Ethereum Layer 2 Networks: Arbitrum, Optimism, and zkSync's Swiss Connections

Ethereum Layer 2 Networks: Arbitrum, Optimism, and zkSync’s Swiss Connections

The fundamental engineering challenge of Ethereum has always been throughput. Ethereum mainnet, in its current form, processes approximately 12-15 transactions per second — a figure that reflects deliberate choices about decentralisation and security, not engineering limitation alone. Every node in the Ethereum network must process every transaction; the throughput of the whole network is bounded by what a single decentralised node can handle. Increase throughput by relaxing this constraint, and you risk centralisation. Maintain the constraint, and you limit what the network can do.

Layer 2 networks resolve this trilemma not by weakening Ethereum’s constraints but by routing activity above them. Instead of executing every transaction on Ethereum mainnet (Layer 1), L2 networks execute transactions in a separate environment with different throughput characteristics, then post compact proofs or data summaries back to Ethereum mainnet. Ethereum mainnet’s validators verify the proof or data; Ethereum’s security properties are inherited without Ethereum processing every individual transaction.

This architecture — execute off-chain, verify on-chain — is conceptually straightforward but technically demanding. The devil is in the verification mechanism.

Rollups vs. Sidechains: The Critical Distinction

Before examining specific networks, a critical architectural distinction: rollups are not the same as sidechains, and the difference matters for security.

Sidechains (such as the original Polygon PoS, Ronin, or BNB Chain) are independent blockchains with their own consensus mechanisms and validator sets. They connect to Ethereum through bridges — but those bridges are independent security systems, not Ethereum security. If a sidechain’s validators act maliciously or its bridge is exploited, funds can be lost. The security of a sidechain is only as strong as its own consensus mechanism, not Ethereum’s. The Ronin bridge hack (March 2022, ~$625M stolen) is the canonical example of sidechain bridge vulnerability.

Rollups execute transactions off-chain but post either transaction data (in optimistic rollups) or validity proofs (in ZK rollups) back to Ethereum mainnet. This data posting is the security mechanism: Ethereum validators verify that the rollup’s state transitions are valid (or, in optimistic rollups, have been unchallenged during the fraud proof window). A rollup that correctly posts its data to Ethereum inherits Ethereum’s security — an attacker who wants to steal funds from an Ethereum rollup must either break Ethereum’s consensus or break the rollup’s proof system. Both are substantially harder than compromising a bridge between independent chains.

The architectural line between rollups and sidechains has blurred somewhat in practice — several networks have begun as sidechains and migrated toward rollup architecture, and some call themselves “validiums” (posting proofs to Ethereum but keeping data off-chain, a middle-ground with different security trade-offs). But the conceptual distinction between Ethereum-secured execution environments and independently-secured chains is real and consequential.

Optimistic Rollups: Arbitrum and Optimism

Optimistic rollups are so named because they optimistically assume transactions are valid — they post transaction data to Ethereum without immediately proving correctness. The security mechanism is a challenge period.

How Optimistic Rollups Work

An optimistic rollup operator (called a “sequencer”) executes a batch of transactions off-chain, produces a new state root representing the result, and posts the transaction data (calldata, or post-EIP-4844, blobs) and the new state root to an Ethereum smart contract. The transaction data is sufficient for anyone observing Ethereum to reconstruct the rollup’s state independently.

If the posted state root is fraudulent — if the operator included invalid transactions or altered outcomes — any party monitoring the rollup can detect the fraud and submit a fraud proof during the challenge period (typically 7 days). The fraud proof demonstrates to the Ethereum smart contract that the posted state root is incorrect, causing the fraudulent state root to be rejected and the fraudulent operator to lose a staked bond.

This mechanism creates a 7-day withdrawal window for optimistic rollup users: funds cannot be withdrawn to Ethereum mainnet with instant finality because the 7-day challenge period must expire first. Third-party liquidity providers (such as Across, Hop, and Stargate) offer fast withdrawal services — they front the funds immediately from their liquidity pools and wait for the challenge period themselves — but this introduces a liquidity cost.

Arbitrum One

Arbitrum One, developed by Offchain Labs, is the leading Ethereum rollup by total value locked (TVL) as of early 2026. Arbitrum uses a variant of optimistic rollup technology called interactive fraud proofs (or “multi-round fraud proofs”), where a fraud challenge is resolved through a binary search process that narrows the disputed computation down to a single instruction, dramatically reducing the on-chain cost of resolving disputes compared to single-round fraud proof approaches.

Arbitrum One is governed by the Arbitrum DAO, which holds the ARB governance token. The DAO controls protocol upgrades, treasury allocations, and the Arbitrum Orbit framework — a system enabling third parties to deploy custom “L3” chains that settle to Arbitrum rather than directly to Ethereum.

Arbitrum’s ecosystem is among the most mature in L2: Uniswap, Aave, GMX (Arbitrum’s native perpetuals protocol), Camelot, Radiant Capital, and hundreds of other protocols are deployed on Arbitrum One. The network’s developer tooling (full EVM equivalence, identical Solidity and Hardhat workflows) has minimised friction for Ethereum developers deploying to Arbitrum.

Optimism (OP Mainnet) and the Superchain

Optimism, developed by OP Labs, takes a different architectural approach to fraud proofs (single-round fraud proofs, simpler to implement but slightly less efficient at dispute resolution) while adding a distinctive governance and ecosystem model.

The Optimism Collective governs OP Mainnet through a two-house structure: the Token House (OP token holders voting on upgrades and treasury allocations) and the Citizens’ House (holders of non-transferable “Citizenship” NFTs voting on retroactive public goods funding — a mechanism for compensating projects that have created value for the ecosystem).

Optimism’s most strategically significant development is the OP Stack — an open-source software development kit for building Optimism-compatible rollups. The OP Stack enables any team to deploy an Optimistic rollup that shares core infrastructure and (optionally) security with OP Mainnet. Coinbase used the OP Stack to build Base, its L2 network, which became the fastest-growing rollup by new user acquisition through 2024-2025 due to Coinbase’s consumer distribution. Dozens of other OP Stack chains have launched, with Optimism positioning this as the “Superchain” — a network of interoperable L2s sharing a sequencing and proof infrastructure.

ZK Rollups: zkSync, Starknet, and Polygon zkEVM

ZK rollups use cryptographic validity proofs — either ZK-SNARKs (Succinct Non-interactive ARguments of Knowledge) or ZK-STARKs (Scalable Transparent ARguments of Knowledge) — to prove that a batch of transactions was executed correctly. Unlike optimistic rollups, ZK rollups do not rely on fraud challenges; the proof itself is the verification.

This architectural difference has profound implications:

• Instant finality: Once an Ethereum block containing a ZK proof is confirmed, the rollup’s state transition is final. There is no 7-day withdrawal window; users can withdraw to Ethereum mainnet in minutes (the time needed to prove a batch and include the proof in an Ethereum block).
• Privacy potential: The same ZK proof technology that proves transaction validity can, in more sophisticated implementations, prove transaction validity without revealing transaction details — enabling private transaction rollups.
• Higher proving costs: Generating ZK proofs is computationally expensive. The proving time and hardware cost is a meaningful operational expense for ZK rollup operators, though hardware acceleration (custom ASICs and FPGAs for proof generation) is actively reducing these costs.

zkSync Era: Matter Labs and Zug

zkSync Era, developed by Matter Labs, is a ZK rollup with EVM compatibility through a custom zkEVM implementation. Matter Labs — and this is the Swiss connection of direct relevance — is incorporated and headquartered in Zug, Switzerland.

Matter Labs was founded by Alex Gluchowski and colleagues in 2019 and has raised substantial venture funding (reportedly over $400 million as of 2023) to develop zkSync. The company’s Zug domicile places it squarely within Crypto Valley’s infrastructure, with access to FINMA regulatory engagement, Swiss foundation structures, and the Zug legal and banking ecosystem.

zkSync Era launched on mainnet in March 2023. Its zkEVM uses a custom approach to EVM compatibility: rather than proving the execution of EVM opcodes directly (which is computationally intensive), zkSync compiles Solidity contracts to a custom intermediate representation that is more efficient to prove. This trade-off achieves high EVM compatibility while maintaining proving performance, at the cost of some edge-case differences from standard EVM behaviour that developers must test for.

The ZK token, distributed in June 2024 through a large airdrop to ecosystem users, launched zkSync’s governance decentralisation. Matter Labs’ Swiss domicile means that the ZK token governance structure, treasury management, and regulatory engagement all interact with Swiss law and FINMA oversight — making zkSync’s governance one of the most significant current examples of Swiss-domiciled ZK rollup governance.

Starknet: StarkWare’s STARK-Based L2

Starknet, developed by StarkWare (Israeli company, co-founded partly by team members with European connections), uses STARK proofs — a cryptographic system with different trust assumptions than SNARKs (no trusted setup, larger proof sizes, but post-quantum security properties). Starknet’s native programming language is Cairo (not Solidity), creating an EVM incompatibility that limits developer accessibility but enables features that EVM-compatible systems cannot implement.

Starknet has attracted academic and research-oriented DeFi teams (including Ekubo, Starknet’s native concentrated liquidity DEX, and protocols focused on sophisticated financial instruments that benefit from Cairo’s expressive type system). The STRK token launched in early 2024, distributing governance power to ecosystem participants.

Polygon zkEVM and Linea

Polygon zkEVM, from Polygon Labs, takes the approach of proving EVM opcode execution directly — maximising compatibility with existing Ethereum contracts at higher proving costs. Linea, from ConsenSys, uses a similar approach with a custom proving system. Both have meaningful TVL and developer adoption, though neither has reached Arbitrum’s scale as of early 2026.

L2 TVL Landscape 2025

The L2 TVL distribution as of early 2026 reflects both technical merit and ecosystem development investments:

• Arbitrum One: Consistently the largest L2 by TVL, typically holding $10-15 billion across DeFi protocols
• Base: The fastest-growing network in terms of new user accounts, driven by Coinbase distribution; lower average TVL per user than Arbitrum, reflecting more retail activity
• OP Mainnet: The governance and Superchain network, holding significant TVL in long-established protocols
• zkSync Era: Significant TVL driven partly by ZK token airdrop incentives, with ongoing ecosystem development
• Starknet: Smaller TVL but technically differentiated ecosystem
• Blast: A network offering native yield on ETH and stablecoins, attracting TVL through incentive programs

L2Beat (l2beat.com) is the authoritative data source for L2 TVL, security classifications, and risk parameters — and has become essential reading for institutional risk assessment of L2 deployments.

EIP-4844: Blob Data and the Cost Revolution

The most significant practical development for L2 economics was EIP-4844 (Proto-Danksharding), activated in Ethereum’s Dencun upgrade in March 2024. EIP-4844 introduced a new transaction type carrying “blobs” — large data payloads attached to Ethereum blocks but not processed by the EVM, subject to a short retention period (~18 days) rather than permanent storage.

The impact on L2 costs was immediate and dramatic:

• Arbitrum transaction fees fell from ~$0.10-0.30 to below $0.01 for simple transfers
• Base and Optimism fees fell to fractions of a cent
• zkSync Era fees similarly collapsed
• Some L2 transactions became cheaper than a Starbucks coffee in many hours

This fee reduction fundamentally changed the economics of L2 deployment. Pre-EIP-4844, the data posting cost was the primary cost for L2 operators and was passed through to users. Post-EIP-4844, data posting costs became a small fraction of total cost, shifting the marginal cost equation toward the L2 operators’ proving and sequencing costs.

For Swiss institutional projects deploying on-chain — private banks exploring tokenised fund settlement, Swiss fintechs building blockchain-based payment rails, DeFi protocols with Swiss client bases — the post-EIP-4844 L2 economics have made previously impractical use cases economically viable.

Which L2s Are Swiss Institutional Projects Deploying On?

The institutional Swiss Web3 ecosystem’s L2 choices reflect several factors:

EVM Compatibility Priority: Swiss institutional projects using Solidity-based smart contract systems (the majority of institutional tokenisation infrastructure) favour fully EVM-compatible networks. Arbitrum, Base, and OP Mainnet receive the highest institutional deployment due to identical developer tooling.

Regulatory Comfort: zkSync Era’s Matter Labs Zug domicile creates a natural affinity for Swiss institutional projects that prefer to engage with Swiss-domiciled infrastructure providers. FINMA engagement with Matter Labs and familiarity with Swiss foundation structures make zkSync a natural choice for projects seeking jurisdictional alignment.

Institutional DeFi Protocols: SIX Digital Exchange (SDX) and Swiss banks exploring DeFi integration have primarily built on Ethereum mainnet and evaluated Arbitrum and Base as the L2 destinations for volume that mainnet cannot cost-effectively process.

Privacy Layer Needs: Projects requiring confidential transaction execution have evaluated Starknet (and privacy-focused L2 variants) for use cases where transaction details must remain private from public mempool observation.

The Path to Danksharding

EIP-4844 is explicitly a transitional step — “Proto-Danksharding” — toward the full Danksharding roadmap. The full vision, named after Ethereum researcher Dankrad Feist, involves:

PeerDAS (Peer Data Availability Sampling): The next step after EIP-4844, planned for a future Ethereum upgrade (Pectra or a subsequent release). PeerDAS enables light clients to probabilistically verify blob availability without downloading complete blobs, using data availability sampling — a technique from the distributed systems literature. PeerDAS can increase the number of blobs per block by an order of magnitude while maintaining the ability of light clients to verify availability.

Full Danksharding: The complete roadmap expands blob capacity from EIP-4844’s 3-6 blobs per block to potentially 64 or more blobs per block, each blob carrying ~125KB of data. At full capacity, Ethereum’s data availability layer could support tens or hundreds of L2 networks processing thousands of transactions per second each, all settling to Ethereum’s security.

The Danksharding roadmap is a multi-year engineering project. It depends on PeerDAS being successfully deployed and operating, on the development of distributed block building (to prevent the increased data load from centralising block production among high-bandwidth operators), and on continued L2 ecosystem development that generates the demand to fill expanded blob capacity.

The combination of the existing L2 ecosystem — Arbitrum, Optimism, zkSync, Base, and the networks still to launch — with the Danksharding data availability roadmap represents Ethereum’s answer to the scaling trilemma: not compromising on decentralisation or security, but building a layered architecture where execution scales above a secure, decentralised base layer. Switzerland’s Matter Labs, from its Zug base, is building a significant part of that architecture.

Frequently Asked Questions

What is the difference between an optimistic rollup and a ZK rollup?

Optimistic rollups post transaction data to Ethereum and assume validity by default, relying on a 7-day challenge period during which fraud proofs can be submitted to dispute invalid state transitions. ZK rollups post cryptographic validity proofs to Ethereum that mathematically prove every transaction batch was executed correctly — no challenge period is needed, enabling near-instant withdrawal finality. ZK rollups have higher proving costs but stronger security guarantees and faster finality; optimistic rollups are simpler to implement but require the withdrawal delay.

Why does the 7-day withdrawal period exist for Arbitrum and Optimism?

The 7-day period is the fraud proof challenge window — the time during which any observer who detects a fraudulent state root can submit a fraud proof to the Ethereum smart contract to reject it. Without this window, a sequencer could submit a fraudulent state and users would have no time to challenge it before funds are withdrawn to Ethereum mainnet. The window ensures that even if the rollup operator behaves maliciously, honest observers have sufficient time to respond. Third-party fast bridge providers (Across, Stargate) offer immediate liquidity by front-running this window.

What is Matter Labs and what is its connection to Switzerland?

Matter Labs is the company that developed zkSync Era, a ZK rollup for Ethereum. Matter Labs is incorporated and headquartered in Zug, Switzerland, making it one of the most significant ZK rollup infrastructure companies with direct Swiss domicile. The company’s Zug presence places it within Crypto Valley’s legal, regulatory, and banking ecosystem, and its ZK token governance structure interacts with Swiss foundation law and FINMA oversight.

What did EIP-4844 do to Layer 2 transaction costs?

EIP-4844, activated in Ethereum’s Dencun upgrade in March 2024, introduced “blobs” — a new data type for rollups to post transaction data to Ethereum at dramatically lower cost. Post-EIP-4844, L2 transaction fees fell 80-99%: Arbitrum simple transfers fell below $0.01, Base transactions dropped to fractions of a cent. The change made previously impractical use cases (micropayments, high-frequency DeFi operations, gaming transactions) economically viable on Ethereum L2 networks.

Which Ethereum Layer 2 networks are most used by Swiss institutional projects?

Swiss institutional projects primarily deploy on EVM-compatible networks — Arbitrum One, OP Mainnet, and Base — due to identical Solidity developer tooling. zkSync Era has a natural Swiss institutional affinity given Matter Labs’ Zug domicile and FINMA engagement. SIX Digital Exchange and Swiss banks exploring DeFi integration have primarily evaluated Arbitrum and Base as the highest-liquidity institutional venues. Projects requiring private transaction execution have explored Starknet and privacy-oriented L2 variants.

Related Coverage

• Ethereum: The Web3 Foundation and Zug’s Protocol Anchor
• Web3 Regulation in Switzerland: FINMA, the DLT Act, and MiCA Interaction
• Substrate: Parity Technologies’ Blockchain Framework and the Web3 Development Stack
• Solana Foundation in Zug: High-Performance Blockchain and Switzerland’s Role in the Solana Ecosystem
• DeFi Governance: How Aave and Compound Run Billion-Dollar Protocols Through Token Voting

Author: Donovan Vanderbilt | The Vanderbilt Portfolio AG, Zurich Published: 25 February 2026