Ethereum L2s at 24,000 TPS now. I think we finally reached escape velocity for L2 scaling. L2 TPS now entering the exponential age. Here's what's new: 1. zk appchains are here. Lighter is a new type of rollup thats post state diffs and zk proofs, requires minimal Ethereum DA. They're a perp appchain with product market fit. Can scale to hyperliquid levels - millions TPS. 2. Rollups to gigagas. First generation rollups (e.g. Base, Arbitrum) are scaling hard. Base to 10,000 TPS in first half 2026. Ethereum's Fusaka makes this possible by Blob Parameter Only (BPO) upgrades that 5x Ethereum DA in 2026. 3. Blurred borders. zksync's upcoming atlas upgrade enables 15k TPS rollups with interoperability that will feel like you're still on the L1. The line between Ethereum L1 and L2s will blur - more zk magic. We'll have enough blockspace supply - where will demand come from? - new zk appchains following the Lighter path - coinbase & robinhood bringing all their users onchain - L1s converting to L2s - institutions building chain strategies - killer apps on breakout rollups we can't anticipate "But these are all separate chains, you're not scaling Ethereum." I have some sympathy for this take. It is true, L2s today are not scaling Ethereum L1 blockspace. But L2s are scaling Ethereum property rights. On a properly constructed L2 user property cannot be stolen or censored - assets are rug resistant. L2 assets are secured by Ethereum - a massive upgrade compared to altL1s and centralized systems. So what's not scaling with L2s? L2s aren't scaling Ethereum L1 composability and ETH economics. New L2 constructions are nibbling at solutions - zkSync Atlas taps into L1 liquidity - Linea gives 20% of fees to ETH holders - future L2 constructions like Based and Native rollups may solve this completely. But L2s aren't the only answer. In parallel, Ethereum is scaling L1 blockspace. With current engineering there's a path to increase L1 TPS from 20 TPS to 100 TPS in 2026. With @drakefjustin's zk powered Lean Ethereum roadmap we start progress toward 10,000 TPS for Ethereum L1. L2s moving fast, many experiments in parallel. L1 moving slower, prioritizing decentralization & security. Ethereum is scaling.

I didn't even mention L2 hyperscalers like @megaeth. MegaETH posts DA to Eigen, not Ethereum - so not a full Ethereum rollup. But this construction allows MegaETH to scale basically without limit and still use ETH for economic security and Ethereum for settlement. If hyperscalers get demand TPS as a metric won't even make sense in that world. There is no TPS limit.