A Brief History of zkVMs

zkVMs have become a very important part of both Ethereum and its Layer 2s. However, there are many misconceptions and confusing bits of information around them.

This article aims to serve as a gentle introduction for users and researchers interested in understanding their importance, as well as for developers looking to get started with zkVMs so they can select the best option for them in case they want to build a project or contribute to an existing one.

From Low-Level Libraries to zkVMs

This section explores the evolution of ZK projects, from their origins in low-level libraries to the rise of zkDSLs and modern zkVM.

, which was used to create and

Main ZK projects falling into every catagory.

The second wave of end user applications were developed using zkDSLs, such as Tornado Cash for private transactions, and MACI for voting. Later, ZK made its way to L2s, making use of it's succinctness properties to reduce transaction costs. L2s often required synthetic virtual machines built on zk libraries and zkDSLs. For instance:

• Polygon zkEVM was built by combining different libraries where it made more sense, zkSTARKs and Plonky2 for producing quick execution proofs and Circom for cheap verification. This resulted on the zkASM + PIL synthetic virtual machine.


• Scroll was born by forking the Ethereum Foundation’s Privacy and Scaling Exploration (PSE)’s zkEVM, using the Halo2 zk library that aimed to be able to prove the Ethereum L1 directly.


• Starkware introduced its own Cairo-VM that differs from the EVM but it's still able to be verified by Ethereum L1.

, a fork of the OP Stack that transforms Optimism into a ZK Rollup. Meanwhile, Risc Zero introduced plans to implement OpenVM on mainnet, making use of its modularity to create an OpenVM module of Ceno, Scroll’s new zkVM initiative.

How zkVMs Generate Value

You might wonder about the revenue model, given that zkVMs are open-source and free. The answer lies in the prover marketplace. As demand for proving grows, reliable and responsive provers will become essential. A decentralized prover marketplace, run by a security and governance token, could be the solution. Risc Zero currently has a closed beta of their solution is a work in progress.

Will zkDSLs Die?

I don't think so. But this question is under debate at this point in time. A big part of the ZK community argue that zkDSLs like Circom or Noir will remain fundamental to Ethereum because they can cover a wider variety of cases and also because, with the right amount of time and resources, they can become more efficient that zkVMs. And by more efficient I mean cheaper to verify and faster to prove. However zkDSLs appear less suited for L2s, with zkVMs taking precedence. Privacy-focused applications will likely continue relying on zkDSLs.

Interestingly, zkVMs may enable new forms of privacy through client-side proving or light clients. For instance, zkVMs could facilitate applications where users run light nodes to verify states locally using zk proofs. This idea, while somewhat related, is worth exploring in a future article.

Is There Competition in the zkVM Space?

Absolutely. And yes, there’s

• Announced: )


• Developer: on X

Risc Zero has been a pioneer in zkVMs and is currently the most advanced project in the field. This has earned it the trust of the community through its significant contributions to the zkVM space.

Get Started with Risc Zero

Run the following a for detailed guidance.

SP1

git clone https://github.com/Turupawn/zkBenchmark.git

cd zkBenchmark

sh generate_proof_sp1.sh