Why Cross-Domain MEV Will Break Current Mitigation Designs

Liquidity is now spread across dozens of L2s and app-chains, but user intent is not. A swap from Arbitrum to Base requires atomic execution across multiple, asynchronous domains.\n- Creates multi-domain atomic bundles that are impossible to execute trustlessly today.\n- Amplifies value leakage as arbitrageurs must race across settlement, data availability, and execution layers.

User-centric architectures like UniswapX and CowSwap outsource transaction construction. This creates centralized points of failure—the solvers.\n- Solvers become cross-domain MEV hubs, controlling flow across Ethereum, Arbitrum, Optimism.\n- Creates information asymmetry; the solver sees the full cross-chain intent and can front-run its own user.

L2s like Arbitrum and Base have sub-second pre-confirmations, while bridges like Across and LayerZero have their own optimistic or light-client verification. There is no universal time.\n- Enables time-bandit attacks where MEV is extracted after a user thinks a tx is final on one chain but before it settles on another.\n- Breaks the atomicity assumption of PBS (Proposer-Builder Separation), as no single proposer controls all involved chains.

MEV bots exploit the latency between L2 state roots posted to L1. A classic cross-domain arbitrage can involve Uniswap on Arbitrum, Curve on Optimism, and Ethereum mainnet in a single atomic bundle. Current sequencer-level PBS (like Flashbots SUAVE) cannot coordinate finality across these sovereign settlement layers, creating a ~12-second vulnerability window for value leakage.

Shift from transaction execution to intent declaration. Protocols like UniswapX and CowSwap abstract routing, but a cross-domain shield requires a shared sequencer network (e.g., Espresso, Astria) or a verification layer like EigenLayer. This creates a coordinated inclusion zone where intents are matched and settled atomically across chains, neutralizing latency-based arbitrage.

• Key Benefit: Atomic cross-domain settlement
• Key Benefit: User receives optimal route, not exploitable price

PBS on a single chain (Ethereum) creates a robust, neutral market. However, a builder winning a slot on Ethereum has no obligation to be neutral on Arbitrum or Base. This allows vertically integrated entities to become cross-domain cartels, extracting value by controlling the flow of transactions across the entire stack. The economic power is multiplicative, not additive.

Architect a PBS system where the builder role is defined across a domain set, not per chain. This requires a canonical reputation system and slashing conditions that penalize cross-domain adversarial behavior. Think EigenLayer for attestations or a layerzero-style oracle network for verifying sequencer commitments. Builders compete on a global cross-chain bundle, not per-domain blocks.

• Key Benefit: Breaks cartel formation
• Key Benefit: Aligns economic incentives across ecosystems

Generalized intent architectures (like UniswapX or Across) expose user preferences to a network of solvers. In a cross-domain world, a solver with visibility into intent flow on Polygon can front-run fulfillment on Avalanche. This creates a meta-MEV game where the most valuable resource is not block space, but predictive knowledge of cross-domain state changes.

The endgame is a cross-domain encrypted mempool. Solvers commit to intent fulfillment without seeing the full details, revealed only upon execution commitment. This requires threshold decryption schemes (e.g., FHE or MPC networks) coordinated by the shared sequencer layer. Projects like Shutter Network are pioneering this for single-chain, but the cross-domain variant is the ultimate shield.

• Key Benefit: Solver competition without leakage
• Key Benefit: User privacy as a systemic property

Cross-domain atomic bundles make isolated PBS and proposer-level auctions obsolete. A searcher's profit is the sum of marginal gains across 3+ chains, rendering single-chain fee burn or redistribution ineffective.\n- Uncaptured Value: A $1M arb bundle pays $50k in gas but only $5k to a single proposer.\n- New Attack Surface: Proposers are incentivized to collude across domains or be outbid by cross-chain entities.

Solvers in systems like UniswapX and CowSwap currently compete within one domain. Cross-domain intents create a meta-game where the most valuable execution path is contested by solvers who must also secure cross-chain liquidity and ordering.\n- Solver Cartels: The capital requirement to win cross-domain auctions favors centralized solver pools.\n- Timing Attacks: Latency between domain finalities (~2s to ~12m) creates new extractable windows for generalized front-running.

Bridges like Across, LayerZero, and Circle's CCTP are natural orderers for cross-domain flow. Their sequencers/relayers become the supreme MEV extractors, capable of censoring or reordering transactions worth 100x their native gas fees.\n- Vertical Integration: Bridge operators will internalize searcher functions, killing competitive markets.\n- Protocol Capture: DApps will be forced to choose bridges based on MEV kickbacks, not security.

Shared sequencers (e.g., Espresso, Astria) for rollups only solve coordination within their cohort. They create a new, larger MEV domain that must still interact with external L1s and other sequencer sets, pushing the problem upstream.\n- Domain Wall MEV: The value transfers between sequencer sets become the new high-stakes auction.\n- Fragmented Sovereignty: Rollups lose control over their own ordering to a consortium, without solving the cross-consortium problem.

Protocols like Chainlink CCIP and Wormhole that generalize message passing don't mitigate MEV—they systematize it. By providing a canonical data layer for cross-chain state, they create a predictable, high-value coordination point for complex extraction strategies.\n- Oracle Manipulation: Cross-domain arbitrage can now be triggered by and settled against oracle updates.\n- Amplified Spoofing: Fake intent traffic on one chain can be used to extract value on another via these channels.

Mitigation must be architected at the interoperability layer from day one. This means:\n- MEV-Aware Bridges: Bridges that auction sequencing rights or enforce fair ordering (e.g., SUAVE-inspired designs).\n- Cross-Domain PBS: A protocol that lets searchers bid for atomic inclusion across multiple domains simultaneously, with fees redistributed to all involved.\n- Intent Coordination Markets: Solvers bid not just for execution, but for guaranteed cross-domain atomicity, internalizing the MEV risk.