Ethereum's Layer 2 networks are hitting a pricing wall, and Offchain Labs co-founder Edward Felten argues that responsive fee models, not patchwork upgrades, are the real path to scalability.
Ethereum rollups have a pricing problem. Despite processing millions of transactions off the mainnet and cutting costs dramatically compared to 2021 gas wars, the fee mechanisms powering most Layer 2 networks remain crude. Edward Felten, co-founder of Offchain Labs, the company behind Arbitrum, believes the industry has been building on top of a pricing model that was never designed for how these networks actually operate. His argument is straightforward: if Layer 2s want to scale to millions of users, they need to stop copying Ethereum mainnet fee logic and start pricing transactions based on real-time network conditions.
The core issue revolves around how rollups charge users. Most Layer 2 networks currently rely on some version of EIP-1559, the fee mechanism introduced on Ethereum mainnet in August 2021. That system uses a base fee that adjusts based on block congestion, aiming to make gas fees more predictable. It works reasonably well on the mainnet. But Layer 2s operate differently. They batch transactions and post them to Ethereum in discrete intervals, meaning their congestion patterns and capacity constraints look nothing like the mainnet. Applying the same pricing logic to a fundamentally different architecture creates friction that users end up paying for.
Felten's proposal, as CoinTelegraph recently reported, centers on what he calls responsive pricing. Rather than relying on the slow, block-by-block fee adjustments inherited from Ethereum mainnet, responsive pricing would allow a rollup to adjust fees based on actual, current demand for batch space and compute resources. Think of it as surge pricing that actually reflects the road you are driving on right now, rather than one updated every few minutes regardless of whether traffic has cleared.
Arbitrum is already testing this model. The network has been quietly experimenting with alternative fee structures that decouple Layer 2 transaction costs from the mainnet's pricing rhythm. The early results suggest that users benefit from lower and more predictable fees during periods of uneven demand, which is most of the time. This matters because Layer 2 growth depends heavily on user experience. A decentralized finance trader or a blockchain gamer does not care about the technical elegance of a fee mechanism. They care whether a transaction costs two cents or twenty cents, and whether that cost is stable enough to plan around.
Why This Matters Beyond Arbitrum
The broader competitive landscape makes this conversation urgent. Layer 2 networks are multiplying fast. Beyond Arbitrum and its sibling Arbitrum Nova, there is Optimism, Base, zkSync, StarkNet, Polygon zkEVM, Linea, Scroll, and dozens more in various stages of deployment. All of them are chasing the same prize: becoming the default execution layer for Ethereum's ecosystem. Fees are a key battleground. In 2023 and early 2024, several of these networks offered near-zero fees as an incentive to attract developers and users. Those subsidies are not sustainable. Eventually, every rollup has to charge something close to the real cost of sequencing transactions and posting data to the mainnet, and the pricing model they choose will determine whether users stay or migrate to cheaper alternatives, including non-Ethereum chains like Solana.
Responsive pricing also has implications for sequencer design. The sequencer is the component that orders and executes transactions on a rollup before they are settled on Ethereum. Today, most major rollups run centralized sequencers, which is a frequent criticism from decentralization advocates. A more sophisticated pricing model could make decentralized sequencer networks economically viable, because participants would have clearer signals about the cost of inclusion at any given moment. This connects to ongoing work by shared sequencer projects like Espresso Systems and Astria, which are building infrastructure to allow multiple rollups to share sequencing resources.
For investors and entrepreneurs watching this space, the takeaway is practical. Network activity on Layer 2s has been climbing steadily throughout 2024, with combined daily transactions on major rollups frequently exceeding Ethereum mainnet volumes. But growth at the current scale is still modest compared to what these networks aspire to handle. If fees remain volatile or unpredictably high during peak periods, the risk is that users simply never develop the habits needed to bring blockchain applications into the mainstream. Responsive pricing will not solve every scalability challenge. Data availability costs, proof generation times, and cross-chain interoperability all remain open problems. But it addresses one of the most tangible pain points that users encounter today, and it does so without waiting for the next major Ethereum upgrade. Expect more rollups to test similar models through the rest of 2024 as the competition for users intensifies.
