Enterprise-operated Ethereum Layer 2 networks are processing hundreds of millions of transactions and collecting substantial user fees while sending comparatively little direct revenue to Ethereum, according to L2BEAT data for the 30 days ended Sept. 7. The gap comes as the largest networks retain upgrade and censorship controls that fall short of the conditions L2BEAT uses for its highest decentralization category, Stage 2.
Base processed 292 million user operations during the period while paying about $8,800 to Ethereum for data availability, proofs and state updates, L2BEAT data shows. That works out to roughly $290 a day. Arbitrum One paid about $2,700 over the same 30-day window.
The figures place the debate over Ethereum’s rollup-centric roadmap beyond raw transaction growth. Major Layer 2s can offer low-cost activity and generate operator revenue, but their present structures leave Ethereum with limited fee capture and users dependent on companies, multisig councils and emergency governance processes for some core protections.
Most value remains outside Stage 2 systems
L2BEAT lists Base with $14.42 billion in total value locked and Arbitrum One with $12.6 billion, placing both among Ethereum’s largest Layer 2 networks. Each is categorized as Stage 1, a label for systems that have moved beyond an early rollup design but retain material dependencies on privileged parties.
The contrast with Stage 2 is sharp. L2BEAT listed only four networks in that category: Facet, with about $661,000 in TVL; Honeypot v2, with about $1,000; Aztec, with less than $1,000; and Ethscriptions, for which the platform showed no TVL data. Combined, the group held under $700,000.
Stage 2 is not simply a measure of transaction speed, fees or locked assets. Under L2BEAT’s framework, a network needs a permissionless fraud-proof system, a minimum 30-day user exit window if a malicious upgrade occurs, and a security council whose powers are restricted to responding to faults verifiable onchain. The framework does not require a decentralized sequencer, the party that orders Layer 2 transactions.
Those requirements place the largest corporate and enterprise-operated systems in a difficult position. Their operators often maintain rapid-upgrade powers, transaction-filtering abilities and emergency procedures designed for compliance, operational incidents or recovery from exploits. Those powers can be useful in a crisis, but they also prevent the system from meeting a standard built around a user’s ability to exit without relying on the operator.
Upgrade controls remain a central trade-off
Robinhood Chain, listed by L2BEAT with $2.9 billion in TVL, demonstrates the governance issue. The network can be upgraded immediately by a 7-of-8 multisig and does not provide users with an exit window following a potentially malicious upgrade, according to its L2BEAT project page dated Sept. 7.
The same page says Robinhood Chain’s fraud-proof system accepts submissions from only two allowlisted entities. Its operator can also censor transactions without delay, including transactions presented as forced inclusion requests. Forced inclusion is intended to give users a route to submit a transaction to Ethereum when a Layer 2 sequencer refuses to process it; an operator-controlled path weakens that protection.
Arbitrum’s April intervention offered a recent example of how extraordinary powers can be used. Its security council approved an atomic transaction, backed by nine of 12 members, that upgraded the Inbox contract and added a temporary permission function. The change enabled a cross-chain message that transferred 30,766 ETH, valued at about $71 million at the time, into a governance wallet. Public reporting linked the funds to a suspected Lazarus Group theft. The permission was later removed.
The episode showed the practical appeal of concentrated authority during a suspected theft: it can move far faster than a fully permissionless dispute process. It also showed why a system with such powers remains outside a model in which users retain guaranteed exit rights regardless of a council’s decisions.
Ethereum’s fee capture is limited
The Layer 2 model gives Ethereum several potential ways to benefit from activity occurring above the base chain, but only settlement-related spending is structurally required by a rollup’s design. Networks may choose to publish transaction data through Ethereum blobs, use ETH-denominated gas, or lean on Ethereum branding and liquidity. Those choices can be changed through business or technical decisions.
Data availability is particularly flexible. A system can alter its configuration to publish data elsewhere without requiring users to migrate their assets. It can also post hashes to Ethereum rather than the full transaction data. Hash posting offers a lower-cost record with reorganization resistance, but it may prevent independent parties from reconstructing the Layer 2 state needed to verify balances or execute exits.
Robinhood Chain’s reported operating data illustrates the difference between application-layer revenue and direct Ethereum payments. The chain is described as generating $3 million to $4 million in daily onchain revenue while sending Ethereum only a few hundred dollars per day. The supplied fee data does not establish a complete profit calculation, since operators have infrastructure, incentive and other costs, but it does show that high Layer 2 activity does not automatically translate into proportional L1 revenue.
L1 development focuses on enforcement-free security
Ethereum’s base-layer roadmap has increasingly emphasized functions that do not depend on Layer 2 operators voluntarily adopting stricter governance structures. A February 2026 “Harden the L1” workflow identified censorship resistance, privacy and security as areas for development, alongside efforts to measure whether those protections are delivered in practice.
EIP-7805, known as FOCIL, is among the proposals aimed at improving censorship resistance. Blob capacity expansion, statelessness research and higher gas-limit targets address the base layer’s ability to process and validate more activity without relying solely on Layer 2 scaling.
Ethereum raised its L1 gas limit to 60 million in 2025 from 30 million, its first major increase since 2021, with internal scaling work targeting capacity above 100 million. The Pectra upgrade in May 2025 doubled blob throughput, while Fusaka’s PeerDAS upgrade in December was described as increasing theoretical blob capacity eightfold.
Lower L1 costs could also change the practical balance for users who want to hold or move assets under Ethereum’s direct consensus rules. Average gas prices reached about 0.16 gwei in April 2026, according to the figures cited in the dataset, reducing the cost difference between certain L1 actions and activity on a managed Layer 2.
Privacy tools are developing along a similar track. Kohaku released an SDK allowing wallets to integrate RAILGUN-style private transactions using separate addresses for each decentralized application. Aztec’s Ignition mainnet has been presented as operating with the full Stage 2 feature set, though its TVL remains tiny beside Base and Arbitrum One.
A meaningful test for the sector would come if a top-five Layer 2 reached Stage 2 with more than $1 billion in TVL and retained its users through the transition. Until then, Ethereum’s largest scaling networks remain defined by a trade-off: low-cost, high-throughput systems with corporate-grade intervention tools, alongside a base layer designed to reduce the need for those tools.
Curious how these dynamics affect traders? Explore our Layer 2 blockchain guide for deeper insight into scaling, security, and fee flows.
Disclaimer: The content on this page is provided for general informational purposes only and does not represent the views or financial advice of Toobit. We make no guarantees regarding the accuracy or completeness of this information and shall not be held liable for any errors, omissions, or outcomes resulting from its use. Investing in digital assets involves risk; users should independently evaluate their financial situation and the risks involved. For further details, please consult our Terms of Service and Risk Disclosure.
