What Are Layer 2 Solutions? Scaling Ethereum and Beyond
If you’ve ever tried to do anything on Ethereum during a busy period — swap tokens, mint an NFT, move funds — you’ve felt the problem firsthand. Gas fees climbing to $50, $100, sometimes more. Transactions stuck in a queue for minutes. The chain that was supposed to revolutionize finance, brought to its knees by a popular JPEG drop or a DeFi yield farm.
Layer 2s are the fix. Or at least, the current best attempt at one.
They’re not a single technology — they’re a category of solutions that all share the same basic idea: move most of the work off the main Ethereum chain (Layer 1), do it faster and cheaper, then post some proof of what happened back to L1 for security. The result is a network that can handle dramatically more transactions without compromising on the security guarantees that make Ethereum worth using in the first place.
Here’s how that actually works.
The Scaling Problem Explained in Plain Terms
Ethereum’s base layer processes transactions roughly one at a time, in order, verified by a global network of nodes. That’s what makes it trustless and censorship-resistant — every node re-executes every transaction and agrees on the result. But it’s also what makes it slow. The chain processes somewhere in the range of 15-30 transactions per second. For comparison, Visa handles around 1,700 TPS on average, with peaks much higher.
When demand exceeds capacity, users bid for block space by paying higher gas fees. During bull market peaks, this has made Ethereum genuinely unusable for anyone transacting amounts smaller than the fee itself — which is most of the planet.
The naive solution is to just make the base layer faster — bigger blocks, faster finality. The problem is that this concentrates validation power among fewer, more powerful nodes. You get throughput, but you sacrifice decentralization. This tradeoff is sometimes called the blockchain trilemma: security, decentralization, and scalability — pick two, at least until you get clever.
Layer 2s are clever.
How Layer 2s Work
The core insight: Ethereum’s real job is to be the final arbiter of truth. It doesn’t need to personally execute every transaction — it just needs to be able to verify that those transactions happened correctly.
So Layer 2s batch thousands of transactions together, process them off-chain, and then post a compressed summary (or a cryptographic proof) to Ethereum. The L1 stores the proof. If anyone tries to cheat, the L1 can catch it or the proof itself makes cheating mathematically impossible.
This reduces the cost of each individual transaction dramatically — often by 90% or more — while still inheriting Ethereum’s security guarantees.
There are several approaches to how Layer 2s handle this, and they’re not equivalent.
The Main Types of Layer 2
Rollups
Rollups are currently the dominant L2 design and Ethereum’s official scaling roadmap. They execute transactions off-chain, then “roll up” a compressed batch and post it to Ethereum. The key split is in how they prove those transactions are valid:
Optimistic Rollups assume transactions are valid by default and only run a dispute process if someone challenges a transaction. The catch: there’s a withdrawal delay (typically 7 days) to allow time for challenges. Arbitrum and Optimism are the two major optimistic rollups. Arbitrum has dominated in terms of TVL and developer activity; Optimism is the foundation of the OP Stack, which powers a growing ecosystem of “superchain” networks including Base (built by Coinbase), Mode, and others.
ZK Rollups (Zero-Knowledge Rollups) use cryptographic proofs — specifically, zero-knowledge proofs — to mathematically verify every batch of transactions before posting to L1. No delay, no challenge period needed. Much faster finality, but far more technically complex to build. zkSync Era, StarkNet, Polygon zkEVM, and Scroll are the main players. ZK rollups are widely considered the long-term winning design, but the tech is still maturing.
State Channels
State channels let two parties transact directly off-chain as many times as they want, then settle the final outcome on-chain. Think of it like opening a tab at a bar — you run a bunch of transactions, then settle the total at the end. The Lightning Network on Bitcoin works this way. For Ethereum, state channels fell out of favor as rollups matured. They’re efficient for specific use cases (micropayments, gaming moves) but not general-purpose.
Plasma
Plasma was an earlier scaling approach that created child chains connected to Ethereum. It had significant limitations around supporting complex smart contracts and required users to actively monitor for fraud. Most projects that launched as Plasma chains (including the early Polygon PoS chain) have since pivoted toward rollup technology.
Sidechains
Technically not L2s (this distinction matters, even if it sounds pedantic). Sidechains like the old Polygon PoS are independent blockchains with their own consensus mechanisms. They connect to Ethereum via a bridge but don’t inherit Ethereum’s security. They can be faster and cheaper, but if the sidechain’s validators collude or fail, you’re exposed. True L2s derive security from L1; sidechains don’t.
The Major L2 Ecosystems in 2026
Arbitrum
The current leader by total value locked and developer activity. Arbitrum One is the flagship chain; Arbitrum Nova is a cheaper variant optimized for gaming and social applications. The Arbitrum ecosystem runs on the Nitro tech stack and has seen significant DeFi migration from Ethereum mainnet. GMX, Camelot, and a substantial chunk of the on-chain derivatives market have made Arbitrum home.
OP Stack / Optimism Superchain
Optimism created the OP Stack as an open-source L2 framework, then let other projects use it to launch their own L2s that share messaging infrastructure. The result is a growing network of chains: Base (by Coinbase), Optimism mainnet, Mode, Zora, and others. Base in particular has seen explosive user growth, driven partly by Coinbase’s distribution and the viral success of friend.tech in 2023. The Superchain concept — interconnected L2s that share security and messaging — is one of the more ambitious roadmaps in the space.
zkSync Era
The flagship ZK rollup from Matter Labs. Fully EVM-compatible (meaning Ethereum smart contracts can be ported over without a full rewrite). Has a large developer community and significant DeFi activity. The zkSync ecosystem token (ZK) launched in 2024 amid some controversy around the airdrop distribution.
StarkNet
Built by StarkWare, StarkNet uses STARKs — a specific type of zero-knowledge proof. Uses its own smart contract language (Cairo) rather than EVM compatibility, which is both a strength (more optimization potential) and a friction point (different tooling). Powers major applications like dYdX’s early version and Immutable X for gaming/NFTs.
Polygon zkEVM
Polygon pivoted hard toward ZK technology. The zkEVM is their flagship, offering EVM compatibility with ZK proof security. Polygon has the advantage of brand recognition and strong enterprise relationships — several major brands have launched on Polygon chains.
Scroll
A newer ZK rollup that emphasizes being “natively EVM-compatible” at the bytecode level, making developer migration easier. Backed by strong technical research and growing in adoption.
What Makes L2s Different From Each Other — And Why It Matters
When evaluating a Layer 2, these are the dimensions that actually matter:
Security model. Is it a true L2 that inherits Ethereum security, or a sidechain with its own validators? Check whether the chain posts proofs or fraud challenges to Ethereum L1.
Withdrawal delay. Optimistic rollups have a 7-day withdrawal window to L1. ZK rollups finalize much faster. For most use cases this doesn’t matter much, but if you ever need to move funds urgently, it can.
EVM compatibility. Can you run standard Ethereum smart contracts without modification? Most major L2s now offer this, but some (like StarkNet) require rewriting in a different language.
Decentralization of the sequencer. Currently, almost all L2s use a single centralized sequencer — a server that determines transaction ordering. This is a meaningful centralization risk that the major L2 teams are working to address through decentralized sequencer roadmaps.
Bridge security. Moving assets between L1 and L2, or between different L2s, requires bridges. Bridges have historically been the single most exploited component in all of DeFi. The mechanics of bridging and which bridges are audited matters.
Ecosystem. Where are the apps you actually want to use? Liquidity follows apps and users — a technically superior chain with nothing to do on it is just vapor.
The L2 and L3 Future
Ethereum’s roadmap has always assumed L2s do the heavy lifting. The base layer handles settlement and data availability; L2s handle execution. More recent developments push this further: L3s (Layer 3s) are application-specific chains that settle to L2s, which settle to L1. This enables even more specialized performance — a gaming chain, a social chain, a derivatives chain — each optimized for its use case.
Arbitrum Orbit, OP Stack, and zkSync’s ZK Stack all support custom L3 deployment. We’re moving toward a modular blockchain world where the chain you’re using is increasingly invisible infrastructure, not a thing you consciously choose.
That’s either very exciting or very confusing, depending on your perspective. Probably both.
Practical Guidance for L2 Users
- For DeFi: Arbitrum and Base have the deepest liquidity and most mature protocols. Start there.
- For lower fees on stable transactions: zkSync Era and Scroll often have competitive fees and fast finality.
- For bridging: Use the official bridge for your target chain where possible. Third-party bridges offer speed and cross-chain routing, but add smart contract risk. Always double-check bridge addresses.
- Keep some ETH on mainnet. Some actions still require L1 — DAO governance, certain NFT contracts, high-value DeFi interactions where trust matters most.
- Watch for airdrop eligibility. Many L2s have distributed tokens to early users. Several (StarkNet, Scroll, Linea) have either recently launched tokens or are expected to. Using these chains genuinely — not just bridging and sitting — tends to be what gets rewarded.
Frequently Asked Questions
Is my money safe on a Layer 2? Safer than on a sidechain, because true L2s inherit Ethereum’s security. That said, L2s introduce their own risks: smart contract bugs in the rollup contracts, centralized sequencers, and bridge vulnerabilities. The major established L2s (Arbitrum, Optimism, Base) have substantial security audits and bug bounties, but nothing in crypto is zero-risk.
Do I need different wallets for different L2s? No. The same Ethereum wallet (MetaMask, Rabby, etc.) works across all EVM-compatible L2s. You just add the network to your wallet and bridge funds over. Your address is the same on every EVM chain.
Why are L2 fees so much cheaper? Because they batch thousands of transactions together and share the cost of posting to Ethereum. Instead of one transaction paying for its own L1 slot, it pays a tiny fraction of a shared batch cost.
What’s the difference between Arbitrum and Optimism? Both are optimistic rollups with similar security models. Arbitrum has historically had more DeFi liquidity. Optimism is the basis for the OP Stack superchain and has Base as a major ecosystem partner. In practice, both are solid choices.
Are ZK rollups better than optimistic rollups? In theory, ZK rollups have superior properties: faster finality and stronger cryptographic guarantees. In practice, optimistic rollups got to market first and have more mature ecosystems today. ZK rollups are catching up quickly, and most roadmaps point toward ZK as the endgame.
Can I stake on Layer 2s? Yes. Many L2s have native DeFi protocols including staking and yield products. See our staking guide for how that works in practice.