Layer 2

Layer 2 is now the default place to do anything on Ethereum that is not a multi-million-dollar transaction. After Dencun (March 2024) brought blob space, after Pectra (early 2025) reduced settlement overhead further, and after a long deflationary war on fees across Arbitrum, Optimism, and Base, the median user transaction on a top L2 now costs a fraction of a cent.

We are particularly interested in the divergence between user metrics and economic metrics. The user metrics look strong everywhere. The economic metrics tell a much more uneven story.

What you will find here

• L2 transaction & activity data — daily transactions and active addresses across the top 8 L2s.
• Sequencer economics — sequencer revenue net of L1 settlement cost.
• L1 burn contribution — how much each L2 actually pays mainnet.
• OP Stack vs ZK Stack vs vanilla rollups — the architecture debate.
• Native vs bridged liquidity — which L2s have native stablecoin issuance.

The L2 landscape today

• Base — the high-velocity consumer L2.
• Arbitrum — the deepest DeFi liquidity.
• Optimism — the OP Stack steward.
• Polygon — PoS chain plus zkEVM.
• Starknet — Cairo VM, ZK-native.
• zkSync — ZK Stack.
• Linea — Consensys-led zkEVM.
• Mantle — modular L2 with native staking yield.

What we think actually matters

• Liquidity depth.
• Sequencer credibility.
• Stablecoin issuance.
• Application gravity.

The economic question post-Dencun and post-Pectra

Post-Dencun, L2 fees to mainnet collapsed by ~95%. The optimistic case is that L2 throughput scales to the point where aggregate mainnet revenue grows. The pessimistic case is that L2s become a value-extracting layer. We think L2 tokens should be valued on sequencer cashflows, not user metrics.

Related coverage

• Ethereum — the settlement layer the entire L2 ecosystem depends on.
• DeFi — the largest use case driving L2 activity.
• Analysis — for long-form pieces on the L2 endgame.

For methodology see /methodology/. The morning briefing tracks L2 daily transactions every Tuesday.