The Hidden Tax of MEV on Modular Execution Layers

Centralized sequencers in rollups like Arbitrum and Optimism have exclusive rights to transaction ordering. This creates a non-competitive, rent-seeking position where MEV is captured off-chain, with zero visibility or revenue sharing for the underlying chain.

• Opaque Extraction: Users pay a hidden tax via front-running and sandwich attacks they cannot see.
• Sovereignty Risk: The L2's economic security is ceded to a single entity's profit motive.

Decouple block building from proposing, as pioneered by Ethereum's PBS, to create a competitive market for rollup block space. This forces MEV value to be auctioned transparently on the public mempool.

• Transparent Auctions: Builders like Flashbots and Titan compete, pushing extractable value back to the proposer (the rollup).
• Credible Neutrality: The winning block is committed to L1, removing the sequencer's ability to censor or reorder.

Solving MEV requires hiding information. Encrypted mempool protocols like Shutter Network and shared auction infrastructure like Flashbots' SUAVE prevent front-running by default, shifting the supply chain from extraction to optimization.

• Pre-Execution Privacy: Transactions are encrypted until included in a block, neutralizing harmful MEV.
• Cross-Chain Liquidity: SUAVE aims to be a decentralized block builder for all chains, aggregating liquidity and intent.

The final shift is from transactions to intents. Systems like UniswapX, CowSwap, and Across let users specify a desired outcome, not a precise path. Solvers compete to fulfill it optimally, capturing MEV as efficiency gains shared with the user.

• User Sovereignty: The supply chain works for the user, not against them.
• Efficiency as Revenue: MEV becomes a positive-sum game, measured in better prices and gasless UX.

Rollup sequencers centralize transaction ordering, creating a censorship vector and a monopoly on MEV capture. This undermines decentralization and forces users to pay a hidden tax.

• Centralized Profit: All MEV flows to a single entity.
• Censorship Risk: Sequencer can front-run or exclude transactions.
• Value Leakage: User value is extracted before reaching L1.

Projects like Astria, Espresso, and Radius are building decentralized sequencer sets that order transactions for multiple rollups. This creates a credibly neutral layer for cross-domain block building.

• MEV Redistribution: Auctions order flow, redistributing value.
• Atomic Composability: Enables cross-rollup transactions.
• Censorship Resistance: No single entity controls the queue.

SUAVE and Radius encrypt transaction content until execution, while Espresso offers fast pre-confirmations. This neutralizes front-running and creates a competitive market for block space.

• Privacy: Encrypted mempools blind searchers and sequencers.
• Fair Ordering: Transactions are ordered without viewing content.
• User Assurance: Pre-confirmations guarantee inclusion and price.

Splitting liquidity across dozens of rollups and L2s creates arbitrage hell. Searchers exploit price differences between domains, extracting value that should remain with LPs and users.

• Inefficiency: Capital is trapped in isolated pools.
• Extraction Multiplier: More bridges and chains mean more arbitrage paths.
• User Cost: Slippage and fees increase to cover MEV risk.

UniswapX, CowSwap, and Across shift from transaction-based to outcome-based execution. Users submit intents; a decentralized solver network competes to fulfill them optimally, capturing MEV for the user.

• MEV Recapture: Solvers internalize value, returning it via better prices.
• Cross-Chain Native: Intents abstract away fragmentation.
• Gasless UX: Users sign messages, not gas-heavy transactions.

Adapting Ethereum's PBS model, rollups like Arbitrum are exploring separating block building from proposing. This creates a competitive market for block construction, democratizing MEV.

• Specialized Builders: Optimize for MEV extraction efficiency.
• Fair Auctions: Proposer selects the highest-value block bundle.
• Protocol Revenue: MEV can be captured and burned or distributed.

Sequencers on rollups like Arbitrum or Optimism can front-run user intents before they are finalized on L1. This creates a two-layer extraction game where MEV is captured on the L2, then again on the L1 during the bridge settlement.\n- Example: A large swap on an L2 DEX is identified, front-run, and the profitable arbitrage is also captured on the L1 bridge like Across.\n- Impact: Users pay MEV tax on both the execution and settlement layers, eroding promised cost savings.

Fast, centralized bridging protocols like LayerZero and Axelar create predictable latency arbitrage windows. Proposers on modular settlement layers (e.g., Celestia) or shared sequencers (e.g., Espresso, Astria) can exploit state differences across rollups before proofs are verified.\n- Mechanism: A validator sees an asset price delta between Rollup A and Rollup B, uses a fast-message bridge to execute the arb, and settles before slow bridges finalize.\n- Result: Value accrues to the fastest, most centralized infrastructure, not to rollup users or token holders.

In a modular stack, the separation of block building (execution) from proposing (consensus/settlement) recreates the PBS problem from Ethereum L1 at every layer. Rollup sequencers become the natural builders, and L1 validators become the proposers, creating a multi-layered cartel.\n- Risk: Sequencers extract MEV from their users, then validators extract MEV from sequencers via inclusion ordering on the settlement layer.\n- Outcome: MEV revenue is syphoned away from the execution layer's security budget, forcing unsustainable token emissions to pay for security.

The only viable endgame is to cryptographically hide transaction content until execution. This requires a dedicated decentralized network for order flow.\n- Approach: Protocols like SUAVE aim to create a neutral, chain-agnostic marketplace for preference expression (intents) and execution.\n- Mechanism: Users submit encrypted intents. Solvers compete on execution quality in a sealed-bid auction, with proofs of correct execution.\n- Goal: Break the information asymmetry that enables front-running, returning MEV as a rebate to the user.

In a modular stack like Arbitrum or Optimism, the sequencer is a centralized profit center. It can front-run, censor, and reorder transactions before they hit the base layer (e.g., Ethereum). The "fair ordering" guarantee is a myth without explicit protocol design.

• Hidden Tax: MEV is captured off-chain, invisible to L1 settlement.
• Centralization Vector: Profits incentivize sequencer centralization, defeating decentralization goals.
• User Cost: Savings from cheap L2 execution are offset by value leakage to the sequencer.

Bake MEV management into the protocol layer. Force a split between the entity that builds blocks (Builder) and the one that proposes them (Proposer/Sequencer). This is the core innovation behind Ethereum's PBS and must be adapted for rollups.

• Transparent Auction: MEV is competed for in a public market, converting hidden tax into explicit protocol revenue.
• Censorship Resistance: Proposer can't easily censor; they just choose the highest-paying, valid block.
• Design Mandate: Requires a native auction mechanism, not a bolt-on.

Move sequencing off the individual rollup to a neutral, shared layer. Espresso Systems, Astria, and Fuel are building this. Ethereum's SUAVE aims to be a universal preference chain for MEV.

• Level Playing Field: Rollups outsource sequencing, gaining neutrality and shared security.
• Cross-Domain MEV: Enables efficient arbitrage between Arbitrum, Optimism, Base.
• Complexity Trade-off: Introduces new latency and liveness assumptions versus integrated sequencers.

Shift from transactional (expose full tx) to intentional (declare desired outcome) systems. Flashbots SUAVE, CoW Swap, and UniswapX are pioneering this. Users submit signed preferences, solvers compete to fulfill them.

• MEV Resistance: Obfuscates transaction content until execution.
• Better Execution: Solvers optimize for user-specified outcomes across venues.
• Paradigm Shift: Requires new standards and infrastructure, breaking compatibility with EVM tooling.