What Is Scroll? The Bytecode-Equivalent zkEVM Ethereum L2, Explained

Most Layer 2s pick a lane: chase the lowest fees, chase the highest throughput, or chase perfect Ethereum compatibility. Scroll picked the third, and picked it harder than almost anyone. It's a zero-knowledge rollup that runs the actual Ethereum Virtual Machine at the bytecode level — so the contract you deployed on mainnet, the tooling you already use, and the opcodes you already know all work unchanged, but with the security of a validity proof instead of a fraud-proof delay. That purism has costs, and Scroll is candid about them. This is how the chain actually works, what the SCR token does and doesn't do, and where it really sits on the decentralisation ladder in 2026.

Who built it, and why

Scroll came out of the research-heavy end of the ZK world. Its co-founders — Ye Zhang, Sandy Peng, and Haichen Shen — built the project in close collaboration with the Ethereum Foundation's applied-ZK team, and that pedigree shows in the design choices. Where some rollups optimised for a splashy TPS number, Scroll optimised for being indistinguishable from Ethereum itself. It open-sourced its circuits, published its research, and took the slower road of proving genuine EVM execution rather than a bespoke virtual machine that developers would have to adapt to.

The philosophy carries into how the team talks about the space. Ye Zhang has publicly called the idea of L2s extracting large fees "toxic," arguing that rollups should push value back to Ethereum rather than treat users as a revenue faucet. Whether or not you buy the framing, it tells you what Scroll is trying to be: an extension of Ethereum, not a walled garden competing with it. Mainnet went live on 17 October 2023.

The architecture: a validity rollup that runs real EVM bytecode

Start with the rollup type, because it's the thing people most often get wrong. Scroll is a zk-rollup — a validity rollup — not an optimistic one. The distinction matters. An optimistic rollup (Arbitrum, Optimism) assumes batches are valid and gives challengers a window, typically seven days, to prove fraud; that's why withdrawing back to Ethereum the canonical way takes about a week. Scroll instead generates a cryptographic proof that each batch executed correctly and posts that proof to Ethereum, where a rollup contract verifies it against the state root and the transaction-data commitment. Once verified, the state is final — there is no fraud-proof waiting period. (For the broader landscape, see our best Ethereum L2s guide.)

The second pillar is EVM equivalence. Scroll is classified as a Type 2 zkEVM: it mirrors Ethereum's behaviour so closely that contracts compiled for mainnet deploy without recompilation, and execution traces match L1 EVM behaviour opcode for opcode. In practice that means Foundry, Hardhat, MetaMask, block explorers, and existing Solidity all "just work." Compare this to zkEVMs that make small semantic changes for prover efficiency — those run faster but occasionally break edge-case contracts or tooling. Scroll chose maximal compatibility, and accepts that proving exact EVM opcodes is more expensive.

Data availability lives on Ethereum. Scroll posts its transaction data to L1 (using blob space, the cheap data lane introduced by EIP-4844), so anyone can reconstruct the L2 state from Ethereum alone — you are not trusting Scroll to hand you the data. That's what makes it a rollup rather than a validium.

The Euclid upgrade: from Halo2 circuits to an OpenVM prover

The most important thing that happened to Scroll's architecture didn't happen at launch — it happened in 2025 with the Euclid upgrade, the biggest change since mainnet.

Originally Scroll proved execution with hand-written Halo2 circuits using KZG commitments — powerful, but painstaking to build and audit, and a bottleneck on how fast the tech could evolve. Euclid deprecated those circuits in favour of a new prover built on OpenVM, a general-purpose RISC-V zkVM. Instead of encoding the EVM directly into bespoke circuits, Scroll now proves a RISC-V program, which is far more flexible to maintain and upgrade. Alongside it, Scroll dropped its custom zkTrie and adopted Ethereum's own Merkle-Patricia Trie (MPT) for state — another step toward being a faithful mirror of L1.

The payoff was concrete: lower fees, higher throughput, better compatibility, and a meaningful security improvement — enough that Euclid pushed Scroll to a Stage 1 rollup classification on L2BEAT. Scroll's own figures put the throughput gain around 4x and fee reductions as high as ~90%, though real-world numbers move with network conditions and Ethereum blob prices.

The deeper significance is architectural direction. Custom circuits were always going to be a maintenance treadmill — every new EVM feature meant new circuit engineering and fresh audits. A general-purpose zkVM changes the economics: improvements to the underlying proving stack accrue to Scroll without a bespoke rewrite, and adopting Ethereum's MPT means Scroll's state model no longer diverges from L1's at all. It's the kind of upgrade that doesn't grab headlines but quietly determines how fast a chain can keep pace with Ethereum over the next several years.

Fees, speed, and the honest performance picture

Day to day, Scroll feels like a cheap, fast Ethereum. Fees are a small fraction of mainnet's — you're paying cents for typical transactions, with the exact cost dominated by the L1 data (blob) cost that Scroll passes through, plus a small execution and proving component. You pay in ETH.

Where Scroll is honest about a trade-off is raw throughput. Because it proves exact EVM behaviour rather than a slimmed-down VM, its sustained transactions-per-second sits below the flashiest L2s — reported averages have been in the low-to-mid tens of TPS, with the roadmap aiming higher as parallelised proving and the OpenVM pipeline mature. If your mental model is "compatibility-maximising zkEVM," that trade-off is the point, not a bug. For a refresher on why any of this beats transacting on L1, see what is a Layer 2.

The SCR token: governance, not gas

This is the part people most often get wrong, so be precise. SCR exists. It launched in October 2024 with a total supply of 1,000,000,000, a Binance listing, and a Binance Launchpool. The airdrop distributed 7% of supply to early adopters in its initial round (part of a larger allocation earmarked for airdrops), with sizeable slices set aside for ecosystem growth and the Scroll Foundation.

But SCR is not a gas token — you do not pay fees in it. It is a governance token for Scroll DAO, giving holders a say in protocol decisions. Its intended longer-term role is to help decentralise Scroll's sequencer and prover, and potentially take on utility functions there — but any fee, staking, or utility role beyond governance is explicitly gated behind future governance votes. As of 2026, holding SCR is a claim on governance, not a claim on network revenue.

The launch was not without drama. The airdrop allocation drew loud complaints from farmers who felt under-rewarded, and there were public accusations around founder-linked wallets moving large amounts around the airdrop — worth knowing as history, and a reminder that token launches and their optics are a genuine risk factor, not just a footnote.

The ecosystem: DeFi-led, compatibility-native

Because anything that runs on Ethereum runs on Scroll, its ecosystem skews toward straightforward ports of proven DeFi primitives rather than exotic chain-specific apps. Nuri has been the leading native DEX — a concentrated-liquidity AMM in the Uniswap v3 mould — and Ambient Finance, a multichain concentrated-liquidity DEX, brought deep swap liquidity. Lending, yield, and bridging infrastructure round out the core, and Scroll ran a points-style incentive program ("Marks," rolled out in sessions) to bootstrap deposits into specific protocols.

On standing, be honest about the trajectory: Scroll's TVL peaked around $585 million in October 2024 and has fallen hard since. In early 2026 its single biggest protocol — Ether.fi, then Scroll's top fee generator — migrated to Optimism, pulling roughly $160 million of TVL and most of the fee revenue with it and dragging on-chain deposits down into the low tens of millions. The DAO responded by moving to cut costs, trim contributor roles, and even wind down its Security Council. Scroll was once talked about as a leading zkEVM by TVL, but the ZK-rollup pack (zkSync Era, Linea, Scroll) trades places as incentives and markets shift — treat any TVL ranking as a snapshot, and note that Scroll's is currently well off its peak.

How to get on

Getting onto Scroll is standard EVM procedure:

1. Add the network to an Ethereum wallet like MetaMask — either manually with Scroll's RPC details, or by letting a Scroll dApp prompt the addition. Everything you know from Ethereum carries over.
2. Bridge assets. The canonical, most trust-minimised route is the official Scroll bridge at scroll.io, which locks your ETH or tokens on L1 and mints the equivalent on L2. Canonical withdrawals back to L1 are slower (they wait on proof finalisation), which is normal for a rollup.
3. Or use a third-party bridge for speed and more source chains — faster, but you're trusting that bridge's liquidity and security, so weigh the trade-off.

The single most important habit: reach the bridge through the official site or your own bookmark, never through a link in a DM, ad, or search result. Fake-bridge phishing is the most common way people lose funds on any L2.

Risks and what to watch

• Centralised sequencer. The team still runs the sequencer that orders transactions. That's normal for the stage Scroll is at, but it means live censorship-resistance and liveness still depend on one operator today. Watch the sequencer-decentralisation roadmap.
• Prover network is early. Open in design, but permissionless proving is not yet mature. Stage 1 is real progress; it is not Stage 2 (the point at which users are fully protected even from the operators).
• Token model is unproven. SCR's value proposition rests on future governance decisions about sequencer/prover decentralisation and utility. Until those land, it's a governance token whose deeper role is a promise, not a fact — and it's volatile.
• Competitive pressure and outflows. zkSync Era and Linea are direct zkEVM rivals with their own war chests; the optimistic giants (Arbitrum, Base) still dwarf the whole ZK category on activity. Compatibility purism is a real edge, but it doesn't guarantee liquidity — and Scroll learned this the hard way in early 2026 when Ether.fi, its top protocol, decamped to Optimism and took ~$160M of TVL with it, prompting DAO cost cuts. Liquidity can leave fast.
• Standard L2 risks. Smart-contract bugs in the rollup or bridge contracts, and phishing around bridging and any future token claims, are the immediate practical dangers.

Bottom line

Scroll is the zkEVM that refused to cut corners on Ethereum compatibility. It proves real EVM bytecode, posts its data and proofs to L1, and — since the Euclid upgrade swapped fragile hand-built circuits for a flexible OpenVM prover — does it faster, cheaper, and at a Stage 1 security level. The honest caveats are the ones Scroll doesn't hide: the sequencer is still team-run, decentralisation is a roadmap rather than a finished product, and SCR is a governance token whose bigger job is still ahead of it. If you value a rollup that behaves exactly like Ethereum and inherits its security through validity proofs, Scroll is one of the most principled bets in the L2 field — just size your expectations to a chain that's still mid-journey.

For related reading: best Ethereum L2s, what is a Layer 2, what is DeFi, and how to bridge to a Layer 2 safely.

Related guides

• Starknet
• Status Network
• Linea
• ZKsync Era
• Optimism

Not financial advice. SCR is volatile and Scroll's decentralisation and token mechanics can change — always verify details on Scroll's official channels and L2BEAT.