Since EIP-4844 shipped with the Dencun upgrade in March 2024, the conversation around Layer 2 costs has moved past raw transaction fees. The real competition now plays out in blob fee markets - a mechanism most users never see but one that increasingly determines which rollups remain viable at scale.
Blobs are temporary data packets that rollups use to post transaction data to Ethereum’s consensus layer. They expire after roughly 18 days, which keeps Ethereum’s storage footprint manageable. More importantly, they operate on their own fee market, separate from the standard execution gas market. When blob demand spikes - typically during periods of high L2 activity - blob fees rise independently, and rollups either eat the cost or pass it to users.
That separation was intentional. But it has produced an asymmetry that favors rollups with deeper treasuries or more aggressive fee-smoothing mechanisms. Arbitrum and Base, which run among the highest transaction volumes of any Ethereum L2, can absorb temporary blob fee spikes more easily than smaller rollups that operate closer to the margin. The fee market, in other words, is functioning exactly as designed - and compressing the competitive field in the process.
The Compression Effect
In the months following Dencun, blob fees stayed close to their minimum target of 1 wei per blob for extended periods. That honeymoon is conditional. Ethereum’s target is 3 blobs per block, with a maximum of 6. Once sustained demand regularly pushes toward the ceiling, the EIP-1559-style pricing mechanism kicks in and fees escalate quickly.
Pectra, which activated on Ethereum mainnet in May 2025, raised the blob count limits - up to 9 blobs per block maximum - which bought more headroom. But the broader trajectory is clear: as L2 adoption compounds, blob space will tighten again, and the rollups with the most efficient data compression will hold a structural cost advantage.
ZK rollups, particularly those using validity proofs, post meaningfully less data per transaction than optimistic rollups. That difference is negligible during low-demand periods. It becomes significant when blob fees climb.
What Actually Changes for Users
Nothing visibly dramatic. Fees on Base or Arbitrum remain low enough that most users won’t notice quarterly blob fee fluctuations. The pressure shows up at the protocol layer - in how rollup teams architect data posting strategies, batch sizes, and whether they pursue EigenDA or Celestia as alternative data availability layers to hedge against Ethereum blob costs entirely.
That last option carries its own tradeoffs around security assumptions, and most major rollups have declined to fully decouple from Ethereum DA. For now, the blob fee market remains the hidden variable quietly sorting which L2s can scale without compromising their cost proposition.