What Are Ethereum Layer 2s? Arbitrum, Base, and Optimism
Arbitrum. Base. Optimism. They process huge volumes, they’re where a lot of crypto activity actually happens, and they’re usually discussed as if they’re competitors to Ethereum — faster, cheaper chains that won where Ethereum lost. That framing is backwards. These networks didn’t beat Ethereum; they are Ethereum’s scaling strategy, the one Ethereum’s own developers deliberately chose. The interesting part is that the strategy worked mechanically while producing an outcome the original blueprint never predicted.
The problem rollups solve
Ethereum is a world computer: anyone can deploy a program and the whole network executes it faithfully. The catch is that “the whole network executes it” is expensive — every node runs every transaction, which is what makes Ethereum secure and decentralized but also slow and, at peak congestion, brutally costly. There were stretches in 2024 when a single token swap could run fifty dollars or more in fees. A world computer that costs fifty dollars a click isn’t one most people can use.
It’s the same core bind Bitcoin hit, the one Lightning was built for: a base layer that’s secure but can’t handle volume. Ethereum’s answer had to be different, because Ethereum isn’t just moving payments — it’s running programs. The path its developers picked is the rollup-centric roadmap: instead of making the main chain process more, push activity to a second layer and let the main chain do what it does best — provide security and final settlement. That second layer is a Layer 2.
How a rollup actually works
A rollup takes hundreds or thousands of transactions, executes them on its own faster network, bundles them, and posts them back to Ethereum. The part that’s easy to get wrong: it doesn’t post a summary or a fingerprint of the result — it posts the actual transaction data, compressed and batched. That’s literally what “rollup” means. The reason matters. If only a summary went to Ethereum, you’d have to trust the operator that the summary was honest. By posting the underlying data, anyone can independently reconstruct the Layer 2’s state and verify it. The data living on Ethereum is what makes the security real instead of trust-based — and it’s why a Layer 2 isn’t a separate chain competing for security. It borrows Ethereum’s. The L2 gives you speed and low fees; Ethereum gives the L2 its trustworthiness. Layers of one system, exactly as the name says.
There are two main types, differing in how they prove validity. Optimistic rollups — Arbitrum, Base, Optimism — assume transactions are valid by default and allow a challenge window, typically seven days, during which anyone can prove fraud; if no one does, the batch is final. ZK rollups — zkSync, Starknet, Linea — instead attach a cryptographic proof to every batch; Ethereum checks the math, and if it holds, the batch is final with no waiting period. Optimistic rollups are older and easier to build on, which is why they hold the deepest ecosystems; ZK rollups offer faster finality and stronger guarantees but are harder to build, and some still lean on centralized components they’re working to shed.
The upgrade that changed everything
In March 2024, Ethereum shipped an upgrade called Dencun, whose key piece created “blobs” — a separate, cheaper data lane built specifically for rollups. Before Dencun, rollups competed with regular Ethereum transactions for the same block space; after, they had their own lane. Layer 2 fees dropped by roughly ninety percent, from a dollar or more down to a few cents or a fraction of one. The world computer got cheap.
The twist most explanations skip
The original vision was pluralistic — a thousand rollups serving different communities, all settling to Ethereum. That isn’t what happened. What happened is brutal consolidation. There are now more than seventy active rollups, but just two of them hold roughly three-quarters of all the money on Layer 2s, and three process around ninety percent of transactions. Most of the rest are what the industry openly calls “zombie chains” — technically alive, activity evaporating. One prominent L2, Blast, saw deposits collapse after a disappointing token launch.
The irony is that Dencun helped cause the consolidation. Once launching a rollup became trivial and fees fell to near-zero everywhere, fees stopped being something you could compete on — so the competition moved to distribution. Base won enormous share not because its tech is dramatically better but because Coinbase operates it and plugs it into Coinbase’s massive user base. Arbitrum won by becoming the deepest pool of DeFi liquidity and compounding the lead. Optimism is the most telling case: its own chain trails both, but its underlying OP Stack is the framework other chains — including Base — are built on. It arguably built the most widely used L2 technology in crypto; it just isn’t winning on the chain carrying its name. That gap, between building the best technology and capturing the most users, is the whole story.
Did the strategy work?
Mechanically, yes. Ethereum needed to scale and it did — fees collapsed, the world computer became usable, activity moved to Layer 2s exactly as the roadmap intended, and all of it still settles back to Ethereum. What it didn’t produce was the decentralized thousand-rollup ecosystem the early vision imagined; it produced a handful of dominant winners chosen largely by distribution muscle. Layer 2s didn’t abandon Ethereum and didn’t lose to it. They extended it — just into a shape the original blueprint didn’t quite predict.
Not investment advice. WTH Crypto is editorial commentary, not financial guidance.
Related
WTHIsWhat Is a Covered-Call ETF, Really?
A covered-call ETF advertises double-digit yield on Bitcoin — but Bitcoin pays nothing. That income comes from selling away your upside. Here's the real trade-off.
WTHIsWhat Is the Ethereum Foundation, Really?
The Ethereum Foundation says it doesn't run Ethereum — then shrank itself to prove it. What it actually is, what it funds, and who pays when it steps back.
WTHIsWhat Is a Perp, Really?
A perp is a futures contract that never expires. Here's how the funding rate keeps it tied to spot — and why leverage makes it crypto's most dangerous instrument.