# Layer 2 - Faster, Cheaper, More Complicated By [Lista DAO](https://blog.lista.org) ยท 2026-07-10 --- Tokyo Station at six p.m. on a Tuesday is, I think, one of the most efficient places on earth. We met by the central pillar in the Marunouchi concourse - Moolah, somehow, had figured out a way to wait for us at a public transit hub without anybody noticing an otter, and I had stopped questioning this - and from there she walked us up to one of the upper-level platforms where the shinkansen tracks ran out toward the north. The platform was full. Briefcases. Suitcases. Salarymen on phones. A school group being herded by a teacher who looked like she had been doing this her entire life. Below us, on a parallel set of tracks, a local commuter line was disgorging a crowd onto a packed platform. "This," Moolah said, "is what Ethereum is up against." "Crowds?" It took me a second to wonder how she could deliver the question in the center of a bustling station. "Throughput. Look around. The commuter line - local trains, every few minutes, every station, packed full of people. That's Ethereum. That's the base layer. It works. It serves everyone. It is also, as you have discovered when you tried to send twenty dollars during a busy moment, sometimes too crowded to be useful." She gestured up to the shinkansen platform we were on. "Now look here. Different track, same station. Express service. Far fewer trains, but each one is much faster, and each carries hundreds of people from one major hub to another in a fraction of the time. The local trains keep running for the small stops. The express handles the long-haul demand. Both feed into the same station system. Neither replaces the other. They handle different parts of the same network's job." "And this is -" "This is the Layer 2 architecture. _Layers_ on top of Ethereum that handle most of the traffic, while still being part of the same overall system. They're called _rollups_ in the jargon, and they have become - in the last few years - where most actual DeFi activity now lives. If you sent a transaction on Ethereum directly today, you might pay one dollar. If you sent the same transaction on a Layer 2, you might pay a tiny fraction of one cent. Same kind of activity. Different track." "Why didn't they just make Ethereum faster?" "Because of a trade-off they take very seriously. To make Ethereum faster, you'd have to make the system smaller - fewer nodes verifying everything, less redundancy, less decentralization. The Ethereum community decided, some years ago, that they would rather keep the base layer slow, expensive, and maximally decentralized, and let _other_ layers handle scale. The base is the vault. The Layer 2s are the lobby." A train pulled in. The doors opened. A precise pulse of passengers exited, another pulse boarded, the doors closed, the train left. The whole exchange took perhaps ninety seconds. I watched it happen and felt, briefly, that I was understanding something I hadn't quite understood before. "Tell me how a rollup actually works," I said. Moolah settled into a different posture. "Imagine you and ninety-nine other people all want to send a letter from Tokyo to Osaka. You could each buy a ticket and ride the train yourselves to deliver the letters. Or - one person could collect all hundred letters, ride the train once, and deliver the whole bag. The bag goes once. The letters all arrive. The cost is split a hundred ways." "So the rollup is the bag." "The rollup is the bag. A Layer 2 collects thousands of transactions, batches them together, and posts a single compressed summary to Ethereum. The summary is small, but it contains enough information that anyone can reconstruct exactly what happened in the rollup. Ethereum doesn't process the transactions individually. It just records the batch. The transactions happened on the rollup; their proof is filed on the base chain." "And the security?" "That is the clever part." She looked up at me. "The rollup posts the _data_ of every transaction back to Ethereum. So even if the rollup itself disappears tomorrow - if the company running it shuts down, or the servers all fail - every user can reconstruct their balance from the data on Ethereum and withdraw their funds. The rollup is fast, cheap, and convenient. But its _security_ is inherited from the base chain. The express train belongs to the express company. The track is the country's track. If the express company fails, you can still walk along the track." I wrote that down. _Rollup = batch transactions, post the proof to Ethereum. Speed from the rollup, security from the base chain._ "There's a complication," Moolah said. "Two main types of rollups, with different ways of _proving_ the batch is honest. They're called _optimistic_ and _zk_. The names tell you a great deal about how they work." "Go on." "An _optimistic_ rollup assumes everything is fine, by default. It posts the batch, says _here's what happened_, and gives anyone in the world a window - a week, usually - to look at the batch and submit a fraud proof if they spot something wrong. If nobody objects, the batch is treated as final. The optimism is structural - the system trusts the rollup, but with a tripwire, and anyone can pull it. Examples: Optimism, Arbitrum, Base. Big, popular, currently the most-used." "And the zk version?" "A _zk_ rollup uses cryptography instead of trust. Every batch is accompanied by a _zero-knowledge proof_ - a mathematical certificate that proves, without revealing the underlying transactions, that everything was processed correctly. The proof is small, fast to verify, and impossible to fake without solving problems mathematicians believe to be unsolvable. There is no week-long window. There is no need for honest watchdogs. The math itself attests. Examples: zkSync, Starknet, Linea, Scroll. Newer, harder to build, increasingly common." "So one trusts the network to police itself, the other proves itself with math." "That is the entire distinction in one sentence. Write it down." I did. "The trade-offs?" "Optimistic rollups are simpler to build, more flexible, and easier to make compatible with existing Ethereum software. But withdrawing funds back to Ethereum takes a week, because of the fraud-proof window. Zk rollups have nearly instant withdrawals - once the proof is verified, the transaction is final - but they are mathematically heavy, take more computation to generate the proofs, and have historically been slower to add features. The two camps argue with each other constantly. Both will probably exist, side by side, for a long time." "And which should I use?" "Whichever has the protocols you want, and the best fees today. Both inherit Ethereum's security. Both let you withdraw to the base chain - slowly, in the optimistic case; quickly, in the zk case. Both are dramatically cheaper than using Ethereum directly. The truth is that for almost everything you'll do as a user, the difference between them does not affect your daily experience. The difference matters for builders, regulators, and the long-term shape of the industry. As a user, you care about which one has a working app for what you want to do." A different train pulled in. This one was a local, on the lower platform - packed shoulder-to-shoulder, the windows fogged. We watched the local crowd pour out, then pour back in, and the train slid away into the tunnels. "There is one thing I want you to remember," Moolah said. "The Layer 2s now hold tens of billions of dollars of activity. Most DeFi users now interact with them, not with Ethereum directly. The transition has happened, mostly without fanfare, in the last few years. Ethereum is increasingly the _settlement layer_ - the place where the largest, most important, most secure things happen - and the Layer 2s are the _user layer_, where everything else lives. The two-layer system, base plus express, is now the actual architecture of how this whole industry works. If you want to understand how DeFi feels in practice, you need to understand that fewer and fewer people are using the base chain anymore. The shinkansen is the system. The local train is the foundation." "And the bridges between them?" "Are the same bridges I warned you about last week. Move with caution. Use the official ones. Don't leave money on chains you don't plan to use. The connective tissue is dangerous. The connective tissue is also the entire point." We rode the escalators back down. The concourse was, if anything, more crowded than when we'd arrived. Somewhere in the noise, a station announcement chimed three soft notes and announced a delay on a line I did not take. "Next week?" Dev asked. "Next week," Moolah said, "Priya is going to lose money. Not very much. Don't worry. But I want her to feel something I cannot make her feel any other way. Bring small amounts to the kissaten on Sunday. We're going to do a trade together." I wrote in the notebook, in capital letters: _NEXT WEEK PRIYA LOSES MONEY ON PURPOSE._ Then, smaller, underneath: _good?_ Moolah saw it. She gave me the slow, slightly amused look that I now understood meant _trust me_. I trusted her. * * * _Next: MEV - The Invisible Hand in Your Pocket_ --- *Originally published on [Lista DAO](https://blog.lista.org/layer-2-faster-cheaper-more-complicated)*