Cross-Domain MEV

What Is Cross-Domain MEV?

Cross-domain MEV is the value that can be extracted by coordinating transaction sequencing across two or more separate blockchain domains. A "domain" can be any independent execution environment: an L1 blockchain like Ethereum, an L2 rollup like Optimism or Arbitrum, a sidechain, or even a centralized exchange. Researchers formally define it as the maximum cumulative balance increase a user can achieve by controlling transaction ordering across several such domains simultaneously.

Traditional MEV operates within a single chain: a searcher spots a profitable opportunity in a pending transaction and reorders, inserts, or censors transactions within that chain's block. Cross-domain MEV adds a new dimension. Because separate chains have independent mempools, different block times, and asynchronous state updates, price discrepancies and information asymmetries naturally arise between them. Searchers who can act across these boundaries simultaneously unlock extraction opportunities that single-domain participants cannot access.

The expansion of Ethereum's rollup ecosystem has made this category increasingly significant. With dozens of rollups sharing the same underlying asset set but operating as isolated execution environments, the attack surface for cross-domain extraction has grown substantially. One academic study identified over 260,000 cross-chain arbitrage transactions across nine chains, generating an estimated lower-bound profit of $9.5 million.

How It Works

Cross-domain MEV exploits three fundamental properties of multi-chain architectures: communication latency between domains, state divergence across execution environments, and information asymmetry between participants who operate on different chains.

Cross-Chain Arbitrage

The most common form of cross-domain MEV is arbitrage across chains. When the same token trades at different prices on DEXs deployed on separate rollups or blockchains, a searcher can buy on the cheaper venue and sell on the more expensive one. Two primary strategies exist:

• Independent-leg arbitrage: the searcher executes opposite-direction trades on two chains without moving assets between them. This works when the searcher already holds inventory on both chains, eliminating the need for bridging.
• Bridge-based arbitrage: the searcher uses a cross-chain bridge to move assets from the cheaper venue to the more expensive one, then sells. Research indicates roughly 32% of observed cross-chain arbitrages involve bridging solutions.

Price gaps on rollups can persist longer than on a single chain. Studies have found that price discrepancies on rollups often last for 10 to 20 blocks, and researchers have identified over 500,000 unexploited arbitrage opportunities on rollups alone. The activity is also highly concentrated: five addresses generated more than half of all observed cross-chain arbitrage transactions.

Cross-Layer Sandwich Attacks

A sandwich attack on a single chain places a front-running transaction before a victim's trade and a backrunning transaction after it, profiting from the price impact. Cross-layer sandwich attacks extend this pattern across domains.

In a cross-layer variant, the attacker places the front-running transaction in a block on one domain (such as Ethereum L1) and targets the victim's transaction on a rollup. Because rollups typically rely on centralized sequencers that control transaction ordering, an attacker who can influence ordering on both the L1 and the L2 can execute sandwich attacks that span layers.

These attacks are particularly difficult to defend against. Standard single-chain protections focus on transaction ordering within a single block, but cross-layer sandwiches place the front-running transaction in the block preceding the target block on a different domain. Researchers estimate that attackers could have earned roughly $2 million through such sequencer-level manipulation.

Cross-Domain Liquidations

Lending protocols deployed across multiple chains create liquidation opportunities that span domains. A searcher monitoring collateral positions across several rollups can trigger liquidations by manipulating oracle prices or slippage on one domain while executing the liquidation transaction on another. When combined with flash loans, these attacks can be capital-efficient and executed atomically within certain domain pairs.

The Role of Shared Sequencers

Shared sequencers are designed to produce a single ordering over transactions destined for multiple rollups. Projects such as Espresso and Astria have proposed architectures where a shared sequencing layer constructs cross-rollup "meta-blocks," finalizing ordering across multiple rollups' transaction namespaces simultaneously.

This architecture creates a fundamental tension. On one hand, shared sequencing can enable cross-rollup composability and atomic execution of bundles spanning different rollups, which benefits users. On the other hand, it creates new MEV surfaces: a shared sequencer that sees transactions across multiple rollups gains the power to extract cross-domain MEV across all domains it serves.

The economic incentives are also challenging. Individual rollup operators earn revenue from sequencing and have limited motivation to delegate this role to a shared service that could redistribute that value. This structural tension has slowed adoption of shared sequencing in practice.

Why It Matters

Cross-domain MEV directly impacts users through worse trade execution. When a searcher front-runs a swap across domains, the user receives fewer tokens than expected. When arbitrageurs race to capture cross-chain price gaps, the resulting network congestion raises fees for everyone.

The problem extends to protocol-level economics. Cross-domain MEV can create centralizing pressures: actors with faster infrastructure, multi-chain presence, and sophisticated strategies gain outsized advantages. Research has documented this concentration trend, with a small number of addresses capturing the majority of cross-chain arbitrage profits.

For Bitcoin Layer 2 solutions like Spark, the design of the execution environment matters. Architectures that minimize reliance on shared mempools and centralized sequencers reduce the surface area for cross-domain extraction. The broader trend toward proposer-builder separation and order flow auctions reflects the industry's recognition that MEV is a systemic challenge requiring protocol-level solutions, not just application-level workarounds.

Mitigations

Encrypted Mempools

Encrypted mempools hide transaction contents until after ordering is committed, preventing searchers from seeing and exploiting pending transactions. Threshold encryption schemes require a committee to collectively decrypt transactions only after their position in a block is finalized. Shutter Network on Gnosis Chain has deployed this approach in production, encrypting transactions to a Keyper committee key that is revealed only after block commitment.

However, encrypted mempools face a tradeoff: by hiding transaction contents, they also suppress beneficial MEV activities like same-block auctions that can return value to users. They also do not fully address cross-domain MEV when the attacker observes state changes on other chains rather than reading mempool contents directly.

Fair Ordering Services

Fair ordering protocols aim to sequence transactions based on objective criteria such as the time they were received, rather than allowing sequencers to reorder for profit. Approaches include first-come-first-served (FCFS) policies and batch auctions that process simultaneous transactions at a uniform price. FCFS ordering shifts competition from ordering manipulation to latency optimization, where participants invest in faster infrastructure to submit transactions first.

Order Flow Auctions and MEV Redistribution

Rather than eliminating MEV entirely, order flow auctions aim to redistribute the extracted value back to users. Flashbots' MEV-Share protocol creates a matchmaking layer where users submit transactions with privacy-controlled hints. Searchers bid for the right to bundle their own MEV-capturing transactions alongside user transactions, and a portion of the resulting profit flows back to the user who supplied the order flow.

Flashbots has also proposed SUAVE as a chain-agnostic coordination layer that could extend this redistribution model across multiple blockchains, though the project remains in early development. Meanwhile, Flashbots' BuilderNet, a decentralized block building network, has been producing a significant share of Ethereum blocks using trusted execution environments (TEEs) for transaction privacy during the block building process.

Application-Level Defenses

Individual protocols can reduce their cross-domain MEV exposure through design choices:

• Using TWAP oracles that smooth price feeds over time, making single-block manipulation less profitable
• Implementing slippage protection and minimum output amounts on swaps to bound the extractable value per transaction
• Routing trades through private submission channels that bypass public mempools
• Designing AMM mechanisms that internalize arbitrage value rather than leaking it to external searchers

Risks and Considerations

Centralization Pressure

Cross-domain MEV extraction favors well-capitalized actors with multi-chain infrastructure. The capital, latency, and coordination requirements create natural economies of scale that concentrate extraction among a small number of sophisticated participants. This centralization extends to shared sequencers, which become a single point of failure if they serve multiple rollups.

Incomplete Mitigations

No existing mitigation fully solves cross-domain MEV. Encrypted mempools address single-chain front-running but cannot prevent extraction based on cross-chain state observation. Fair ordering services shift competition from ordering to latency. Order flow auctions redistribute value but do not eliminate extraction. The cross-domain nature of the problem means that solutions deployed on only one chain leave extraction opportunities on other domains.

Bridge and Sequencer Trust

Many cross-domain MEV strategies depend on bridges and sequencers, both of which introduce trust assumptions. Centralized sequencers can extract MEV directly, while bridges introduce latency and smart contract risk that searchers must price into their strategies. The interaction between bridge security and MEV extraction remains an active area of research.

Evolving Attack Surface

As the multi-chain ecosystem grows, the potential for cross-domain MEV grows with it. Each new rollup, bridge, or interoperability protocol adds domains that searchers can coordinate across. Defenses that work for two-domain interactions may not scale to environments with dozens of interconnected chains. For a deeper analysis of MEV dynamics on Layer 2 networks, see the research on Bitcoin L2 MEV extraction and Ethereum MEV and proposer-builder separation economics.

This glossary entry is for informational purposes only and does not constitute financial or investment advice. Always do your own research before using any protocol or technology.