysten Labs, also shared his views. He stated, "I think adopting a RISC-V backend is a good choice for Ethereum (because it needs to support existing EVM contracts). But if I were to design a new chain from scratch, I would still choose Move over multi-language support. Many of Sui's advantages come from using strongly typed objects as a common abstraction layer throughout the stack."

This reflects the historical factors of different chains for the "virtual machine selection strategy", Ethereum's development was first designed in the early stage, and many future needs and developments could not be expected, and compatibility and transition design are currently being emphasized in response to changes; The new public chain Sui focuses on full-stack integration from the language to the bottom layer, so that development and security are tightly integrated.

Typus Finance's growth long Kyrie also shared a conversation he had with Vitalik during the EthTaipei event. He recalled, "At that time, I asked Vitalik: 'Do you think the Move language and object-oriented settings can enhance blockchain security?'"

He replied, "I don't think this changes anything; a project being hacked is still a hack, no matter what language it is in."

But Kyrie immediately rebutted that Move can indeed reduce the chances of development errors, is easier to learn than Rust, and the object-oriented model helps limit the scope of risk. "When contracts are hacked, the losses may be a limited amount rather than infinite exposure," he added.

Although Vitalik did not express his stance at the time, it seems that his willingness to propose RISC-V as a more strongly typed and modular alternative indicates a slight shift in attitude towards the importance of language design in blockchain security.

This article discusses Ethereum heart surgery? Vitalik proposed that the Ethereum execution layer may completely replace the EVM, switching to RISC-V, first appeared in ChainNews ABMedia.