The Mechanism Spec: Designing L1-L2 Economic Coordination That Might Actually Work

tl;dr

• The primary recommendation is the blob-consumption burn as a mandatory floor — a protocol-enforced burn of 1.5x the current blob fee for data posted. It is Sybil-resistant, requires no revenue oracles, eliminates the verification problem, and can be implemented today. This is the one mechanism that does not depend on composability having value.
• Composability Tiers are conditionally promising — but only if a product thesis emerges to anchor them. Tier 1 (“Ethereum Aligned” status, fast messaging, priority blobs) costs Base ~$3.77M/year and is economically rational today. But the higher tiers rest on synchronous composability having real value, which is empirically unproven. Intents already solve 90-95% of cross-chain needs.
• Trade agreements must precede technical architecture, not follow it. The EEZ built the highway and hopes traffic will come. The right sequence is: define what L1 and L2s owe each other (fees, MEV, exit terms, dispute resolution), then build the infrastructure to serve those terms. StarkWare’s dAMM is the model — a product that demanded shared liquidity by design, rather than shared liquidity as a feature looking for a product.
• Based sequencing is right for some L2s, not a universal endgame — at a 50-70% MEV rebate, based L2s lose 6-29% of revenue versus self-sequencing. Taiko proves feasibility but is currently unprofitable (83.9 ETH loss in two weeks). App-specific and non-DeFi L2s gain nothing from L1 composability.
• Two mechanisms were eliminated before reaching the final three — Progressive DA Pricing dies on Goodhart’s Law and Sybil vulnerability. Security Bond imposes deadweight capital costs without composability benefits. Both inherit an unsolved TVL oracle problem.
• The honest conclusion: synchronous composability uniquely enables only MEV efficiency (centralizing) and cross-chain flash loans (niche). If no product like dAMM emerges that genuinely demands shared infrastructure, the composability carrot is not big enough to anchor an economic zone. Start with the trade agreements. Let the technical architecture follow.

Table of Contents

1. Introduction: From Metaphor to Mechanism
2. Five Candidates, Two Eliminated
3. Mechanism 1: Revenue-Indexed Burn — The Mandatory Floor
4. Mechanism 2: Composability Tiers — The Voluntary Incentive
5. Mechanism 3: Based Sequencing with Revenue Rebate — The Structural Endgame
6. How the Three Layers Stack
7. Real-World Parallels: Where These Structures Have Been Tested
8. Recommendation: Trade Agreements First, Architecture Second
9. The Honest Caveat
10. Data Sources & Methodology

1. Introduction: From Metaphor to Mechanism

The first report in this series asked whether synchronous composability is a solution looking for a problem. The answer: it is real but narrow — the trading venues that won market share did so through product excellence on single chains, not through cross-chain atomic execution.

The second report asked what L1-L2 economic coordination would need to look like. The answer: something closer to a trade agreement than a handshake — binding terms with genuine benefits, not voluntary arrangements that produce shallow integration. We found that every prior attempt at cross-chain economic coordination (Astria, Espresso, AggLayer, Cosmos IBC, Polkadot) has failed to produce durable economic relationships.

This report is the mechanism design specification. The question is no longer what needs to happen, but how. Five candidate mechanisms were evaluated. Two were eliminated on fundamental design flaws. Three advanced, and this report recommends how they combine into a layered system — with a clear primary recommendation.

Two constraints frame everything that follows:

The Base constraint. Base controls ~47% of L2 TVL and ~62% of L2 fee revenue. Any mechanism Base rejects is dead on arrival. This is not an opinion — it is arithmetic. Mechanism design that ignores the largest participant’s incentives is not mechanism design. It is wishful thinking.

The Celestia constraint. Celestia offers a 2.8-55x cost advantage over Ethereum for data availability ($0.07-$7.31/MB vs. $3.83-$20.56/MB). This sets a hard ceiling on what Ethereum can extract through pricing alone. Any mechanism that makes Ethereum L1 too expensive simply accelerates the migration to alternative DA layers. The mechanism has to create value, not just capture it.

2. Five Candidates, Two Eliminated

Before walking through what survived, here is what did not — and why.

Eliminated: Progressive DA Pricing

The idea: index blob fees to the L2’s TVL. Larger, more successful L2s pay proportionally more for data availability.

The fatal flaw is Goodhart’s Law. The moment TVL determines price, TVL becomes a metric to be gamed rather than a measure of economic activity. An L2 can split into multiple smaller L2s to stay below pricing thresholds (Sybil splitting). TVL can be inflated or deflated through bridge mechanics, recursive lending, and point farming — we have watched DeFi protocols do exactly this for four years. More fundamentally, TVL is measurable on-chain but manipulable by anyone willing to pay the gas cost of a deposit.

The supervisor finding confirms this: resource-based pricing (per-blob) is inherently Sybil-resistant because you cannot fake posting data. TVL-based pricing is not. This mechanism also punishes success — the more valuable an L2 becomes, the more expensive its DA costs, creating a perverse incentive to migrate to Celestia precisely when Ethereum should be trying hardest to retain its most valuable tenants.

Eliminated: Security Bond / Stake-to-Play

The idea: require L2s to stake ETH proportional to their TVL as a security bond. Essentially a capital commitment that signals alignment.

This fails for the same reason plus one more. It inherits the TVL oracle problem from Progressive DA Pricing — same measurement issues, same Sybil vulnerability. But it adds a second problem: deadweight capital cost. The staked ETH does not do anything for the L2. It does not improve composability, reduce latency, or unlock new products. It is a pure tax on capital that could otherwise be productively deployed. In a market where L2s already retain $28-321 for every $1 they pay Ethereum, imposing a capital lockup without corresponding benefits is asking L2s to accept worse economics for nothing in return.

The EIP-8141 rejection precedent matters here. That proposal was killed as “too complex for what it delivers.” Any mechanism that imposes significant costs without corresponding benefits will face the same resistance — and should.

3. Mechanism 1: Revenue-Indexed Burn — The Mandatory Floor

The Design

Every L2 posting data to Ethereum burns a percentage of its gross sequencer revenue in ETH. This is the floor — not the whole structure, just the minimum economic contribution that all L2s make to the settlement layer that provides their security.

Original Specification

The Revenue Verification Problem

Here is where honesty matters more than elegance. Revenue verification is, in the general case, essentially unsolvable. An L2 sequencer can route revenue through off-chain channels, restructure fee collection to minimize on-chain visibility, or simply underreport. The commit-reveal-challenge scheme works against naive cheating but not against a sophisticated actor who restructures their revenue pipeline.

This is not a theoretical concern. It is the same problem that makes income tax enforcement expensive and imperfect — and that is with nation-state-level audit infrastructure. A protocol-level mechanism cannot rely on accurate revenue measurement when the entity being measured controls the revenue pipeline.

The Better Alternative: Blob-Consumption Burn

Rather than indexing to revenue (which is gameable), index to a resource the protocol can already measure: data posted. The mechanism works like this — settlement requires burning ETH equal to 1.5x the current blob fee for the data posted. You cannot fake how much data you post. The burn scales naturally with L2 activity. No revenue oracle needed. No Sybil vulnerability (splitting into multiple L2s does not reduce total data posted). And critically, it transforms the burn from a revenue tax into a minimum total economic contribution that includes blob fees as a component.

This is a cleaner design. It does not solve every problem — a 1.5x multiplier is still a parameter that requires governance — but it eliminates the hardest problem (revenue verification) while preserving the core function (mandatory floor on L1 value capture).

The gamma rate should be dynamically indexed to blob costs rather than fixed. A static 5% will be too low in bull markets when L2 revenue surges and too high when revenue compresses. Dynamic indexing tied to blob market conditions keeps the floor proportional without requiring constant governance intervention.

4. Mechanism 2: Composability Tiers — The Voluntary Incentive

This is where the design gets interesting — and where the recommendation lives.

The Structure

Three tiers, each offering progressively more composability in exchange for progressively more alignment. The critical design feature: no L2 is worse off than today. Tier 0 is the status quo. Higher tiers are voluntary opt-ins that unlock concrete benefits.

Why Tier 1 Is the Immediate Priority

Tier 1 is the wedge. The costs are low — for Base, roughly $3.77M/year at a 5% burn rate, against ~$55M in annual profit. The benefits are concrete and deliverable today:

Fast messaging. Sub-30-second cross-chain communication is genuinely useful for bridging, governance, and cross-chain account management. It does not require synchronous composability — it requires infrastructure that Ethereum can build and offer as a Tier 1 benefit.

“Ethereum Aligned” branding. This sounds soft. It is not. A verifiable, on-chain status that wallets, bridges, and aggregators can surface to users has real economic value. Users choosing between two L2s — one displaying “Ethereum Aligned” and one not — will bias toward the credentialed option, all else equal. The value is not in the label itself but in the verification behind it: Stage 1 security, ETH as gas, standards compliance. It is a quality signal that differentiates in a market drowning in undifferentiated L2s.

Priority blob inclusion. When blob demand spikes (and it will as L2 activity grows), Tier 1 L2s get priority access. This is insurance against the scenario where blob space becomes contested — a benefit that costs nothing in normal conditions but is extremely valuable in stressed conditions.

Base’s Path

Base adopts Tier 1. The cost ($3.77M/year) is ~6.9% of its $55M profit. For that, it gets fast messaging, brand credentialing, and blob inclusion priority. This is economically rational today — before composability has proven its value.

Tier 2 is the aspiration. Based sequencing or a composability bond unlocks synchronous composability with L1 and other Tier 2 L2s. But Tier 2 has a bootstrap problem: synchronous composability is only valuable when there is a critical mass of Tier 2 participants. The EEZ’s founding members (Aave, Centrifuge) plus Taiko provide an initial network, but the honest assessment is that Tier 2’s value proposition remains unproven until products emerge that genuinely require cross-chain atomic execution.

5. Mechanism 3: Based Sequencing with Revenue Rebate — The Structural Endgame

The Design

L2s delegate sequencing to L1 proposers. In exchange for surrendering sequencer sovereignty, they receive a rebate of R% of the MEV attributable to their transactions.

Feasibility

Taiko proves based sequencing works technically. It also proves it is currently unprofitable — 83.9 ETH lost in two weeks. The most aligned architecture in Ethereum’s toolkit is losing money. This is not a death sentence (early infrastructure often runs at a loss), but it is a fact that mechanism designers cannot ignore.

The MEV attribution problem is real. Counterfactual simulation — “what would the MEV have been if this L2’s transactions were excluded?” — is the best available approach, but it is imperfect. MEV is emergent, not additive. Cross-domain MEV from the interaction between L1 and L2 orderflow cannot be cleanly decomposed into “L1’s share” and “L2’s share.” Any revenue-sharing mechanism built on MEV measurement rests on unreliable inputs.

Who Based Sequencing Is For (And Who It Is Not For)

Based sequencing makes sense for DeFi-heavy L2s where L1 composability unlocks concrete value — cross-chain liquidations, unified lending pools, atomic arbitrage. These L2s benefit from L1 proposer inclusion because their users and protocols benefit from synchronous access to L1 state.

It does not make sense for app-specific L2s, gaming chains, social protocols, or non-DeFi use cases. An L2 running an order-book exchange with no L1 dependencies gains nothing from based sequencing and loses sequencer revenue. Framing based sequencing as the universal endgame is wrong. It is one path to maximum alignment for L2s that want it — available within Tier 2, not mandated across the ecosystem.

6. How the Three Layers Stack

┌─────────────────────────────────────────────────────────────────┐
│  Layer 3: Based Sequencing with Rebate                         │
│  (long-term, voluntary, highest alignment — within Tier 2)     │
├─────────────────────────────────────────────────────────────────┤
│  Layer 2: Composability Tiers                                  │
│  (voluntary opt-in for composability benefits — Tier 0/1/2)    │
├─────────────────────────────────────────────────────────────────┤
│  Layer 1: Blob-Consumption Burn                                │
│  (mandatory floor, protocol-enforced — 1.5x blob fee)         │
├─────────────────────────────────────────────────────────────────┤
│  Layer 0: EIP-7918 Blob Fee Floor                              │
│  (already live — minimum blob pricing)                         │
└─────────────────────────────────────────────────────────────────┘

The critical design principle: maximum total cost envelope, not additive stacking. If an L2 participates in based sequencing (Layer 3) and is Tier 1 aligned (Layer 2) and pays the blob-consumption burn (Layer 1), the total cost to that L2 must be capped. Layers credit against each other. The burn floor (Layer 1) is absorbed into the Tier 1 requirements (Layer 2). The MEV rebate in based sequencing (Layer 3) accounts for the burn already contributed. No L2 should face a scenario where compliance with all three layers produces costs that exceed the rationality threshold.

Without this envelope, the system becomes a political negotiation surface. Three years spent arguing about gamma rates, rebate percentages, and tier thresholds — instead of shipping the technology that makes composability worth paying for. Complexity is the enemy. The mechanism should be simple enough that an L2 operator can calculate their total cost in five minutes and decide whether the benefits justify it.

7. Real-World Parallels: Where These Structures Have Been Tested

These mechanisms are not novel in structure — only in application. Each has precedent in traditional economic coordination.

Franchise Models (Revenue-Indexed Burn)

McDonald’s franchisees pay 4-5% royalty on gross sales plus ~4% for advertising, totaling 8-13%. The industry average franchise royalty is 6.7%. The 5% burn rate proposed here sits within this range. The analogy is imperfect — McDonald’s provides a brand, supply chain, and training infrastructure that Ethereum does not yet offer to L2s — but the revenue-indexed structure is battle-tested across thousands of franchise relationships.

7-Eleven’s model is closer to based sequencing: a 50-60% gross profit split where the franchisor (L1) captures a majority share but also handles core operations (sequencing). The L2 equivalent: surrender sequencing, receive a rebate, benefit from L1 infrastructure.

Platform Commitment Discounts (Composability Tiers)

AWS reserved instances offer 40-72% discounts for long-term commitment. Committed customers are 3.5x less likely to switch providers. The composability tier structure mirrors this: higher commitment (Tier 1 → Tier 2) unlocks progressively better terms (fast messaging → synchronous composability). The economics of platform lock-in through superior service, not through exit penalties.

Apple’s App Store applies a 30% platform fee that dropped to 15% for small developers — a progressive structure that acknowledges different participants have different cost sensitivities. The tier system implicitly does this: small L2s stay at Tier 0 with minimal costs, large L2s opt into higher tiers where the benefits justify the alignment costs.

Collective Security Contributions (Blob-Consumption Burn)

NATO’s 2% GDP defense spending target is the most instructive parallel — and the most cautionary. The target was set in 2014. It took 11 years and a Russian invasion of Ukraine for most members to approach compliance. Social pressure alone did not work. An external threat catalyst was required.

The Ethereum equivalent: Celestia’s cost advantage is the external threat, and L2 defection to alternative DA layers is the catalyst. The blob-consumption burn must be set at a level where paying it is clearly less costly than the alternatives (losing “Ethereum Aligned” status, losing fast messaging, losing blob priority). If the burn exceeds that threshold, L2s will simply leave — and unlike NATO, there is no mutual defense clause forcing them to stay.

Resource Royalties (Dynamic Indexing)

Alberta’s oil royalty structure starts at 5% and scales upward with commodity prices. This is the model for dynamic gamma indexing — the burn rate adjusts with blob market conditions. Low demand, low burn. High demand, higher burn. The mechanism self-calibrates without requiring constant governance votes.

8. Recommendation: Trade Agreements First, Architecture Second

The primary recommendation is the blob-consumption burn as a standalone floor — the one mechanism that works regardless of whether composability proves its value. Composability Tiers are conditionally promising, with Tier 1 as the immediate priority. But the higher tiers, and the entire composability thesis, depend on products emerging that genuinely demand shared infrastructure.

Here is the full assessment.

The blob-consumption burn is the only mechanism that stands on its own. It does not depend on composability having value. It does not require L2s to surrender sequencer revenue. It is Sybil-resistant, protocol-enforced, and implementable today. It is a tax — but a fair one, tied to the resource actually consumed. Set at 1.5x the current blob fee, dynamically indexed to blob market conditions. This is the floor.

Composability Tiers are conditionally promising — but only if a product thesis emerges. Tier 1 offers fast messaging, brand credentialing, and blob priority at costs L2s can afford (~$3.77M/year for Base). That is real value. But the higher tiers rest on synchronous composability justifying their sovereignty cost. Our own analysis found that sync composability uniquely enables only two things: more efficient MEV extraction (which risks further centralization among sophisticated actors) and cross-chain flash loans (commercially niche). Intents already solve 90% of cross-chain UX, and that number is climbing as intent infrastructure improves.

The distinction between dAMM and the EEZ is instructive. StarkWare and Loopring’s dAMM (distributed AMM) started with a product — an L2-powered AMM that aggregates liquidity on L1 and distributes state across L2s. The product thesis demanded shared liquidity by design; the infrastructure requirement followed from the product. The EEZ inverts this: shared infrastructure first, products TBD. We have seen this pattern before — Cosmos IBC, Polkadot parachains, CORBA, Astria. Technically sound infrastructure without a product thesis produces ghost chains.

Based sequencing is right for some L2s, wrong for most. At 50-70% MEV rebate, the revenue loss is 6-29%. Taiko proves feasibility but is unprofitable. Non-DeFi L2s gain nothing from L1 composability. It is the structurally deepest alignment — and the one fewest L2s will accept.

The priority sequence — what should actually happen:

1. Ship the blob-consumption burn at 1.5x. Protocol-enforced floor. No revenue oracle. Dynamically indexed. This can move through the EIP process now.

2. Define Tier 1 criteria and ship the “Ethereum Aligned” registry. ETH as gas, Stage 1 security, ERC-7683, blob-consumption burn compliance. Make it verifiable on-chain so wallets and aggregators can surface it. This is the quick win that creates real brand value.

3. Define the trade agreements before the technical architecture. Before building Tier 2 composability infrastructure, specify the economic terms: what does each party give, get, and give up? Fee schedules, MEV allocation rules, exit procedures, dispute resolution. The EU’s 70,000-page acquis communautaire came before the single market’s full implementation, not after.

4. Let products drive the composability demand. dAMM-style designs — products that genuinely require shared liquidity — should lead. If and when those products emerge at scale, Tier 2 composability becomes a natural extension of demonstrated demand. If they do not emerge, composability tiers beyond Tier 1 should remain aspirational rather than engineered.

5. Enforce a total cost envelope. No L2 participating across all layers pays more than X% of revenue in total L1 contributions. Make the ceiling explicit and public.

9. The Honest Caveat

Everything above rests on an assumption that is empirically unproven: that composability has real economic value.

The composability thesis report found that the trading venues which won market share did so through product excellence on single chains. Cross-chain bridge volume represents ~8-9% of intra-chain DEX activity. The products that genuinely require synchronous cross-chain execution — cross-chain flash loans, unified lending pools with cross-chain liquidation — exist in theory and in academic papers, but not yet in production at scale.

If no product emerges that makes synchronous composability genuinely load-bearing, the composability tier system degrades. Tier 1 becomes “pay a burn for a badge.” Tier 2 becomes an empty category. The revenue-indexed burn operates as a standalone tax with no corresponding service. The entire incentive structure becomes a more sophisticated version of what Polkadot built — technically impressive infrastructure that nobody uses because nobody built the products that need it.

This is not a reason to abandon the design. It is a reason to sequence correctly.

The strongest version of this proposal has two parts, and the order matters:

First, define the trade agreements. Before building composability infrastructure, specify the economic terms. What do L1 and L2s owe each other? Under what conditions? With what enforcement? The blob-consumption burn is step one — a fair floor that can exist independently. Beyond that, the terms need to be negotiated, not assumed.

Second, let products drive the technical architecture. dAMM showed the right approach: a product design that demanded shared liquidity, not shared liquidity offered as a feature in search of a product. Fast messaging that actually works in under 30 seconds. Cross-chain state proofs that enable novel DeFi primitives. Shared liquidity mechanisms — like dAMM’s L1 pool with distributed L2 state — that demonstrably improve capital efficiency. These are product and engineering problems, not mechanism design problems. Solve them, and the composability tiers become a natural extension of demonstrated demand. Fail to solve them, and no amount of tier design or burn calibration will matter.

The bootstrap problem is not solved by incentive design. It is solved by building products so obviously useful that L2s adopt shared infrastructure because users demand it. And it is solved by defining the economic terms clearly enough that every participant knows what they are signing up for before the infrastructure is built.

Trade agreements first. Products second. Technical architecture third. We have been doing this in reverse.

Data Sources & Methodology

Primary Data Sources

• L2 Revenue & Profitability: Token Terminal, 21Shares, Messari (State of the Superchain H1 2025), Reflexivity Research
• L1-L2 Payment Flows: Token Terminal, Pine Analytics, CryptoSlate, Fidelity Digital Assets
• Blob Fee Economics: L2BEAT Scaling Costs, CryptoSlate, Fidelity Digital Assets — “The Fusaka Upgrade: Scaling Meets Value Accrual”
• Celestia DA Costs: Celestia documentation, independent benchmarking (Zeeve, CoinLaw)
• Taiko Economics: Taiko Labs (Mirror), public on-chain data
• EEZ Announcement & Members: The Block, CoinDesk, Gnosis official communications (March 29, 2026)
• EIP-7918: Ethereum Improvement Proposals repository, Fidelity Digital Assets analysis
• Based Rollup Analysis: DWF Labs, Sygnum Bank, Taiko Labs
• Shared Sequencing Precedents: Phemex/MEXC (Astria shutdown December 2025), Blockworks/CoinDesk (Espresso), Zeeve/CoinLaw (AggLayer developer defections)
• Cross-Chain MEV: “Cross-Chain Arbitrage: The Next Frontier of MEV in Decentralized Finance” (arXiv:2501.17335, ACM SIGMETRICS 2025)

Real-World Parallel Sources

• Franchise economics: FDD, Franchise Business Review, McDonald’s Corporation FDD (2024)
• AWS reserved instance pricing: AWS public pricing documentation
• NATO defense spending: NATO official statistics, Brookings Institution
• Apple App Store commission structure: Apple Developer documentation, Epic v. Apple court filings
• Alberta oil royalties: Government of Alberta, Natural Resources Revenue Data
• 7-Eleven franchise structure: 7-Eleven FDD, franchise economics literature

Methodology Notes

• Revenue projections for burn calculations use 2025 full-year L2 sequencer revenue data from Token Terminal and Messari, applied at the specified gamma rates.
• Based sequencing revenue impact estimates use Taiko’s public financial data and MEV research literature for rebate range modeling.
• Celestia cost comparisons use publicly available pricing data; ranges reflect blob size variability and network conditions.
• The 1.5x blob-consumption burn multiplier is a design parameter, not a derived figure. It represents a starting point for governance calibration.
• No data points have been fabricated. Where specific numbers were unavailable, the gap is noted. Where projections are used, they are labeled as such.
• This report is the third in a trilogy: “The Composability Question” (March 29, 2026), “The Missing Treaty” (March 29, 2026), and this mechanism specification. Cross-references to prior findings are thematic, not repetitive.