ZK Settlement is Coming to Arbitrum

Earlier this year, a roadmap of the engineering work underway across the Arbitrum Platform outlined the path to help make it the best infrastructure for building financial products in the programmable economy. A key part of that roadmap is a multi-proving model powered by zero-knowledge (ZK) technology.

Bringing ZK proving to Arbitrum One* and dedicated blockchains built with Arbitrum Platform can give the ecosystem a faster path to L1 withdrawals. Instead of relying solely on the traditional seven-day challenge window, ZK proving can reduce asset settlement times from days to hours. This helps increase capital efficiency for users, bridges, and protocols, without compromising the security of the Arbitrum Platform.

The following milestones highlight the latest progress toward enabling live ZK proving.

1. Arbitrum blocks can now be ZK proven

ZK proving now works in Arbitrum block validation. Point it at a real mainnet block and it generates a zero-knowledge proof that the block executed correctly by running the actual Arbitrum state transition function inside SP1, Succinct's zkVM. This is not a reimplementation or an approximation of Arbitrum, it is the ZK prover running the exact same code as Arbitrum’s optimistic prover. 

2. Stylus works inside the prover

Arbitrum isn't EVM-only. In fact, Arbitrum Stylus extends the EVM with support for Rust, C/C++, Move, and other languages that compile to WASM, opening the door to ~20 million existing developers and their battle-tested libraries.

The WASM-based prover implementation has been extended to prove Stylus contracts alongside solidity contracts. Arbitrum is the first battle-tested production stack to make its existing EVM and WASM execution ZK-provable.

3. A new high-performance validator, written in Rust

Validators are the nodes that re-execute the blockchain and check every state update. Arbitrum's core validation machinery has long been written in Rust for performance, but it ran wrapped inside a Go service. SP1, the ZK prover being utilized, is also built around Rust. So, the validator was rebuilt as a standalone Rust service, and it has completed its first rounds of QA.

Rebuilding the validator in Rust accomplished two jobs at once. Firstly, it's a leaner validator that slots into the existing Nitro architecture. Secondly, it makes ZK proving a first-class validation mode rather than a bolt-on: standard validation and proof generation will live behind the same interface, so a node can validate a block or prove it through one path.

4. ZK plugs directly into BoLD settlement

BoLD is the protocol Arbitrum uses to settle to Ethereum: anyone can post an assertion (a claim about the blockchain's new state), and it's confirmed after a challenge window passes - with any dispute resolved by an interactive challenge game that guarantees the honest claim wins.

BoLD has been extended to accept ZK proofs. Instead of waiting out the challenge window, an assertion is confirmed as soon as a valid ZK proof lands and a new Fast Confirmation Committee attests to the same result. And because BoLD's dispute game remains underneath, ZK is purely additive: if no proof shows up, the blockchain settles exactly as it does today. This is multi-proving at its finest: ZK proofs, attestations, and fraud proofs working together.

For a deeper dive into how multi-proving works with BoLD, see this breakdown:

5. Systematically lowering proving costs 

Arbitrum proving was benchmarked head-to-head against vanilla Ethereum proving, and the extra costs were traced to their sources, down to where every cycle goes.

From there, big expense drivers have been addressed one by one: the proving pipeline optimized for WASM execution, runtime overhead inside the prover reduced, intensive cryptography offloaded to the zkVM's optimized precompiles. Further structural wins, like multi-block proving, come next.

6. Moving from research branch to production codebase

ZK proving is now merging into the main Arbitrum Nitro codebase, the same codebase every Arbitrum blockchain runs today. The work landed first on a dedicated feature branch, and it's being moved into the main branch piece by piece. Once complete, ZK execution of Arbitrum lives in the production code path, not a side experiment or separate repository, ready for dedicated blockchains on the Arbitrum Platform to adopt.

Next steps

To achieve ZK proving in production, there are three key upcoming work streams:

1. Drive down costs and latency by further optimizing proving and moving to Reth-based execution
2. Close the loop on end-to-end verifiability by proving the blockchain's L1 message inbox 
3. Wire these capabilities into the node so blockchains can easily toggle ZK settlement as a configuration. 

Together with other initiatives underway, executing on these core milestones will pave the way for efficient, multi-prover settlement across the Arbitrum Platform.

*Offchain will bring a proposal to the DAO to upgrade Arbitrum One to ZK settlement. If passed, other blockchains built with Arbitrum Platform will be able to upgrade at will.

Disclaimer: The information provided is for informational purposes only and does not constitute financial, technological, or any other form of advice. Please conduct your own independent research and consult with a qualified professional before making any decisions. Statements regarding product roadmap, planned functionality, developments in progress, and future direction are forward-looking, reflect current expectations only, and are not commitments or guarantees.