Did Ethereum's Rollup Strategy Enrich L2s While Impoverishing L1?

General
1

I’ve been thinking a lot about Vitalik’s February statement that the rollup-centric roadmap “no longer makes sense,” and the more I dig into the data, the more complex this picture becomes.

The Revenue Problem Nobody Wants to Talk About

Let’s start with some numbers that made me pause:

  • Solana: $1.03M per day in chain fees
  • All Ethereum L2s combined: $182K per day

That’s not a typo. A single L1 blockchain is generating 5.7× more daily fee revenue than Ethereum’s entire Layer 2 ecosystem combined.

But here’s where it gets really interesting—or troubling, depending on your perspective. Ethereum L1 revenue has collapsed from $3.8 billion in Q4 2021 to just $28.4 million in Q4 2025. That’s a 99.3% drop.

Meanwhile, Base alone generated $19 million in revenue in Q4 2025. And how much did they pay back to Ethereum L1 for security? $363,000. That’s a 1.9% passthrough rate.

What We Were Promised vs. What We Got

When the rollup-centric roadmap was introduced in 2020, the vision was clear: Ethereum L1 would become a secure, decentralized base layer, and L2s would handle execution at scale. Security comes from L1, speed comes from L2s, everyone wins.

But there were two critical assumptions that haven’t played out as expected:

1. L1 would scale more slowly than it has

The Pectra and Fusaka upgrades, combined with aggressive gas limit increases, mean that L1 fees have stayed surprisingly low. The pressure that was supposed to push users to L2s? It’s not as intense as we thought it would be.

2. L2s would decentralize faster than they have

As of 2025, only three L2s—Arbitrum, OP Mainnet, and Base—have reached Stage 1 decentralization. Most are still at Stage 0, meaning they have centralized operators with admin keys that can censor transactions or even potentially steal funds.

The Security Theater Problem

Here’s what bothers me most as an L2 engineer: We market these networks as “secured by Ethereum,” but what does that actually mean when the sequencer is centralized?

Yes, L2 state is posted to Ethereum L1. Yes, you can theoretically exit if the operator goes rogue. But the day-to-day reality is that most L2s are controlled by centralized entities that capture all the MEV, set their own fee structures, and could censor your transactions.

That’s not the decentralization we promised.

Different Perspectives on Success

Now, let me complicate this further with some counterpoints to my own argument:

Stablecoin deployment: Solana has $14B in stablecoin supply vs. $9.05B in L2 DeFi TVL. Solana wins on active usage.

Total Value Secured: Ethereum L2s hold $40.5B in TVS vs. Solana’s total ecosystem. L2s win on stored value.

These metrics tell different stories. Are L2s winning with institutions while losing with retail? Are they optimized for holding rather than trading?

So… Did We Build the Wrong Thing?

I don’t think the rollup-centric strategy was wrong. But I do think we’re at an inflection point.

Vitalik is right that L2s need to find value propositions beyond “we’re faster and cheaper than mainnet.” Because if mainnet itself gets fast and cheap enough, that value prop disappears.

The L2s that will survive are the ones that specialize:

  • Privacy-focused execution environments
  • App-specific chains optimized for gaming or social
  • Compliance-friendly environments for institutions
  • Novel VM architectures that enable new primitives

Generic “Ethereum but faster” L2s are going to struggle.

The Question I Keep Coming Back To

Did the rollup-centric roadmap enrich L2 operators while impoverishing Ethereum L1?

The data suggests yes, at least in the short term. L2s captured billions in value while passing back millions to L1.

But did L2s enable innovation, experimentation, and ecosystem growth that wouldn’t have happened otherwise? Also yes.

The real question isn’t whether L2s were a mistake. It’s whether the current economic relationship is sustainable, and whether we can fix the decentralization gap before it becomes a critical security risk.

What do you all think? Am I being too harsh on L2s, or not harsh enough?

2

This hits home for me in a really personal way.

When I first started building DeFi apps in 2022, deploying a simple ERC-20 contract to Ethereum mainnet cost me over $500. For someone who was barely scraping by in a shared apartment in San Francisco, that was… not doable. I literally couldn’t afford to be a mainnet developer.

L2s changed everything for me. Suddenly I could deploy to Optimism or Arbitrum for a few dollars. I could iterate, experiment, make mistakes without bankrupting myself. The democratization of development that L2s enabled? That’s real. That matters.

But the Economics Are Messy

You’re absolutely right that the value capture is lopsided. Base making $19M while paying $363K back to Ethereum is… yeah. That math doesn’t feel sustainable.

But here’s my question: Should we prioritize L1 economics or developer accessibility?

Because if Ethereum L1 had scaled sooner and kept fees low from the beginning, we wouldn’t have needed L2s as desperately. But it didn’t. And for developers like me who couldn’t afford $500 deploy costs, L2s were the only option.

The Decentralization Delay Worries Me Too

I’ll admit I didn’t fully understand the Stage 0 vs Stage 1 vs Stage 2 framework until last year. When you said most L2s have admin keys that could theoretically steal funds… that made my stomach drop.

I’ve been building on these platforms assuming “secured by Ethereum” meant full security inheritance. The reality that operators can censor transactions or have emergency override keys? That’s not what I thought I was signing up for.

What Should Developers Do Now?

I have a few projects launching in the next 6 months. Should I:

  • Target Ethereum L1 since fees are lower now?
  • Stick with established L2s (Arbitrum, Base, Optimism)?
  • Wait for specialized L2s that Vitalik mentioned?
  • Look at alternative L1s like Solana?

Honestly, the whole situation feels uncertain in a way it didn’t a year ago. But I guess that uncertainty is just… part of being in crypto?

I don’t think L2s were a mistake. They enabled an entire generation of developers who couldn’t afford mainnet. But I do think we need to fix the decentralization and economic alignment issues before they become critical problems.

What do other devs think? Are you changing your deployment strategies based on this discussion?

3

Let me bring a business perspective to this, because I think we’re mixing up two different questions:

  1. Did L2s extract value? Yes, absolutely.
  2. Should we be mad about it? That’s where it gets interesting.

L2s Are Businesses, Not Charities

Base generated $19M in Q4 2025. That’s not theft—that’s product-market fit.

Users chose to use Base instead of Ethereum L1. Nobody forced them. They made an economic decision that paying Base’s fees was better than paying Ethereum’s higher fees (even with recent reductions).

From a pure business standpoint, Base:

  • Identified a market need (lower fees, better UX)
  • Built a product to serve that need
  • Captured value from users who willingly chose their platform
  • Paid for the security infrastructure they used (blob fees to L1)

That’s… just how businesses work? I’m not sure what the controversy is.

The Real Question: Why Didn’t Ethereum L1 Scale Sooner?

Here’s my controversial take: Maybe Ethereum should have prioritized L1 scaling instead of outsourcing execution to L2s.

If Ethereum had:

  • Increased the gas limit more aggressively in 2021-2022
  • Implemented Pectra/Fusaka upgrades faster
  • Optimized execution client performance sooner

…then maybe we wouldn’t have needed a dozen competing L2s, each with their own token, their own community, and their own economic incentives that don’t align with L1.

Users Voted With Their Wallets

Solana is generating $1.03M/day in fees. That’s 5.7× more than all Ethereum L2s combined.

Why? Because Solana bet on monolithic scaling from the start. Users get:

  • Fast transactions
  • Cheap fees
  • One chain to manage, not a dozen L2s with bridges
  • No fragmentation, no cross-chain UX nightmare

Ethereum chose modularity and complexity. Users chose simplicity and speed. The market is giving us feedback—are we listening?

What Should Ethereum Do Now?

From a business strategy standpoint:

Option 1: Accept the L2-centric model

  • Focus L1 on being the best settlement layer possible
  • Let L2s compete and innovate
  • Accept lower L1 revenue as the price of ecosystem diversity

Option 2: Reclaim the user experience

  • Scale L1 aggressively (Glamsterdam and beyond)
  • Make L1 competitive with L2s on fees
  • Let L2s specialize or die

Right now we’re trying to do both and succeeding at neither.

My Advice to Fellow Founders

If you’re building a Web3 startup in 2026:

  • Don’t assume Ethereum L2s will always be cheaper – L1 is catching up
  • Don’t assume your L2 of choice will exist in 2 years – consolidation is coming
  • Do prioritize user experience over maximalist ideology – users don’t care about decentralization purity if the app doesn’t work
  • Do have a multi-chain strategy – Ethereum, Solana, and emerging L1s all have different strengths

The market rewards products that solve real problems, not products that perfectly align with our original vision of how things “should” work.

Sometimes the right answer is admitting the plan needs to change.

4

I need to push back on the “secured by Ethereum” narrative, because the technical reality is much less secure than the marketing suggests.

Stage 0 Decentralization Is Centralized Theater

Let me be blunt: Most L2s are centralized systems with Ethereum-flavored security marketing.

Here’s what Stage 0 actually means:

  • A single entity controls the sequencer
  • Admin keys can upgrade contracts without delay
  • The operator can censor transactions
  • In many cases, the operator can freeze or confiscate funds
  • Users have no recourse except “trust the team”

That’s not “secured by Ethereum.” That’s “secured by Coinbase” or “secured by Offchain Labs” or whoever runs the sequencer.

The Data Doesn’t Lie

As of 2025:

  • Stage 0: Most L2s (centralized operators, admin keys, no fraud proofs)
  • Stage 1: Arbitrum, OP Mainnet, Base (limited decentralization, emergency overrides)
  • Stage 2: Zero. None. Not a single L2 is fully decentralized.

We’ve been building L2s for 4+ years and we still don’t have a single example of a fully decentralized rollup in production.

The Economic Implications Are Even Worse

You mentioned Base making $19M while paying $363K to Ethereum. Let me add context:

Where does that $19M come from?

  • Sequencer fees (centralized extraction)
  • MEV capture (all goes to Base, none to Ethereum validators)
  • Priority transaction ordering (sold by centralized operator)

Ethereum stakers earn ~3-4% APR for securing the network. L2 sequencer operators earn billions while free-riding on L1 security.

This is fundamentally unsustainable.

Why would I stake 32 ETH to earn 3% when I could run an L2 sequencer and capture 1000× the revenue with a fraction of the capital?

The Security Model Is Broken

Here’s what really concerns me: L2s market themselves as inheriting Ethereum’s security, but they introduce entirely new trust assumptions:

  1. Sequencer liveness: If Base’s sequencer goes down, Base stops. Ethereum keeps running.
  2. Upgrade risk: Centralized operators can push malicious upgrades. Ethereum L1 requires social consensus.
  3. Censorship: L2 operators can censor transactions indefinitely. Ethereum L1 has credible neutrality.
  4. MEV extraction: L2 operators capture all MEV privately. Ethereum has PBS and validator competition.

The security model isn’t “Ethereum security + fast execution.” It’s “Ethereum data availability + centralized sequencer risk.”

What Needs to Happen

If L2s want to claim they’re “secured by Ethereum,” they need to:

1. Achieve Stage 2 decentralization

  • Decentralized sequencer networks (not a single operator)
  • Permissionless fraud/validity proof submission
  • Time-delayed upgrades with community oversight
  • Forced inclusion mechanisms to prevent censorship

2. Share MEV with Ethereum L1

  • Based rollups with L1 validators sequencing L2 blocks
  • MEV-share mechanisms that distribute value to ETH stakers
  • Sequencer revenue tied to blob fee payments

3. Increase value passthrough to L1

  • Higher blob fees as demand grows
  • Protocol-enforced minimum security payments
  • Alignment between L2 profit and L1 sustainability

The Uncomfortable Truth

L2s became rent-seeking sequencers, not scaling solutions.

They captured the economic upside of Ethereum’s brand and security while avoiding the costs of decentralization.

And the ecosystem celebrated this as “scaling success” instead of recognizing it as centralization with extra steps.

Vitalik’s statement that the rollup-centric roadmap “no longer makes sense” is him finally admitting what many of us have been saying for years: We chose complexity and fragmentation, and we got centralized sequencers that don’t share the values that made Ethereum valuable in the first place.

I’m optimistic we can fix this. But it requires honest assessment of where we are, not continued security theater and misleading marketing.

The first step is admitting the problem. The second step is building based rollups, enshrined PBS, and forcing L2s to either decentralize or stop claiming Ethereum’s security guarantees.

5

Brian’s technical analysis is spot-on, but I need to add the regulatory dimension that institutional clients care deeply about.

Centralized L2s Create Regulatory Ambiguity

When an L2 has a centralized operator who controls:

  • Transaction sequencing
  • Fee collection
  • User fund access (via admin keys)
  • The ability to censor or freeze transactions

…that operator starts looking a lot like a money transmitter under U.S. law.

And if they’re a money transmitter, they need:

  • Money transmitter licenses in 49+ states
  • AML/KYC compliance programs
  • Regular audits and reporting
  • Bonding and capital requirements

Most L2s aren’t structured for this. They’re operating in a regulatory gray zone, hoping that “decentralization is coming soon™” will be enough to avoid enforcement.

The “Secured by Ethereum” Marketing Problem

From a compliance perspective, the disconnect between marketing and reality creates serious risks.

When an L2 markets itself as “secured by Ethereum” but operates with centralized control, that creates:

  1. Disclosure risk: Are users adequately informed about centralization?
  2. Custody risk: Who actually controls the assets on the L2?
  3. Liability risk: If the operator freezes funds, who is responsible?
  4. Regulatory classification risk: Is this a decentralized protocol or a centralized service?

Institutional Concerns

I work with institutions considering Ethereum exposure. Here’s what they ask:

“If we custody assets on Base, do we trust Coinbase or Ethereum?”

The answer should be “Ethereum” if the security claims are true. But the reality is “Coinbase,” because Base controls sequencing, contract upgrades, and emergency functions.

That changes the risk assessment entirely.

“What happens if an L2 operator is served with a court order to freeze assets?”

On Ethereum L1, this is extremely difficult due to decentralization. On a Stage 0 L2? The operator can comply immediately because they have admin keys.

Institutions need to understand this distinction.

Regulatory Clarity Requires Honest Disclosure

If L2s want to attract institutional capital, they need:

1. Clear disclosure about centralization levels

  • Stage 0/1/2 framework should be prominent in all marketing
  • Explicit statements about operator capabilities (censorship, freezing, upgrades)
  • Risk disclosures about trust assumptions

2. Roadmaps with accountability

  • Specific dates for Stage 1 and Stage 2 milestones
  • Consequences if milestones are missed
  • Public commitments to decentralization timelines

3. Legal structures that match operational reality

  • If operators control funds, structure accordingly
  • If operators can censor, disclose regulatory compliance approach
  • If operators capture MEV, explain why that’s compatible with claims of decentralization

Will Regulators Treat Stage 0 L2s Differently?

Here’s my prediction: Yes, and soon.

Regulators are starting to understand the difference between:

  • Truly decentralized protocols (minimal regulation)
  • Centralized operators marketing as decentralized (full regulation)

Stage 0 L2s with centralized operators will increasingly face:

  • Money transmitter licensing requirements
  • Securities law scrutiny (especially if they have governance tokens)
  • Consumer protection enforcement
  • Mandatory disclosure rules

Stage 2 L2s that are genuinely decentralized will have much better regulatory treatment.

This creates a forcing function for decentralization. L2s that stay at Stage 0 will face regulatory costs that Stage 2 L2s avoid.

The Path Forward

I actually agree with Brian’s three-point plan:

  1. Achieve Stage 2 decentralization (regulatory necessity)
  2. Share MEV with Ethereum L1 (economic sustainability)
  3. Increase value passthrough (alignment of incentives)

But I’d add a fourth:

4. Honest disclosure and regulatory compliance

  • Classify L2s accurately based on actual decentralization
  • Build legal structures that match operational reality
  • Proactively engage regulators instead of hoping to avoid notice

The L2s that succeed long-term will be the ones that:

  • Decentralize for real (not just roadmap promises)
  • Align economics with Ethereum L1
  • Comply with regulations honestly
  • Disclose risks transparently to users and institutions

The alternative is increasing regulatory scrutiny, enforcement actions, and institutional hesitation.

Compliance enables innovation. Ambiguity creates risk. Legal clarity unlocks institutional capital.

L2s need to choose: Decentralize or accept regulation as centralized service providers.