ZK Co-Processor: Redefining a New Paradigm for Blockchain Applications

The Historical Origins of Co-processors

Co-processors have a long history in the computer field, primarily serving the role of offloading specific tasks from the CPU. For example, Apple's M7 motion co-processor launched in 2013 significantly enhanced the motion sensing capabilities of smart devices, while the well-known GPU was proposed by Nvidia in 2007 as a graphics processing co-processor. This architecture is commonly referred to as "heterogeneous computing" or "hybrid computing."

The main advantage of a coprocessor is that it can handle specific tasks that are complex and have high performance requirements, allowing the CPU to focus on more flexible and variable computing work.

Challenges Facing Ethereum

Ethereum currently faces two main issues:

1. High Gas fees limit the development of complex applications. A standard transfer requires 21,000 Gas, and more complex operations cost even more, which seriously hinders the widespread adoption of applications and users.

2. Smart contracts can only access a limited amount of recent Block data. Future network upgrades may further restrict the storage of historical data, making it difficult for many innovative applications that rely on large amounts of data to be realized.

These issues stem from the fact that Ethereum did not consider handling large-scale computing and data-intensive tasks from the beginning of its design. To accommodate these applications, the concept of co-processors needs to be introduced. The Ethereum main chain acts as the "CPU", while the co-processors are similar to "GPU", which can flexibly handle computing and data-intensive tasks. Combined with zero-knowledge proof technology, co-processors can perform trusted computing off-chain.

Application Prospects of Co-processors

The application fields of ZK co-processors are very broad, almost covering all real decentralized application scenarios, including but not limited to:

• Social Application
• Game
• DeFi Protocol
• On-chain data-based risk control system
• Oracle
• Data Storage
• Large Language Model Training and Inference

In theory, any Web2 application can be implemented on the blockchain through ZK co-processors, while also leveraging Ethereum as a secure settlement layer.

Overview of Mainstream Collaborative Processor Projects

Currently, well-known co-processor projects in the industry mainly focus on three major application scenarios:

1. On-chain data index
2. Oracle
3. ZKML ( Zero-Knowledge Machine Learning )

The general-purpose ZK coprocessor (General-ZKM) project can cover all three scenarios. The underlying virtual machine architectures used by different projects also vary, such as Delphinus focusing on zkWASM, while Risc Zero is based on the RISC-V architecture.

Technical Architecture of Typical Co-Processor Projects

Risc Zero

The ZK coprocessor of Risc Zero is called Bonsai, and its goal is to become a chain-agnostic general-purpose coprocessor. Bonsai is based on the RISC-V instruction set architecture and supports multiple programming languages such as Rust, C++, Solidity, Go, and more.

The main components of Bonsai include:

• Prover Network: Receives and generates ZK proofs
• Request Pool: Stores proof requests initiated by users
• Rollup Engine: Collect and submit proof results to the mainnet
• Image Hub: Visual Development Platform
• State Storage: Off-chain Key-Value Database
• Proof of Market: A platform that matches computing power supply and demand.

Lagrange

Lagrange aims to build a coprocessor and a verifiable database that contains historical data on the Blockchain to support the development of computation and data-intensive applications.

Its main functions include:

1. Verifiable Database: Indexing On-Chain Smart Contract States
2. Parallel Computing Based on MapReduce Principles

Lagrange adopts a new data structure to store contract states, account states, and Block data to meet the requirements of zero-knowledge proofs. Its ZKMR virtual machine achieves distributed computation and proof aggregation through two steps: Map and Reduce.

Succinct

The goal of Succinct Network is to integrate programmable facts into various aspects of the blockchain development stack. Its off-chain ZKVM is called Succinct Processor (SP), supporting LLVM languages such as Rust.

The core features of SP include:

1. Recursive Proofs Based on STARKs
2. Wrapper from SNARKs to STARKs
3. Precompiled Centered zkVM Architecture

Co-processor Project Comparison

When comparing general-purpose ZK co-processor projects, the following aspects are mainly considered:

1. Data Indexing/Synchronization Capability
2. Underlying technology route ( SNARKs vs STARKs )
3. Does it support recursive proofs?
4. Proof System Efficiency
5. Ecological Cooperation Situation
6. Financing Background

Overall, mainstream projects have relatively similar technical paths, all adopting wrappers from STARKs to SNARKs, supporting recursive proofs, and building their own prover networks. In a situation of severe homogenization, ecological cooperation and resource acquisition capabilities may become key competitive factors.

The Difference Between Co-Processors and Layer 2

Unlike user-oriented Layer 2, co-processors are mainly aimed at application development. They can serve as acceleration components or modular components, applied in the following scenarios:

1. As an off-chain virtual machine component of ZK Layer2
2. Provide off-chain computing power for applications on the public blockchain
3. As a verifiable oracle for cross-chain data
4. Serve as a bridge for cross-chain messaging

The potential of the co-processor goes far beyond this; it is expected to reconstruct various middleware in the Blockchain ecosystem, including oracles, data queries, cross-chain bridges, and more.

Challenges Faced by Co-processors

1. High development threshold: ZK technology is complex and requires mastery of specific languages and tools.
2. The track is still in its early stages: each project is still exploring the optimal technology route.
3. Insufficient hardware support: Dedicated ZK hardware has not yet been commercially used on a large scale.
4. Technological Homogeneity: It is difficult to form significant technological advantages, and competition may shift towards resource acquisition capabilities.

Future Outlook

ZK technology is expected to reshape the development paradigm of blockchain applications. The application prospects of general-purpose ZK co-processors are broad and, in theory, can realize corresponding versions of any Web2 application on the blockchain.

The two key indicators for achieving "mass adoption" are: a fully on-chain real-time provable database and low-cost off-chain computation. This goal requires gradual iterative implementation, and the commercialization of ZK computing chips is also an important prerequisite.

The current cycle lacks genuine innovation, providing a window of opportunity to build the next generation of "mass adoption" technologies and applications. It is expected that in the next round of the cycle, the ZK industry chain will achieve commercial landing, laying the foundation for on-chain interactions in Web3 that can accommodate billions of users.