Skip to main content

This site is for educational purposes only. Nothing here constitutes financial advice.

Topic 57 of 189

zkSync Era Architecture

How zkSync Era's prover (Boojum), native account abstraction, and unique EVM-compatibility-via-zkEVM design differ from competing rollups — and what the architectural choices mean for users and developers.

Beginner
8 min readUpdated November 2026Block Clarity Hub Editorial Team

The basics

What zkSync Era Is

zkSync Era is a zero-knowledge rollup launched in March 2023 by Matter Labs. It compresses many L2 transactions into a single ZK proof that Ethereum verifies — meaning you get Ethereum's security at a fraction of the cost. Era is part of a broader zkSync ecosystem: zkSync Lite (the original 2020 rollup for payments, swaps and NFTs, whose block production stopped on 4 May 2026) and Era (general-purpose smart-contract rollup). In a 4 September 2026 notice, Matter Labs said that within the next six months chains running EraVM, the execution environment Era runs on, will begin a transition in which EraVM is going to be retired, and named ZKsync Atlas, which runs the EVM natively, as EraVM's successor; exact steps and dates for funds held in smart contracts were still to be published.

What Makes Era Different from Optimism / Arbitrum

Optimism and Arbitrum are optimistic rollups — they assume transactions are valid unless someone proves otherwise via a fraud proof, with a 7-day withdrawal delay. zkSync Era is a ZK rollup — it generates a cryptographic proof of every batch's validity, posted to L1. The trade-off: ZK rollups have higher proving costs but allow withdrawal in about a day instead of a week, and rely on validity proofs rather than on someone noticing and challenging invalid state.

Native Account Abstraction

Era has native account abstraction at the protocol level: accounts can initiate transactions like an EOA but can also carry arbitrary logic like a smart contract, and regular EOA wallets still work. (Starknet also has native account abstraction.) This means features that require ERC-4337 plumbing on other chains — paymasters, multi-sig wallets, signature abstraction — are built into Era at the protocol level. Pay gas in USDC? Built in. Use a passkey instead of a seed phrase? Built in. Era's UX advantages flow from this single architectural choice.

  • ZK rollup with withdrawals in about a day (a 24-hour execution delay as of 10 November 2026) rather than optimistic rollups' 7-day delay
  • Native account abstraction at the protocol level — no separate ERC-4337 layer needed
  • EVM-compatible but not EVM-equivalent — compile for EraVM with zksolc, or run unmodified EVM bytecode through the EVM interpreter (some opcodes unsupported, higher fees)
  • Boojum (the current prover) generates STARK-based proofs wrapped in SNARK for cheap L1 verification

Key Takeaways

  • zkSync Era is a ZK rollup that ETHBEAT (formerly L2BEAT) lists at about $216 million total value secured and Stage 0 (10 November 2026)
  • Cryptographic validity proofs replace fraud-proof challenge windows
  • Native account abstraction enables better UX without separate protocols
  • Compatible with most Solidity contracts: compile for EraVM, or deploy unmodified EVM bytecode through the interpreter at higher fees

Going deeper

Boojum — The Current Prover

Boojum is zkSync's STARK-based prover, introduced in 2023, replacing an earlier SNARK-based prover. In September 2026 Matter Labs said it is developing EraBender, an Airbender-based prover, as a potential second proving system to run alongside Boojum, subject to testing, audit and production readiness. Boojum proves execution traces using the Goldilocks field (a small finite field that fits in 64-bit registers, enabling faster proving on commodity CPUs) and FRI (Fast Reed-Solomon Interactive Oracle Proof) commitments. The final STARK proof gets wrapped in a SNARK for cheap on-chain verification on L1 — this hybrid approach captures the best of both worlds: fast STARK proving + small SNARK proofs for cheap Ethereum verification.

EVM Compatibility — What Works and What Doesn't

Era's compiler (zksolc) takes standard Solidity source and compiles it to EraVM bytecode, and since the V27 upgrade (executed on mainnet in May 2025) an EVM bytecode interpreter can also run unmodified EVM bytecode without recompilation, at fees ZKsync's docs say may be 1.5x to 4x higher than for native EraVM contracts. Most contracts work without changes. The catches: some opcodes have different gas costs (because they're more expensive to prove inside a ZK circuit), some opcodes are unsupported (mostly low-level CALL variants and SELFDESTRUCT), and the precompiled contracts have different addresses or aren't available. Contracts using inline assembly or unusual opcodes may need adjustments. Popular protocols such as Uniswap V3 and Aave V3 are deployed on Era (DefiLlama lists both). ZKsync's September 2026 announcement says Atlas, EraVM's successor, 'runs EVM natively, follows Ethereum's protocol rules, works with mainstream compilers'.

ZK Stack and Hyperchains

zkSync Era is one chain built on the 'ZK Stack' — Matter Labs's open framework for deploying additional ZK rollups. Other chains built on the ZK Stack ('Hyperchains'; the current ZKsync docs call them 'ZKsync chains') include several application-specific deployments, and they do not all use the same prover: as of 10 November 2026, ETHBEAT (formerly L2BEAT) lists Abstract on Boojum and ADI Chain on Airbender. The vision: many ZK rollups on shared ZKsync infrastructure, interoperating via shared L1 settlement. This is zkSync's answer to OP Stack's Superchain concept — a modular framework where new chains can spin up with battle-tested ZK infrastructure.

Native AA in Practice

On Era, accounts can be smart contracts from day one (regular EOA wallets also work). This means a smart-account wallet can offer features (gasless onboarding, social recovery, session keys, paying gas in any token) that are first-class citizens of the protocol, not bolted on via ERC-4337. The trade-off: deploying a smart account has a one-time gas cost. Era's UX for new users is notably good — gasless first transactions are routine via paymasters.

Key Takeaways

  • Boojum uses STARK proving wrapped in SNARK verification — fast proving + cheap on-chain verify
  • Most Solidity contracts work, either compiled for EraVM or as unmodified EVM bytecode via the interpreter
  • ZK Stack lets new chains deploy on shared ZKsync infrastructure (provers include Boojum and Airbender) — Era's answer to OP Stack
  • Native AA produces noticeably better UX than ERC-4337 retrofits on EOA chains

Under the hood

Sequencer Decentralisation Roadmap

Like most production rollups, Era currently runs a single centralised operator (ETHBEAT: 'The system has a centralized operator'). Decentralised sequencing work is now described in terms of ZKsync OS: Matter Labs' Q3 2026 report says BFT consensus for the ZKsync OS sequencer moved from design to implementation. Until decentralisation lands, sequencer trust assumptions apply — the sequencer can technically delay or reorder transactions, and ETHBEAT notes that users can submit transactions to an L1 queue but cannot force them, and that the operator can censor withdrawals through a transaction filterer without delay.

Bridge Architecture and L1 Settlement

Era's bridge to L1 is canonical — contracts on Ethereum verify the rollup's state root and proof. Deposits are near-instant; withdrawals require a proof confirmation and then an execution delay, a security window before a batch is final. That delay was 3 hours until ZKsync governance proposal ZIP-17 let each EraVM chain raise it to 24 hours, and as of 10 November 2026 Era's own execution delay is 24 hours (ETHBEAT recorded the change on 8 October 2026), so a withdrawal now takes at least a day rather than a few hours. Compared to optimistic rollups' 7-day delays, this is a major UX improvement. The bridge has been audited multiple times and battle-tested — but as with any cross-domain bridge, smart-contract risk is non-zero. Wormhole, Multichain, Ronin all showed bridge risks; canonical rollup bridges have a strong record but aren't risk-free.

ZK Rollup Economics

ZK rollup costs decompose into: data availability (now published as cheap EIP-4844 blobs or calldata, per ETHBEAT), L1 verification cost (per batch settlement), prover cost (off-chain, but real). Era amortises these across the batch. The prover cost is hidden from users but borne by the sequencer; it's part of why the sequencer's revenue is meaningful.

The ZK Rollup Field Generally

zkSync Era is one of several ZK rollups. Others: StarkNet (Cairo VM, different ecosystem), Scroll (different EVM-compatibility approach), Linea (Consensys-backed, different stack), Taiko (an Ethereum-equivalent rollup using several proof systems), Aztec (privacy-focused, and still labelled 'Alpha' as of its V6 release on 3 November 2026). Polygon zkEVM, which used a different ZK proof system, was sunset in July 2026. Each makes different tradeoffs in EVM compatibility, prover speed, decentralisation, and ecosystem strategy. As of 10 November 2026, ETHBEAT (formerly L2BEAT) lists StarkNet at Stage 1 and ZKsync Era, Linea and Scroll at Stage 0.

ZK Doesn't Mean Privacy

Despite the name, ZK rollups are not privacy chains. The ZK proof is proving that the state transition was valid — not hiding what happened. All transaction details on Era are public, just like Ethereum. Privacy-focused ZK chains (Aztec, Aleo, Penumbra) use ZK for confidentiality; rollups use ZK for scaling. The shared cryptography led to widespread confusion; the use cases are unrelated.

Key Takeaways

  • Era's sequencer is still centralised: users can queue transactions on L1 but cannot force their inclusion (ETHBEAT, 10 November 2026), and decentralisation work is now described for ZKsync OS
  • Withdrawal latency is about a day (Era's execution delay is 24 hours as of 10 November 2026), not the 7 days of optimistic rollups
  • User fees are kept low by EIP-4844 blob space + batched proof verification
  • The ZK rollup field is competitive, but most are still early on decentralisation: ETHBEAT lists StarkNet at Stage 1 and ZKsync Era, Linea and Scroll at Stage 0 (Polygon zkEVM was also a ZK rollup until its July 2026 sunset)
  • ZK in rollups means 'cryptographically verified scaling,' not privacy

Related Content

MEV-Share Mechanics

How Flashbots' MEV-Share lets users share metadata about their transactions in exchange for a cut of the MEV searchers extract — a structural shift from 'searchers prey on users' to 'searchers and users share value.'

Solver Networks

How solver markets shape DEX prices, the competitive structure across CoW Swap, 1inch Fusion, Hashflow, and other venues, and why solver consolidation is a central debate in DeFi market structure.

Liquidation Cascades

The mechanical structure of cascading liquidations — Black Thursday, May 19 2021, August 2024, October 2025 — what amplifies them, who profits, and why over-leverage creates correlated risk across the entire DeFi stack.

Flash Loan Economics

Why flash loans exist, the legitimate uses (arbitrage, collateral swaps, refinancing), and how the same primitive that enables capital-efficient DeFi also enables governance attacks and oracle manipulation.

StarkNet Architecture

How StarkNet's Cairo VM, account-abstraction-native model, and STARK prover differ from EVM rollups — and what Madara, prover decentralisation, and the broader Starkware ecosystem mean for the network's future.

Arbitrum Nitro and Stylus

How Arbitrum Nitro evolved from Arbitrum One's custom architecture to running Geth-on-WASM, what BoLD fraud proofs change, and what Stylus's WASM extension brings (Rust/C/C++ smart contracts).

Optimism Bedrock and the Superchain

How OP Mainnet's Bedrock architecture, fault-proof rollout, and Superchain interop strategy reshape Optimism from a single rollup into a widely deployed modular L2 framework (Optimism says 54 mainnet chains ran on the OP Stack as of October 2026).

Base Architecture

How Coinbase's Base L2 launched on the OP Stack and has been moving to its own Base-operated stack, how Coinbase's exchange-side integration drives uniquely smooth UX, what the sequencer's centralised reality means today, and the path to a 'stage-2' rollup.

References & further reading