Introduction
A stark academic warning is challenging the economic foundations of Ethereum’s primary scaling strategy. New research from zkSecurity, Prooflab, and Imperial College London reveals that the fee models used by major layer-2 rollup networks are fundamentally flawed, creating systemic mispricing that exposes billions of dollars in secured assets to spam attacks and user exploitation. The study, which benchmarks networks like Arbitrum (ARB), Optimism (OP), and Polygon zkEVM (MATIC), concludes that simplified pricing mechanisms are inadvertently subsidizing malicious activity while overcharging honest users, posing a significant threat to the ecosystem’s stability.
Key Points
- Five major Ethereum rollups use divergent fee models that collapse computation, data, and proof costs into simplified formulas vulnerable to exploitation
- Mispricing enables attackers to spam networks with cheap transactions while honest users overpay, creating denial-of-service risks for billions in secured assets
- Researchers recommend multidimensional pricing with dynamic adjustment and cost transparency to align fees with actual resource consumption
The Illusion of Simple Fees
At the heart of Ethereum’s scaling roadmap, rollups like Arbitrum, Optimism, and zkSync Era operate by bundling transactions and processing them off-chain before settling a compressed proof on the main Ethereum (ETH) blockchain. This process is designed to reduce costs and increase throughput. However, the new study, titled “Unaligned Incentives: Pricing Attacks Against Blockchain Rollups,” identifies a critical oversimplification. To function, a rollup incurs three distinct and independently variable costs: computation (executing transactions), data availability (posting data back to the base chain), and proof verification (settling the batch). The researchers found that most rollup networks collapse these three costs into a single, simplistic fee formula.
This one-size-fits-all approach fails to reflect the actual, fluctuating resource consumption of different transaction types. For instance, a simple token transfer on Solana (SOL) or Bitcoin (BTC) might be computationally cheap but require significant data posting. By not accounting for these dimensions separately, the current fee models create widespread mispricing. “The problem lies not in faulty code, but in economic design choices that shape incentives,” the paper states, highlighting that the issue is foundational to the current rollup economic model rather than a mere software bug.
Exploitable Models and Systemic Risks
The benchmark analysis of five major networks—Polygon zkEVM, zkSync Era, Scroll, Optimism, and Arbitrum—revealed a patchwork of divergent and potentially exploitable fee rules. Some networks fix fees when a transaction is submitted, while others determine the final cost only when a batch is sealed. Certain systems, like those on Optimism and Arbitrum, even incorporate refund mechanisms if the actual cost is lower than initially charged. While seemingly user-friendly, the study warns that these mechanisms can be gamed. Attackers can submit a flood of transactions, reclaim a portion of the fees, and still consume substantial network resources at a subsidized cost.
This mispricing creates a direct path to denial-of-service (DoS) attacks. An attacker can exploit underpriced transactions to spam the network, clogging it with low-cost, high-volume transactions. This degrades performance for all users and can artificially inflate costs for honest participants. The stakes are high; as the paper notes, rollups today secure tens of billions of dollars in digital assets, making them high-value targets. The systemic risk is not theoretical—it is an economic vulnerability baked into the fee structures of the very networks Ethereum is counting on for its future growth.
A Call for Multidimensional Pricing
As a solution, the researchers from Imperial College London and their partners urge a shift towards “multidimensional” fee mechanisms. This would involve separately pricing computation, data posting, and proving costs to accurately align user fees with the actual resources their transactions consume. Such a model would naturally disincentivize spam, as attackers would have to pay the true cost of the network resources they consume. It would also lead to fairer and more predictable costs for users.
The paper suggests practical tools to achieve this, including dynamic fee adjustment based on real-time demand, partial batching of transactions, and greater transparency around the breakdown of cost components. Some rollup teams are already experimenting with adaptive fee curves, but the study emphasizes that industry-wide standards are lacking. This recommendation is particularly urgent with the advent of zero-knowledge virtual machines (zkVMs), which, while offering stronger security guarantees, introduce proving costs that can be highly variable. A fee model that cannot handle this variability risks breaking down under network stress, leading to inconsistent fees and unreliable service for exchanges, wallets, and end-users.
The study’s final message for developers and investors is to look beyond headline-grabbing metrics like transactions per second or nominally low fees. The true measure of a rollup’s robustness and long-term viability lies in the sophistication of its economic design. As Ethereum’s roadmap becomes increasingly rollup-centric, ensuring these layer-2 networks are economically secure is not just an optimization—it is a necessity for safeguarding the entire ecosystem.
📎 Related coverage from: decrypt.co