📜 "Who is paying rent to Ethereum?" 》 L2 Rollups 2014-2025 Fees at a Quick View Since writing the white paper in 2014, Ethereum has planned to turn the "settlement layer" into an ATM. It was later Layer-2 that really paid money to L1. The following data from 2024-07-19 looks back for 12 months, allowing you to intuitively understand who burns the most gas. 💥 Core conclusions 2015-2019: Plasma and sidechains test the waters, and almost no one really puts money to L1. 2020-2021: Loopring/zkRollup settles in batches for the first time, with an annual fee of less than a few hundred thousand dollars. 2022-2023: Optimistic & zk exploded, and L2s "paid rent" one after another. 2024-2025: The emergence of "super landlords" - Base and Arbitrum have been able to influence the Ethereum calldata market. 💸 L1 costs handed over to Ethereum in the last 12 months (USD) 🔹 Base — $4.91 M 🔹 Arbitrum One — $2.65 M 🔹 Polygon zkEVM — $1.25 M 🔹 Scroll — $0.81 M 🔹 Optimism Mainnet — $0.66 M 🔹 zkSync Era — $0.30 M 🔹 Starknet — $0.23 M 🔹 Kroma — $0.14 M 🔹 Linea — $0.11 M 🧮 A total of ≈ $11 M / year - less than a fraction of Ethereum's fees burned in the same cycle, but the growth rate is visible to the naked eye. ⚙️ Why is there so much difference? 1️⃣ Settlement frequency: Base/Arbitrum updates quickly, with large data volumes, and is naturally expensive. 2️⃣ Compression technology: zk tethered (zkSync, Starknet) itself has less calldata. 3️⃣ "Fee subsidy": Some new chains (Linea, Kroma) used L1 fees as marketing costs in the early days, deliberately lowering the batch frequency. 🔮 Looking ahead to 2026 and beyond After EIP-4844, blob quotes will continue to lower per capita rent, but the total expenditure will most likely only increase - because more people use it. The competition between L2s will continue: whoever can do the best compression, batching, and proof will be able to move more transactions to the chain. 🙏 Thank you for reading, give me ❤️ a &🔁, encourage me to continue to produce hard-core dry goods!