Taiko is a ZK-Rollup alternative to Ethereum, fully EVM compatible, incorporating all EVM operational codes at the Layer2 level, while maintaining the security level characteristic of Ethereum.
The project attracted $22 million in investments from major players like Seqouia China, IOSG Ventures, OKX Ventures, KuCoin Ventures, and others, received mentions from Vitalik Buterin, and support from leading funds and tier-one projects.
Taiko Protocol and Features
In the Taiko protocol, the initiator must find a confirmer for the block proposal, and the designated confirmer deposits TKO collateral to ensure delivery confirmation. Taiko does not impose rules on how initiators find and reward verifiers, allowing them to meet in person and complete the transaction in cash, for example. Thus, supply chain operates as a free market. Initiators can choose any confirmer they prefer.
Besides the economic advantage, several technical features make Taiko ideally suited for a multi-client system:
Taiko is a type-1 ZK-EVM, offering two benefits:
- For execution diversity, existing EVM implementations (Geth, Besu, Reth, etc.) can be directly integrated into L2.
- For verifying the anchored L1 design, a standardized ZK-EVM is required for open verification by multiple clients, as verifiers need to check the same transition to achieve consensus on their verification results. In this case, a type-1 ZK-EVM would be most suitable, as it closely follows Ethereum specifications. For rollup-specific logic, Vitalik also mentions how ZK-EVM can be modified with precompilation support, and this would be sufficient to use these precompiled for supporting Taiko's BBR (Based Booster Rollup) design.
Taiko publishes data on Ethereum, unlike some L2s exploring alternative data availability options. As long as the data is hosted on L1, Taiko can easily adapt to Vitalik's proposed implementation of ZKEVMClaimTransaction to cover the state transition, proof, and data availability.
Taiko works with multiple proof systems. Existing testnets already support ZK-EVM from PSE, SGX, and Reth. The infrastructure is set up for using multiple execution clients and proof systems, which will be discussed in the last section. Based on this infrastructure, position in the ZKP field will focus on modular compilation.
Modularity and Openness Development Plan
Multiconfirmation in Taiko is built on principles of modularity and openness.
Modularity
In the context of ZKP and considering multiple clients, the team uses modern compilers to create universal assembly instructions, such as Risc-V or WASM. These instructions are translated into arithmetic forms for various proof systems (AIR or PIL), and then the arithmetized execution trace is encoded using different SNARKs.
Client compilation with modern compilers brings the following benefits:
| Number | Description |
|---|---|
| 1 | Client updates are independent of proofs, as there is no need to implement schemes for the latest EIPs or hard forks; maintaining the source code up to date is sufficient. |
| 2 | The company achieves code optimization through compilation tools, such as LLVM. |
| 3 | Cross-compilation increases diversity; for instance, Geth or Reth are compiled into RISC-V or WASM instructions, already having four sets of proofs. |
The company's future efforts will focus on compiling SNARKs. Arithmeticization methods such as PLONK and R1CS, with backends like Halo2, eSTARK, or Supernova, are not limited to a single ZK protocol, unlike monolithic ZK-VM/EVMs, which are tied to specific ZKPs for backend implementation. As more projects begin to use components from each other to improve performance, the monolithic tech stack may become modular. The ZKP research field is evolving so rapidly that flexibility becomes more important than direct implementation of the latest results. To stay flexible, the company cooperates with projects like Powdr Labs and Risc Zero on their cross-compilation pipeline, achieving maximum modularity. For the technically savvy readers, here are the specific advantages:
- Developers can apply optimizations to compilers for different backends, for example, favoring high-degree gates or using more lookup arguments.
- Accelerated schemes, such as keccak and Poseidon hash functions, can be implemented as libraries.
- The ability to gradually add ZK features, such as LogUp, to the language and include corresponding backend support.
- Integration of new ZK frameworks for backends becomes faster. In some ZK research projects, only the concept is developed, making it difficult to use in production. By making the compiler the main tool, we can easily apply frameworks at early stages.
- Existing backend schemes, such as ZK-EVM components by PSE, written on Halo2, can be reused through direct invocation.
In collaboration, Taiko has already integrated zeth and ZK-VM from Risc Zero, developing an additional SGX backend for it. Taiko engineers also integrate Powdr into the multi-proof system, develop the language and PIL library, optimize compilation, add more backends, and perform overall low-level optimization. At the hardware level, our ZK Acceleration Layer (ZAL) aims to standardize collaboration between proof systems (Halo2, Arkworks, Risc Zero, Polygon, etc.) and acceleration libraries (CPU, GPU, FPGA, etc.).
Openness
The more clients, proof systems, and integrated backends — the better, so the team strives to bring everyone together into a community. The platform developers have a long history of working with others, for example, with PSE on ZK-EVM and Risc Zero. Now, by creating a more modular ZK stack, it's possible to effectively work on API abstraction for better generalizations and integration. Taiko will serve as a portal for deploying proof systems into production and their testing in-chain.
Taiko Stack
A scalable and flexible infrastructure is necessary for the multi-proof paradigm. The initial element of ZK proof of validity is the client state trace and storage proofs, which are used to create a witness and public data.
Hence, the platform uses a lightweight host to extract traces from multiple clients and transmit these traces to multiple provers. From the prover's side, the design supports both modular and monolithic stacks, while extracting the same public data from the target client (currently Geth).
In the future, Geth as a Taiko node can be replaced by another node if the trace format is compatible. Additionally, the lightweight client running on proof systems (currently Reth) can also be replaced with any implementation that compiles into an acceptable assembly language.
The Taiko protocol is ideally suited for the multi-client system, as it has an open supply chain of multi-proofs with type-1 execution, which ensures data availability on L1.
Taiko represents a multi-proof architecture with modularity and openness. The resource works with Powdr Labs to utilize cross-compilation with clients and ZKPs, and collaborates with Risc Zero for implementing Taiko execution on their ZK-VM and TEEs. Taiko's flexible infrastructure considers both modular and monolithic ZKP stacks.
