Scroll’s bytecode equivalence and performance shift

The Bytecode Advantage

Scroll’s recent September proof recursion upgrade changed the performance hierarchy after it overtook the transaction per second benchmarks of zkSync Era and Polygon. I find that the decision to prioritize bytecode equivalence over the LLVM-based approach of zkSync Era defines the user experience. While zkSync Era uses its compiler to reach Type 4 compatibility, Scroll targets Type 2 status. This means a developer can deploy an Ethereum smart contract directly to Scroll with zero changes to the execution semantics. You should look at the bytecode compatibility if you want to avoid audit re-validation for protocols already audited on Ethereum. Polygon zkEVM remains a Type 2 or Type 3 solution, but Scroll provides a similar level of equivalence. Scroll supports the Cancun EVM target and the latest Solidity version 0.8.26. It also supports most precompiled contracts. In 2026, Scroll maintains a throughput of 20 to 40 transactions per second, while zkSync Era processes between 40 and 60 transactions per second. Scroll holds approximately 800 million dollars in total value locked, whereas zkSync Era holds around 2 billion dollars. zkSync transaction fees are typically 1 to 5 percent of Ethereum mainnet costs.

Architecture and Security Fixes

The technical architecture relies on the Scroll Node, which includes the Sequencer, the Coordinator, and the Relayer. The Sequencer retrieves a batch of transactions from its mempool, generates a new L2 block, and creates a new state root. The Coordinator receives the execution trace and dispatches it to a randomly selected Roller to generate proofs and then aggregates them into a single aggregate proof. The Relayer monitors the rollup contract for data availability and relays messages between Ethereum and Scroll. The Rollup and Bridge contracts connect Scroll to the Ethereum base layer because they store state roots and block data in the Ethereum state for data availability and provide block finality. I saw the Scroll Security Council execute an emergency on-chain verifier update on May 26, 2025, to fix a bug in a Plonky3 dependency. The Plonky3 team found two bugs in the FRI verifier, and the Scroll team worked with Axiom to deploy the patch. This vulnerability involved missing randomness in mixed domain handling during FRI folding. This bug could not be meaningfully exploited because the sequencer and provers are not permissionless. The Rollers use hardware acceleration to process proofs in parallel.

Economic Trade-offs and Fees

The fee structure for Scroll transactions includes an L2 gas fee and an L1 data fee. The L2 gas fee equals the L2 gas price multiplied by the L2 gas used. The L1 data fee covers the cost of sending transaction data to Ethereum and only applies to L2-initiated transactions. I find the gas fees on Scroll to be high, particularly for users using the native bridge to move assets from Ethereum. This remains a frequent complaint for the network. Since the DarwinV2 upgrade, the maximum reorg depth is 17 blocks. Unlike zkSync, which has a ZK token for governance, Scroll has no native token and no announced plans for one. zkSync’s ZK token has a total supply of 21 billion and reached 695,000 wallets during its 2024 airdrop. zkSync also has native account abstraction built into its protocol, while Scroll relies on ERC-4337 via bundlers. This provides a different user experience compared to zkSync. Will the lack of a governance token hinder the growth of the Scroll ecosystem?