zkEVM Update (PSE, Taiko, ConsenSys, Polygon, Scroll, zkSync, StarkNet)

We use Vitalik's zkEVM classifications to map the projects according to EVM equivalency. Type-1 is exactly Ethereum-equivalent, which means that the rollup can re-use 100% of EVM infrastructure. Type-2 and Type-3 fall in the middle of the spectrum, trading some minor changes for compatibility. Type-4 is the least EVM compatible, striving to compile some EVM code but sacrificing compatibility with most Ethereum apps.

Type 1

• The Ethereum Foundation’s Privacy and Scaling Explorations (PSE) team and Taiko chose to take on the hefty task of building a Type-1 zkEVM. Type-1 zkEVMs make no compromises in achieving equivalency. Even after completing the implementation and auditing process will probably take a minimum of several months to complete.
• Taiko spun out of Loopring to build an EVM-compatible ZK-Rollup.
• The Scroll and PSE team have closely collaborated to build their zkEVMs (literally sharing a repository), with the former team focusing on speed to market while sacrificing some elements of full compatibility. Because of the complex nature of a Type-1 zkEVM, there likely won’t be a mainnet anytime this year.

Type 2

• ConsenSys, Polygon zkEVM, and Scroll fall into the Type-2 bucket. As of now, Polygon and Scroll are in Type-3, slowly working towards further equivalency. Polygon is expected to launch mainnet in March, while Scroll is slotted for the end of Q2.
• ConsenSys is the newest zkEVM to enter the field. It launched a permissioned testnet in December 2022. The team hopes to address the rollup training wheel problem, where a multisig that can force a particular outcome in case there are bugs in the code, through a multi-prover system. With this system, a rollup would utilize several proving mechanisms with different levels of security to remove the risk of a single point of failure, which is the case with single-prover rollups.

Type 3

• Under the StarkNet umbrella, there is Kakarot. It is a zkEVM written in Cairo and has launched as a smart contract on StarkNet. Because it functions as an interpreter for EVM bytecode, it is best categorized as a Type-3 zkEVM. It is expected that Kakarot might eventually become a Layer-3 on StarkNet.

Type 4

• Type-4 zkEVMs are only language-level compatible, meaning high-level code, such as Solidity or Vyper, is compiled to be ZK-friendly.
• zkSync is on mainnet and in progress to launch to projects and users before the conclusion of Q1. In its existing instantiation, StarkNet qualifies as a Type-4 zkEVM using Warp, a compiler for Solidity to Cairo. So far, StarkNet is the only zkEVM that has decentralized its sequencer and prover.

Looking Ahead

Scroll leads the way in adoption but the increase in unique wallet addresses and transactions could result from speculation surrounding an upcoming token launch. The traction may not remain the same as projects go to mainnet.

Even at mainnet, the zkEVMs might continue to keep their training wheels as they make sure their network is debugged and they fully decentralize operators.