Kakarot: A Zk-EVM that Vitalik Buterin thought was cool and invested in

Written by: Yinan

Introduction

Kakarot zkEVM is an Ethereum Virtual Machine (EVM) written in Cairo. Cairo is a Turing-complete language associated with CairoVM. CairoVM utilizes polynomial and ZK-STARK proof systems to achieve provable computation.

The characteristic of zkEVM is that it can generate provable transactions, so it can also generate provable blocks. Kakarot is built on top of CairoVM, so every transaction executed on Kakarot is provable.

In general, Kakarot zkEVM is written in Cairo language and utilizes the ZK-STARK proof system to provide provable computing power. It enables developers to deploy EVM applications on top of it, and enables end users to interact with these applications using their usual toolchain.

Finally, Kakarot will also provide interoperability with the Starknet protocol, providing composability for protocols in different fields such as DeFi and GameFi.


route map

The project started in October 2022 with several goals: open source from day one and, most ambitiously, create a zkEVM in less than 10,000 lines of code. This was achieved in just six months. And the testnet will be released in late summer of 2023.

The roadmap of Kakarot zkEVM is divided into three phases:

• Phase 1: Kakarot zkEVM on Starknet — Bringing EVM to Starknet

Kakarot zkEVM will first exist in Starknet's L2 as a built-in EVM, enabling developers to directly deploy their Solidity (or any EVM compatible language) smart contracts on Starknet and use their familiar tools (such as Foundry, Hardhat , Wagmi, etc.). Their end users will be able to interact with their dApps using their familiar toolchains such as Metamask, Wallet connect, etc. Overall, the developer and user experience on Kakarot will be exactly the same as Polygon, Scroll or Ethereum L1

• Phase 2: Kakarot x Madara — L3 zkEVMs

Kakarot and Madara will be combined into a unified stack to enable L3 zkEVMs and where it makes sense to enable L4, L5, etc. Teams will be able to deploy their zkEVM application chain and utilize proof of validity to settle transactions on Starknet. Computations are performed off-chain or at a higher layer, and verified on-chain. These proof-of-validity L3s (like Kakarot) have an interesting and underrated property: the ability to decouple security and decentralization. Users can benefit from the security of Ethereum L1 without requiring the same level of decentralization (i.e. thousands of validators). However, for rollups, decentralization is beneficial because it brings liveness and censorship resistance, two very important properties. This can be achieved by setting the sequencer in the hundreds instead of thousands. L2 has a much lower gas cost and higher performance (TPS) than L2 due to computation being done at another layer. Note that L2 is already much cheaper than L1. Rollups are stacked and multiplied for scalability. To further reduce gas costs, proof verification and data availability (DA) can be separated. Starknet L2 can be used only as a proof verification layer, while new data availability solutions, such as Celestia or EigenDA, can be used to publish transaction data. Users will have the option to opt-in to either, depending on their security needs. Publishing proofs and transaction data on Starknet is a safer option, while using DA solutions to publish transaction data is a more economical option

• Phase 3: Kakarot x Madara — type 1 zkEVM

Kakarot and Madara can also be combined to implement type 1 zkEVM. If you can:

1. Write the Ethereum consensus rules with Cairo in the Madara x Kakarot full node, so that the L1 consensus can be proved.
2. Switch from Pedersen Merkle Patricia Trie (MPT) to Keccak MPT. Then, Kakarot will become a type 1 zkEVM client capable of attesting to L1 blocks. This is a more advanced use case that depends on the roadmap of Ethereum, especially Verge. After Verge, keccak may be replaced by poseidon as the hash function of choice for Ethereum. This will help the zkEVM team to be type 1, since the main compatibility hurdle for zkEVM is storage layout, i.e. implementing Keccak MPT at a provable and reasonable cost.

Team:

In a hacker's house in Lisbon, about a week of dedicated work, about ten people built a zkEVM, so far, Kakarot core team was born.

With the help of the Starknet Foundation and Onlydust, today Kakarot has more than 40 unique contributors.

**Elias Tazartes: **Kakarot's co-founder and CEO, a freelance developer specializing in blockchain engineering (Dapp, Ethereum, Algorand) and full-stack development (Svelte/React, NodeJS, Rust, Python).

**Clément Walter:**Scientist, engineer, entrepreneur, handyman, curious about everything, this is Clément's introduction to herself. At the same time, he is the co-founder and chief technology officer of Kakarot


Financing:

Currently Kakarot has carried out pre-seed round financing. The specific financing amount has not been announced. The teams and personnel involved in the investment are: **Starkware, LambDa. **Angel investors include Vitalik, Nicolas Bacca, co-founder of Ledger, and Rand Hindi, co-founder and CEO of ZAMA.


Summarize:

At present, the ecological development of public chains that are not compatible with EVM is in a slow stage. At present, most of the public chain ecology is developed around the Ethereum ecosystem. Kakarot, as the Ethereum Virtual Machine (EVM) written in Cairo on Starknet, can quickly A good link between the Starknet ecology and the Ethereum ecology, on the Starknet that is not compatible with EVM, enables the Ethereum ecology to be well built on Starknet, and lowers the entry barrier for developers to develop, accessing the security of Ethereum and reducing costs. In the second and third stages of Kakarot, more composability can be achieved, and modules such as Celestia or EigenDA can be combined modularly.

As the ZK-EVM written on Starknet, the connection with the Starknet ecology is like a relationship between lips and teeth, and both are prosperous. At present, Starknet is also constantly iteratively upgrading to improve the user experience, and relying on the ZK-EVM built by Kakarot, it can be very good. To contribute to the ecological prosperity of the project, on the contrary, if the development of Starknet is not satisfactory, Kakarot will also be affected.

Another advantage of Kakarot is that it has the support of V God and StarkWare, not only spiritually but also financially.


Compared with other ZK-EVMs, Kakarot is at a disadvantage—a team of less than 10 people, with a much smaller budget and brand. This is one reason why Kakarot is not deployed as a general-purpose zkEVM L2.

At present, Kakarot is still in the early stage, and the testnet will be released in the near future, please stay tuned.