L2s are raking it in

L2s are raking it in—what about Ethereum?

Foresight News特邀专栏作者
2026-09-08 11:30
This article is about 10909 words, reading the full article takes about 16 minutes
Amid the commercial frenzy, it’s time to re-examine the Rollup narrative and the CROPS development roadmap.
AI Summary
Expand

• Core Thesis: While leading enterprise-grade L2s are posting impressive commercial results, most remain stuck at Stage 0/1 and have yet to truly achieve “Stage 2” final settlement, contributing minimally to Ethereum in fees. The commercial success of L2s cannot be equated with Ethereum’s success—it is essentially compliance arbitrage and brand borrowing. Ethereum’s R&D focus should shift toward CROPS (censorship resistance, privacy, security)—areas that are inherently non-replicable.
• Key Elements:
  1. Only 4 chains have truly reached Stage 2 standards, with combined TVL under $700,000; Base holds $14.42 billion in TVL and Arbitrum One holds $12.6 billion, both remaining at Stage 1.
  2. Base processed 292 million user operations, paying Ethereum only approximately $8,800 in fees over 30 days (about $290 per day); Robinhood chain generates $3–4 million in daily revenue but pays Ethereum only a few hundred dollars per day.
  3. Stage 2 requires permissionless fraud proofs, a 30-day exit window, and restricted Security Council powers—yet enterprise operators (licensed brokerages, banks, public companies) must retain the ability to freeze, censor, and execute emergency fixes, making the set of Stage 2-compatible business models an empty set.
  4. In April 2026, the Arbitrum Security Council passed an atomic transaction upgrading the Inbox contract, moving 30,766 ETH (approximately $71 million) suspected to have been stolen by the Lazarus Group to a governance wallet—highlighting operators’ incentive to retain intervention powers.
  5. Optimism co-founder Mark Tyneway admitted that “Stage 2 deviates from users’ true demands”—the real users are chain operators, whose need for features that minimize legal risk runs counter to Stage 2 requirements.
  6. Ethereum’s value to L2s is realized only through three voluntary channels—data availability rent, ETH monetary premium, and brand licensing—none of which have mandatory enforcement; only true settlement (Stage 2) can establish structural binding, but enterprise operators subjectively refuse it.

Original Author: _gabrielShapir0

Original Translation: Saoirse, Foresight News

Supported by extensive public data and regulatory logic, this article offers a systematic reflection on the assumptions of the “rollup-centric” roadmap of the past few years. Its core contribution is revealing a frequently avoided question: How should Ethereum position itself when L2 operators choose to remain in an intervenable state based on their own interests? Regardless of whether readers agree with his conclusions, the misaligned incentives and boundary conditions proposed in the article provide a verifiable framework for discussion. Attached at the end is a differing perspective from Ethereum community member Ryan Berckmans for readers’ reference.

The Paradox: L2s are a Huge Commercial Success, Yet Fundamentally a Failure for Ethereum

Robinhood Chain is the most prominent example, sparking renewed debate: whether, and how, Ethereum should adapt to the demands of L2s. This global top-tier stock trading platform serves nearly 28 million users across 38 countries, turning its core business into tokenized stocks deployed on a self-custody wallet supporting over 120 countries. On its first day, it integrated with Uniswap for 24/7 token trading, offered developers the GonzoFi experimental environment, and allowed users to deposit tokens into lending pools.

Other L2s are also growing rapidly. Base, within a publicly listed company, holds $14.42 billion in Total Value Locked (TVL); Arbitrum One holds $12.6 billion. With the advent of the Orbit and OP development suites, launching a public chain has been simplified into a procurement decision.

This is no accident: operating an L2 is itself a lucrative business. Just from this chain, Robinhood generates $3-4 million in daily on-chain revenue, not to mention potential earnings from its commercial partnerships related to its trust token wrapping business (which underpins the tokenized stock offering).

Faced with such booming commercial results, one might ask again: where exactly does the value of these L2s lie? And what does Ethereum get out of it?

L2BEAT categorizes L2s into various types. The most important category is Rollups, which Robinhood Chain, Base, Arbitrum, and Optimism are classified as. Rollups are divided into Stage 0, Stage 1, and Stage 2. Only Stage 2 Rollups truly delegate final settlement to Ethereum smart contracts for execution. Stage 0 is mostly fully controlled by the sequencer operator; Stage 1 is generally controlled by a Security Council holding multi-signature management authority. Despite these controls being masked by various superficial designs (e.g., some chains claiming to have implemented “forced transaction inclusion”), the operator retains absolute control over the entire chain; Ethereum can only help identify fraud, but cannot enforce sanctions.

Several years have passed since Vitalik published “A Rollup-Centric Ethereum Roadmap” in October 2020 and proposed the staging criteria in November 2022. Currently, only four chains carry the Stage 2 label: Facet with $661k TVL, Honeypot v2 with $1k TVL, Aztec with less than $1k TVL, and Ethscriptions with no stats. The combined TVL of these four chains is less than $700k. In contrast, Base and Arbitrum One have TVLs of $14.42 billion and $12.6 billion respectively, yet both remain at Stage 1. Robinhood Chain has a TVL of $2.9 billion, but its contracts can be instantly upgraded by a 7-of-8 multisig account, leaving users with no exit window if a malicious upgrade occurs; its fraud proof system only accepts submissions from two whitelisted entities; and the operator can censor any transaction without delay, including those supposedly “forced inclusions” (L2BEAT project page, 2026-09-07).

There is no reason to believe the status quo is temporary. In January 2026, Optimism co-founder Mark Tyneway wrote: “Stage 2 deviates from what users actually want, and everyone is afraid of community backlash, so no one dares to speak the truth.” He pointed out that the real “users” are the chain operators, who want features that minimize their own legal risks, which often contradict Stage 2 requirements. Imagine explaining in court why you couldn’t pause the bridge contract while users’ funds were being drained, or why you couldn’t prevent North Korea-linked entities from holding assets on your platform. The prevailing industry view is that almost all enterprise-operated top L2s (like Robinhood, Base) will never evolve to Stage 2.

If the original intent of Rollups was to progress to Stage 2, then this L2 vision has failed. Even if L2s are commercially thriving, this statement sounds discordant. Ethereum finds itself in an awkward position: it has found product-market fit, but the customers it’s serving are using Ethereum only “incidentally” and “symbolically.” L2s treat Ethereum as an optional backup settlement layer, not truly utilizing it. So, where do we go from here?

The rest of this article will elaborate on the structural mismatch between Ethereum’s interests and the L2 market. Although L2s bring some benefits to Ethereum and ETH, these are insufficient to make L2s the central focus of Ethereum’s R&D efforts. Seeing through the dazzling commercial success of L2s and resisting the temptation to chase trends is difficult but necessary.

Four Ways L2s Give Back to Ethereum: Why Only “True Settlement Rent” Qualifies as a

L2s create value for Ethereum through four channels:

1. Data Availability Rent: L2s purchase Blob data space. Blobs are essentially fungible commodities with alternatives in the market: external data availability layers or operator-run committees. Switching to an alternative only requires configuration changes, not asset migration, making switching costs nearly zero. More problematically, Ethereum has publicly committed to continuous scaling, which is itself the goal of its scaling roadmap. A commodity with continuously expanding supply and buyers who can switch anytime cannot generate scarcity rent. Data confirms this. In the 30-day period ending September 7, 2026, Base, the largest L2 by TVL, processed 292 million user operations but paid only approximately $8,800 to Ethereum for data, proofs, and state updates – roughly $290 per day. Arbitrum One paid only about $2,700 over the same period. By this calculation, Robinhood Chain generates $3-4 million in daily revenue but pays Ethereum only a few hundred dollars per day. Furthermore, for chains where users cannot freely exit, Blobs don’t even fulfill their intended design value. The purpose of posting data on-chain is to allow anyone to rebuild the chain state and exit; operators who only upload hashes to L1 can achieve the same reorg resistance at a lower cost. No matter how Blob pricing is optimized, reality remains unchanged: with Ethereum promising ample supply, it cannot command scarcity prices.

2. ETH as Gas, Capturing Monetary Premium: This path is entirely voluntary with no mandatory enforcement. Operators can denominate fees in stablecoins, deploy custom gas tokens, or even subsidize fees down to zero, and the protocol cannot prevent it. So-called “ETH-L2s” are mostly marketing concepts with a weak “ETH as money” effect, and this soft effect cannot generate stable revenue.

Multiple factors continue to weaken this path. First, enterprise operators have financial statements denominated in USD; the ETH received by their sequencers is working capital, not long-term reserves. Coinbase has been criticized for selling the ETH fees generated by Base but refuses to disclose the flow of funds. Public financial reports show: from Q1-Q2 2026, Coinbase’s ETH holdings for investment changed from 150,193 to 150,279 tokens. Even though Base was the largest L2 during this period generating fees, there’s no sign of significant ETH accumulation. Second, L2 teams face continuous pressure to issue their own tokens: investors and employees need liquid assets to realize returns; the product itself needs an incentive mechanism controlled by the operator. Third, proprietary tokens need designed use cases: commercially, tokens without utility have no demand; regulatory-wise, tokens with actual functions are easier to argue are not securities. The most natural function for an L2 token is paying sequencer fees, and both the Orbit and OP development suites natively support custom gas tokens. Each funding round further diminishes the expected returns on this path.

3. Settlement Rights: Ethereum holds a right that operators cannot revoke, exercised by L2 users. Out of the four paths, only this one constitutes a property right, not a simple buyer-seller relationship. Only through this path can Ethereum act as a franchisor rather than just a commodity supplier.

4. Brand Licensing: This refers to the “secured by Ethereum” label. Ethereum has no trademark, no licensing body; this license is granted free and irrevocably to everyone. For projects that already control distribution channels, community consensus cannot constrain them. Even if Ethereum wanted to charge for this, the protocol’s pricing mechanism is uniform and indiscriminate, and there’s no business development department specifically catering to companies like Robinhood. A “license” that cannot be revoked and has no pricing is essentially a gift, not a source of revenue.

Data availability rent and ETH network effects are loose and unenforceable; brand licensing is entirely non-commercializable. Only when Ethereum performs true final settlement can structural, endogenous interest alignment be built, creating a moat. But true settlement capability only activates at Stage 2, which enterprise L2 operators subjectively refuse. This creates an intractable misalignment of interests.

Without True Settlement, L2s and Ethereum Can Only Be a Mutually Beneficial Relationship – But Nothing More

Ethereum does receive a small amount of genuine data availability revenue; massive numbers of users on these chains hold and use ETH, generating some monetary premium; developers, capital, and attention that would otherwise flow to other chains are attracted; and there’s a reputational dividend: a brokerage launching an “Ethereum L2” signals to regulators that Ethereum is infrastructure, not contraband.

However, all these benefits are voluntary, revocable at any time, unpriced, and loosely coupled. That’s what it means for L2s to treat Ethereum as an “optionality.” At best, it amplifies the influence of the ETH asset, but not the Ethereum chain itself. Options hold enormous value for their holders; but the issuers of options can only look at the premiums they received – and in this case, the premium is zero.

There’s another side to this dividend, which is the core incentive for large general-purpose L2s to adopt the L2 shell: regulatory arbitrage. Strip away the marketing, and what remains is a centralized ledger operated by a corporate entity, with pause switches, transaction filtering, and contract upgrade capabilities. If you ran this system bare, you’d be a custodian or a money transmitter, clearly visible to regulators. But attach an Ethereum bridge and call it a Rollup, and you can leverage social consensus to claim it’s part of Ethereum, sufficiently decentralized that intermediary regulations seemingly no longer apply – while the operator retains all the power of an intermediary.

A common counterargument: large L2s do L2 to save costs. The question is: what costs? It’s either the cost of achieving decentralization, or the compliance cost of *not* being decentralized. Either way, it’s essentially cost savings on the regulatory front. Mark Tyneway’s post voiced what everyone knew implicitly. He later admitted that many institutional L2s have a “vibe mismatch with reality” – they are clearly not Ethereum, lack credible neutrality, and aren’t trustless; operators can switch protocol versions at will. The label carries a legal function, not a cryptoeconomic one. Operators are simply responding to incentives, which is understandable.

Why Stage 2 is the True Critical Dividing Line

L2BEAT defines three hard requirements for Stage 2: ①Permissionless fraud proof systems; ②A user exit window of at least 30 days in case of malicious upgrades; ③The Security Council can only handle errors that are verifiably judged on-chain. Notably, this standard doesn’t require decentralized sequencers. Stage 2 constrains upgrade authority and user escape hatches, but doesn’t interfere with transaction ordering rights or sequencer fee profits.

Commercially: Stage 2 creates an asymmetry. An L2 chain itself can leave Ethereum, but it can no longer hold user assets hostage, because users hold exit rights that operators cannot erase through upgrades. Only with this asymmetry can Ethereum potentially raise its “rent” without directly losing customers. Without this, Ethereum is like a month-to-month supplier whose clients can switch to alternatives at any time.

Legally (the author’s professional domain): Referencing FinCEN’s 2019 crypto asset guidance and case law under 18 U.S.C. §1960, the core criterion is whether the entity has complete independent control over the assets being transmitted. As long as an operator can filter transactions, freeze balances, and unilaterally upgrade bridge contracts without an exit window – regardless of how polished the documentation is – it effectively holds complete control. Conversely, if permissionless proofs, a 30-day exit window, and Security Council powers limited to on-chain verifiable bugs are implemented, operators have a strong basis to argue they are merely publishing software, not conducting money transmission. Therefore, Stage 2 architecture is the strongest technical defense against being classified as a “centralized fund services provider wearing a Rollup costume.” This also explains why enterprises voluntarily abandoning Stage 2 is worth deep consideration, not just disappointment.

Governance-wise: L2s that don’t support forced transaction inclusion do not inherit Ethereum’s censorship resistance. They only inherit data availability and state roots – capabilities that can expose operator misbehavior but cannot reverse its consequences. Censorship occurs at the sequencer level, where Ethereum’s properties end.

If Stage 2 became the industry default, “secured by Ethereum” would represent a genuine legal relationship. The saying “even if Coinbase fails, user assets on Base remain safe” would no longer be just a nice aspiration. The reality is that Stage 2 hasn’t become mainstream, with almost no major projects adopting it.

Why Stage 2 Can’t Be Implemented: Community Pressure Cannot Solve the Fundamental Problem

The real cost of Stage 2 for operators is the loss of operational discretion. A common misconception is that Stage 2 would strip sequencers of their fee profits. It doesn’t. Stage 2 limits upgrades and user escape mechanisms, but doesn’t interfere with the monopoly on transaction ordering. Even running a fully centralized, highly profitable sequencer can satisfy all Stage 2 conditions while collecting every bit of fee revenue.

What’s truly lost is operational discretion. Regulated financial institutions must be able to execute court orders to freeze accounts, block addresses per OFAC sanctions lists, halt and reverse fraudulent transfers, urgently patch vulnerabilities before funds are stolen, and provide a specific responsible party when regulators inquire – not throw up their hands. Stage 2 architecture is designed to eliminate all these capabilities – which is precisely its core purpose.

Let’s do a reductio ad absurdum: as long as operators retain the ability to freeze accounts, they don’t meet Stage 2 definitions; if they give up freezing capabilities, corporate legal counsel can’t sign off, and the enterprise becomes unable to fulfill legal obligations. Therefore, for operators holding brokerage licenses, money transmitter licenses, banking charters, or publicly listed status, the set of business models compatible with Stage 2 is an empty set – not small, but completely non-existent.

From another angle, imagine lobbying Elon Musk to have X Money run as a Stage 2 rollup, instead of a Stage 0 chain masquerading as an L2, a permissioned L1 with a small validator set (the Hyperliquid model), or just a regular database. How would you convince him? He trusts himself to manage user assets well; minimal trust mechanisms hold no appeal for him; all options are sufficient for his needs. Under the current legal framework, as long as you package the product as an L2 or a permissioned L1 with a few “independent validators,” even a fully centralized chain faces no fines or jail time; running an unlicensed database bare, however, carries legal risks. So building a blockchain is necessary, but choosing a genuine Stage

ETH
Layer 2
Arbitrum
Welcome to Join Odaily Official Community
Subscription Group：
https://t.me/Odaily_News
Chat Group：
https://t.me/Odaily_GoldenApe
Official Account：
https://twitter.com/OdailyChina
Chat Group：
https://t.me/Odaily_CryptoPunk
Recommended Articles

YZi Labs retakes control, BNB treasury CEA Industries undervalued

Opening bell imminent: Memory chips rally across the board, Bloom Energy added to S&P 500, Dell added to S&P 100, double boost on the rise—what other news deserves attention

币股风向标丨Strategy年内融资209亿美元，位列美股第四；Strive跻身上市公司BTC持仓规模前五，年底有望跃升至第二（9月8日）

Search
Article Catalog
Author Library
Foresight News
Bitcoin’s “Bounce Has Ended,” Formally Entering Late Bear Market Phase?
Ethereum Q1 2026 Review: On-Chain Activity Hits All-Time High, Tokenized Assets Lead the Industry
Robinhood Chain Ecosystem Overview: Which Projects Are Worth Getting In Early On?
Article Ranking
Daily
Weekly

No time to sit idly on-chain, which Robinhood Chain assets offer a second chance to get in?
Robinhood链上Meme币股的3种打开方式
64-year-old Man Tops Fomo Leaderboard, “Diamond Hands” Bonkguy’s Meme Legend
Bitcoin Up 25% in August, But the Real Battle Lies at $81,000
Where Should You Provide LP for Stock Tokens? A Quick Look at Five Emerging AMMs on Robinhood Chain
Download Odaily App
Let Some People Understand Web3.0 First
IOSAndroid
About UsContact UsJoin UsFriendly LinksAdvertising Cooperation
User AgreementPrivacy PolicyDisclaimerOdaily Brand Media Kit | Official Logo & Visual Guidelines
Beijing Ruike Culture Media Co., Ltd.
京ICP备 2026027382号京公网安备11010502060861号

Source: www.odaily.news