Why Ethereum needs Layer 2s: the rollup-centric roadmap explained

Ethereum can only settle about 15 transactions a second, and making the base chain handle far more would push out the everyday computers that keep it decentralized. Layer 2 networks solve this by doing the heavy work off-chain and settling back to Ethereum. Ethereum's plan is to stay the secure base while L2s scale on top.

What problem do Layer 2s solve?

Ethereum's main chain is secure and decentralized, but it is slow by design: it processes only about 15 transactions per second, because every full node must run and store every transaction. When demand is high, that limit means competition for space, and fees rise. Layer 2s exist to add capacity without weakening the base.

The obvious fix would be to make the main chain bigger, letting each block hold far more transactions. The problem is what that does to decentralization. The more work each block demands, the more powerful and expensive a computer has to be to keep up. Push that too far and only a few large operators could afford to run full nodes, which is exactly the centralization Ethereum was built to avoid.

So Ethereum took the other path: keep the base chain lean and secure, and move the heavy lifting up a level. Instead of one chain doing everything for everyone, the main chain becomes a secure settlement layer, and separate networks built on top handle the high volume of day-to-day activity. Those networks are the Layer 2s.

This is not a stopgap; it is the stated plan. Ethereum's roadmap is often called rollup-centric, meaning the base chain is meant to specialize in security and settlement while rollups carry the everyday transactions (source: Ethereum scaling documentation). Understanding that split is the key to the rest of this guide. (How the base chain itself works is covered in our companion guide on how Ethereum works.)

Here is the everyday version of the problem. When Ethereum gets busy, competition for those few transactions per second pushes fees up, sometimes to several dollars or more for a single action. That is bearable for large transfers, but it makes small, frequent ones painful. L2s exist so that a five-dollar swap does not cost five dollars in fees.

What is a Layer 2, exactly?

A Layer 2, or L2, is a separate network that runs transactions off the main Ethereum chain, then posts the results back so Ethereum secures them (source: Ethereum Layer 2 documentation). You get fast, cheap transactions on the L2, while Ethereum stays the final record. The main chain is Layer 1; anything built on top to scale it is a Layer 2.

The split of duties is the important part. The L2 handles execution, meaning it runs the actual transactions and keeps its own running tally of balances. Ethereum handles two things the L2 sends back. The first is settlement, the final agreement on what happened. The second is data availability, a public record of the transaction data so anyone can check the L2's work. The L2 does the busy work; Ethereum is the court of record.

A quick way to picture a rollup: instead of a thousand people each mailing a separate letter to Ethereum, the rollup collects all thousand, puts them in one envelope, and mails that. Everyone shares the postage. The saving comes from sharing the expensive step, which is posting to Ethereum.

This is what people mean when they say L2s inherit Ethereum's security, though that phrase needs care, and a later section unpacks it. For now, the simple version is that an L2 does not ask you to trust a new, separate blockchain in the way a wholly independent network would. It leans on Ethereum for the part that matters most: making sure the final record cannot be quietly rewritten.

The type of L2 that matters most today is the rollup. A rollup bundles many transactions together, runs them off-chain, and posts a compressed summary plus the underlying data back to Ethereum. The name comes from that bundling: many transactions are rolled up into one batch, so the cost of using Ethereum is shared across all of them instead of paid one by one.

Optimistic vs ZK rollups: what is the difference?

Rollups come in two main kinds, differing in how they prove their transactions are honest. Optimistic rollups assume every batch is valid and allow a window to challenge a bad one. ZK rollups, short for zero-knowledge, attach a cryptographic proof of validity to every batch. The trade-off is mostly about how fast you can move funds back to Ethereum.

An optimistic rollup posts its batches and treats them as correct unless someone proves otherwise, using what is called a fraud proof. To leave room for a challenge, withdrawals back to Ethereum through the network's own bridge carry a waiting period, usually around seven days. Arbitrum and Optimism are the best-known optimistic rollups, and both are highly compatible with existing Ethereum code, which made them easy for apps to move onto.

A ZK rollup instead posts a validity proof with each batch, a small piece of cryptography that proves the transactions were processed correctly without revealing every detail. Because the proof settles the question immediately, there is no week-long challenge window, so funds can return to Ethereum quickly. The trade-off has been that this technology is harder to build and was, for a while, less compatible with existing code, though that gap has narrowed.

For a beginner, the difference matters less than it sounds. The three biggest networks today are all optimistic rollups, and for everyday use they feel the same as any other: fast and cheap. The main practical effect of the type is how long a withdrawal through the network's own bridge takes. And even that rarely matters when you use an exchange, as the next point explains.

Here is how the two compare, alongside the main chain:

From Blofin's operational perspective, the roughly seven-day challenge window on optimistic rollups is a protocol property, not a wait you face when you use an exchange. When you withdraw ETH on Arbitrum from Blofin, you do not wait seven days. That window is the time the rollup needs internally to let anyone challenge a bad batch before it settles to Ethereum. Our withdrawal flow manages the L2 positioning on our side and holds that exposure for you, so the protocol-level wait does not become your wait.

What does the 2026 Layer 2 landscape look like?

By 2026, a handful of L2s hold most of the activity, and three of them dominate. Arbitrum is the largest by a wide margin, with Base and Optimism close behind, and a second tier of ZK-based networks growing underneath them.

A tracker called L2BEAT, which monitors the value held across rollups, counted dozens of live L2s holding tens of billions of dollars combined in April 2026 (source: L2BEAT), a picture echoed in industry coverage of the 2026 market (source: 2026 Layer 2 outlook). Arbitrum has led on value held for some time, accounting for a large share of all L2 activity. Base, built by Coinbase, has grown quickly into second place. Optimism, which shares much of its underlying design with Base, sits just behind. Below them, ZK-based networks like zkSync, Starknet, Scroll, and Linea make up a growing second tier. The list shifts as networks rise, merge, or fade, so treat any ranking as a snapshot rather than a fixed order.

The table below shows the rough picture in 2026, drawn from public L2 dashboards (source: DefiLlama). The figures move with the market, so they are a guide to scale, not exact balances:

For a beginner, the practical takeaway is simpler than the rankings. The big three optimistic networks are where most everyday users, apps, and liquidity already are, which makes them the easiest places to start. Whether to scale Ethereum this way or move activity to a separate Layer 1 chain entirely is a different debate, covered from the investor's angle in our guide on Layer 1 alternatives.

Why are Layer 2 fees so much cheaper?

Layer 2 fees are low because a rollup spreads the cost of using Ethereum across a whole batch, and because a 2024 upgrade made the data part of that cost far cheaper. Instead of each transaction paying full price for Ethereum space, hundreds share one posting. The result is fees often a fraction of a cent.

The upgrade that changed the math is called EIP-4844, which shipped in March 2024. It introduced a new, cheaper way for rollups to post their data to Ethereum, using temporary attachments called blobs rather than competing for regular block space. That single change cut typical L2 fees by roughly 80 to 90 percent, bringing many L2 transactions down to well under a cent (source: EIP-4844 specification). It is the main reason L2s went from "cheaper" to "almost free for small transactions."

There is an honest caveat worth adding for 2026. Ethereum's main chain is also cheap most of the time now, because so much activity has moved to L2s that mainnet is rarely congested. So on a quiet day, the gap between a mainnet transaction and an L2 transaction can be small in absolute terms. The L2 advantage is largest exactly when you need it most: when the network is busy and mainnet fees spike, L2 fees barely move. How base-chain fees work in the first place is covered in our guide on Ethereum gas, and a closer fee-by-fee comparison lives in a dedicated guide on L2 fees versus mainnet.

One more thing surprises beginners: the fee does not depend on how much you send. Moving 1,000 dollars costs the same gas as moving 10 dollars, because you pay for the computation, not a percentage of the amount. So a single transfer might cost a dollar or more on a busy mainnet day, and a fraction of a cent on Arbitrum or Base, whether the transfer is large or small. That flat, tiny cost is what makes L2s practical for everyday spending.

What does "inheriting Ethereum's security" actually mean?

It means a rollup relies on Ethereum for the record of what happened and the final say on it, not for running the transactions themselves. This is a real and valuable guarantee, but it is narrower than "the L2 is exactly as safe as Ethereum," and the difference is worth understanding before you trust an L2 with large sums.

The part an L2 genuinely inherits is settlement and data availability. The transaction data is posted to Ethereum, so anyone can independently check the L2's work. And because Ethereum holds the final record, no one can quietly rewrite the L2's history once it settles. That is a strong foundation. It is what separates a real rollup from a simple separate chain that just asks you to trust it.

The part that is not automatically inherited is execution and the machinery around it. The L2 still runs its own software, usually relies on a single operator called a sequencer to order transactions, and depends on its proof system working as intended. A bug in the L2's own code, or a sequencer that goes offline, is an L2 problem, not something Ethereum steps in to fix.

A simple example makes this concrete. If an L2's sequencer goes offline for a few hours, your transactions can stall even while Ethereum itself runs fine. Your funds are not lost, because the record still settles to Ethereum, but the convenience pauses until the L2 recovers. That is the kind of risk that lives with the L2, not with the base chain. The detail of how each network handles those responsibilities, and how much trust each one still requires, has its own dedicated guide on Layer 2 security.

What does it mean if you use a Layer 2?

For everyday use, an L2 feels like a cheaper, faster Ethereum, with two things worth knowing: you may need to bridge funds onto it, and tokens come in native or bridged forms. Bridging means moving assets from Ethereum or an exchange onto the L2. Once there, you transact for a fraction of the cost, using the same kind of wallet.

The native-versus-bridged distinction trips people up. Some tokens, like certain versions of USDC, are issued directly on an L2 by the original issuer, the native form. Others arrive by being locked on Ethereum and represented on the L2 by a bridge, the bridged form. They usually trade close to each other, but they are not always interchangeable, and sending the wrong one to the wrong place is a common and avoidable mistake. When in doubt, check which version a destination expects before you send.

From Blofin's operational perspective, our four supported networks for ETH and ERC-20 transfers are Ethereum mainnet, Arbitrum, Optimism, and Base, three of which are L2s. That gives us a live read on where people actually move funds. On a typical 2026 day, well over half of our outbound ETH and ERC-20 withdrawals go to an L2 rather than to mainnet. The visible reason is the fee saving. The quieter reason is that many of the apps users are withdrawing to now live mainly on L2s, including much of the trading that happens on decentralized exchanges. Picking which network to withdraw on is, in practice, a question of where you are going next, and the deposit and withdrawal mechanics are covered in our operational guide.

If you are new to this, start simple. Pick one established network, move a small amount across, and run a test transaction before anything larger. Confirm you are sending the token version the destination expects, native or bridged. None of this is hard once you have done it once, and the low fees mean a test run costs almost nothing.

Frequently asked questions

Are Layer 2s safe to use?

The leading L2s are widely used and hold large sums, but "safe" depends on the specific network. A rollup inherits Ethereum's settlement and data guarantees, which is a real strength, yet it still runs its own software and usually relies on a single sequencer to order transactions. That means an L2 carries some risks Ethereum itself does not. For everyday amounts on a major, established L2, the risk is modest; for large sums, it is worth understanding each network's setup, which our Layer 2 security guide covers.

Can I keep funds on a Layer 2 long-term?

Yes, and many people do, since major L2s have run reliably for years. The trade-off is that you are trusting the L2's software and operator in addition to Ethereum, so the same caution applies as with any protocol: prefer established networks and keep amounts in proportion to your comfort with the extra risk. Your assets remain yours on the L2; you can bridge them back to Ethereum whenever you choose.

What is the difference between optimistic and ZK rollups?

Optimistic rollups assume batches are valid and allow a challenge window, usually around seven days, before native-bridge withdrawals to Ethereum finalize. ZK rollups attach a cryptographic validity proof to each batch, so there is no challenge window and funds can return to Ethereum quickly. Optimistic rollups were easier to build and adopt early; ZK rollups offer faster finality and are catching up on compatibility.

Why are Layer 2 fees so much cheaper than mainnet?

Because a rollup shares the cost of posting to Ethereum across a whole batch of transactions, and because the 2024 EIP-4844 upgrade made that data far cheaper to post using blobs. Together these push many L2 fees below a cent. Note that in 2026, mainnet is often cheap too, so the L2 advantage is largest when the network is busy and mainnet fees spike.

Will Layer 2s replace Ethereum mainnet?

No; the plan is the opposite. Ethereum's rollup-centric roadmap keeps the main chain as the secure settlement and data layer that L2s depend on, while L2s handle most everyday transactions. The two are designed to work together: L2s cannot exist without a base chain to settle to, and Ethereum scales by supporting them rather than by replacing itself. In that sense, more L2 usage makes the base chain more important, not less, because every L2 ultimately settles back to it.

Researched and written by the Blofin Academy editorial team with AI-assisted drafting. Primary sources include the ethereum.org Layer 2 documentation, the EIP-4844 specification, and L2BEAT for Layer 2 data. Network standings and fee figures are snapshots as of April 2026 and were independently verified against cited dashboards; the Layer 2 landscape changes frequently.

This article is for informational purposes only and does not constitute financial advice, investment guidance, or a recommendation to buy, sell, or hold any digital asset. Layer 2 networks involve risks beyond those of Ethereum mainnet, including sequencer and smart-contract risk, and bridging assets carries its own risks. Conduct your own research and consult qualified professionals before making decisions. Blofin Academy content reflects the state of public information at time of publication; protocols and ecosystem data change frequently.