What is a Layer 2 (rollup)?
A Layer 2 processes transactions cheaply and settles them on a main chain like Ethereum. Here is how rollups work, the two main types, and the trade-offs.
Ethereum is secure, decentralised, and slow. When everyone wants in at once, fees climb and the whole thing feels expensive. A Layer 2, or L2, is the workaround: a separate network that does the actual transacting, then hands a compressed summary back to Ethereum to anchor it.
Most of them are rollups, which is a literal name. They roll thousands of transactions up into one batch and post that single batch to the base layer, so everyone in the batch splits the cost of settling it. The result is transactions for cents instead of dollars, while still leaning on the base chain for security.
By the mid-2020s the majority of Ethereum activity had moved to L2s, so understanding them is no longer optional. This page covers the problem they solve, how a rollup actually works step by step, the two main designs, and what to watch for as a user.
The problem Layer 2s solve
A base chain can only fit so many transactions in a block, and blocks come at a fixed pace. When a popular launch or a market panic hits, everyone bids for the same scarce block space, and the price of getting in spikes. At the worst moments people have paid more in fees than the thing they were trying to buy was worth.
You cannot simply make the blocks huge without asking every participant to store and verify more data, which pushes out the ordinary people who keep the network decentralised. That tension, between capacity and decentralisation, is the scaling problem. L2s sidestep it by moving the work off the main chain rather than enlarging the main chain.
How a rollup actually works
You send your transaction to the rollup instead of to Ethereum. The rollup has its own sequencer, a piece of software that orders incoming transactions and executes them quickly and cheaply on the L2.
Periodically the rollup takes a large batch of these transactions, compresses them, and posts the batch to Ethereum along with the new state that results. Thousands of users share the single base-layer fee for that post, which is why the per-transaction cost collapses.
The base chain does not re-run every transaction. Instead the rollup provides a guarantee that the batch is valid, and how it provides that guarantee is exactly what separates the two main designs.
Optimistic vs zero-knowledge rollups
Optimistic rollups assume every batch is valid and publish it, but allow a challenge window, usually about a week, during which anyone can submit a fraud proof showing a transaction was invalid. They are simpler to build, but that challenge window means withdrawing funds back to the main chain can be slow unless you use a third-party bridge to jump the queue.
Zero-knowledge rollups instead post a cryptographic validity proof with each batch, mathematically demonstrating the batch is correct without re-executing it. Settlement can be much faster because there is nothing to dispute, but the technology is harder to build and the proofs are computationally heavy.
Both aim for the same outcome: cheaper transactions that still settle on a trusted base layer. The choice is a trade-off between engineering simplicity and withdrawal speed, and the industry has been steadily moving toward the proof-based approach as the tooling matures.
The trade-offs, and what to watch
Cheap fees are not a moat on their own. What actually keeps users on an L2 is liquidity, the applications built there, and trust in how it settles back to the base chain. A ghost town with low fees is still a ghost town.
Decentralisation varies a lot. Many L2s still rely on a single sequencer to order transactions, which means one operator could in theory censor or reorder your transaction, and could go down and halt the chain. Whether that operator can be forced to behave, and how easily you can exit to the base layer without its cooperation, is the real security question.
As a user, two things matter most: the security of the specific bridge you use to move funds on and off, which is where most large L2-related losses have happened, and how decentralised the sequencer is. Cheap and fast is the easy part. Trust-minimised is the part still being built.
Frequently asked questions
Is a Layer 2 safe?
A well-designed rollup inherits much of its base chain's security because it settles transactions there. The real risks lie in the operator that orders transactions and, above all, in the specific bridge you use to move funds between the layers.
Why are Layer 2 fees so much lower?
An L2 batches thousands of transactions together and posts a single compressed record to the main chain, so the base-layer cost is shared across everyone in the batch instead of paid in full by each user.
What is the difference between Layer 1 and Layer 2?
Layer 1 is the base chain that provides security and final settlement, such as Ethereum or Bitcoin. Layer 2 is a faster, cheaper network built on top that processes transactions itself and settles them back down to Layer 1.
How long does it take to withdraw from a Layer 2?
On a zero-knowledge rollup it can be quick. On an optimistic rollup the native withdrawal can take around a week because of the fraud-proof challenge window, though third-party bridges offer faster exits for a fee and some added trust.
Are Layer 2 tokens the same as the chain?
Not necessarily. Many L2s issue a governance or fee token, but the network can run without you holding it, and the token's price is separate from how well the chain works. Treat the token and the technology as two different questions.
Ethereum is secure, decentralised, and slow. When everyone wants in at once, fees climb and the whole thing feels expensive. A Layer 2, or L2, is the workaround: a separate network that does the actual transacting, then hands a compressed summary back to Ethereum to anchor it.