The Kakarot zkEVM project has been started in October 2022.
In December 2022, it reached 100% opcodes equivalences, and in January 2023 we completed 8/9 precompiles support.
The EVM itself is written in Cairo and currently deployed as a smart contract on Starknet (like Aurora on NEAR).
We are now working on the RPC nodes to enable the usual EVM tooling and wallets to work out of the box with Kakarot. If everything goes well, we should be able to open a testnet in the coming months.
other links since the original post can only contain 2
and latest deployment on Starknet testnet: Contract - Starkscan
is Kakarot technically an “L3”?
it’s a good question, and actually the response is not obvious and may depend on Starknet traction and performance as well.
I think that the fact that we currently live on top of Starknet is an implementation detail that the user may not be aware of (like how many Aurora users are aware of it being only smart contracts on Near). It does not necessarily mean that kakarot is an L3 because there is still a “direct” L1 to kakarot gateway (using under the hood the Starknet one). If I’m right, the level (2, 3) depends technically on where you commit your chain data. If kakarot is pure starknet contracts, then kakarot, as a part of Starknet, settles on L1.
Depending on the performance (throughput, finality, cost) of being on Starknet, and business opportunities, we may develop our own infra and commit either to starknet (which would make it an L3) or to L1 (which would keep it an L2).
Overall, in the current context of both Kakarot and zkEVM L2s, I think that Kakarot stands amongst other competitors as an alternative “EVM L2”, ie a place where you can deploy your solidity stack as usual, and which eventually settles on L1.