Why Bridge Fees Create Perverse Governance Incentives

A dominant subset of validators or sequencers (e.g., in an optimistic or zk-rollup bridge) colludes to artificially inflate latency or censor transactions. This forces users towards higher-fee, instant liquidity solutions that the cartel also operates, creating a circular revenue stream.\n- Attack Vector: Censorship & Latency Manipulation\n- Real-World Parallel: Miner Extractable Value (MEV) on Ethereum, applied to inter-chain messaging.

Governance intentionally deprioritizes development of native fast-withdrawal pools to protect the business model of third-party liquidity providers (LPs) who pay a share of their profits as a protocol fee. This creates a permanent tax on user urgency.\n- Attack Vector: Strategic Roadmap Stagnation\n- Example: A bridge with $1B+ TVL choosing not to deploy its own capital for instant settlements to protect LP fee revenue.

Core developers, influenced by fee revenue stakeholders, weaken cryptographic security assumptions or delay upgrades to more efficient proof systems (like moving from MPC to ZK). This maintains higher operational costs, which are passed to users as fees, justifying the bridge's premium.\n- Attack Vector: Deliberate Technical Debt\n- Mechanism: Rejecting upgrades like zk-SNARKs for light clients to preserve higher-cost, fee-generating guardian node models.

Governance delegates from major bridges (LayerZero, Wormhole, Axelar) form a tacit agreement to avoid competing on fee reduction. They instead compete on marketing and integration, creating an oligopoly with inflated, sticky pricing. Cross-chain dApps are forced to accept high fees as a cost of business.\n- Attack Vector: Soft Price Fixing\n- Evidence: Consistently high 10-30 bps fees across major bridges despite vastly different technical architectures.

Governance allocates a disproportionately small treasury for security audits and formal verification while promoting a large public bug bounty. This creates a perverse incentive for whitehats to delay disclosure, allowing exploitable code to remain in production longer and generate more fee-based revenue before a fix is forced.\n- Attack Vector: Underfunding Proactive Security\n- Result: Critical vulnerabilities live in production for months, with fees accruing the entire time.

Tokenholders vote against fee reduction proposals because their staking yields are directly funded by bridge revenue. Reducing fees collapses the token's cash-flow valuation, creating a direct conflict between user experience and tokenholder profit.\n- Attack Vector: Misaligned Tokenomics\n- Universal Symptom: Governance tokens for bridges and L2s consistently vote for maximizing sequencer/validator revenue over user cost savings.

The core flaw is paying the security providers from the fees they are meant to secure. The solution is to separate the fee beneficiary from the signers.

• Fee diversion to a DAO treasury creates a buffer, but governance must still manage it.
• Fee burning (e.g., via buyback-and-burn) permanently removes the incentive, but reduces protocol sustainability.
• Third-party fee auction models, like those used by Across and CowSwap, let external solvers compete, separating execution from validation.

Move from transaction-based bridging (where the bridge is the asset custodian) to intent-based filling (where the user declares a desired outcome). This fundamentally alters the fee model.

• UniswapX and CowSwap use solvers who post bonds and compete on price; fees are for execution, not custodianship.
• Across uses a bonded relayer model with on-chain fraud proofs; fees reward filling speed, not securing a vault.
• This shifts risk from a monolithic bridge treasury to distributed, slashed actors, removing the massive honeypot.

Use a shared security layer like EigenLayer to restake ETH and provide cryptoeconomic security for the bridge's validation. This separates the bridge's governance (fee setting, upgrades) from its underlying crypto-economic security.

• Bridge validators become Actively Validated Services (AVSs) slashed by the restaked ETH pool.
• Governance can no longer dilute security for profit because the stake is independently managed and slashable.
• This creates a market for security where bridges rent safety, aligning validator incentives with the restaking pool's reputation.

For existing monolithic bridges, impose hard governance limits on fee extraction until the system is sufficiently decentralized. This is a transitional mitigation.

• Fee caps as a % of TVL prevent runaway extraction (e.g., max 0.5% annual revenue from fees).
• Time-locked governance upgrades require delays for major parameter changes, allowing community reaction.
• Transparency mandates force real-time reporting of fee revenue vs. security spend, creating public accountability pressure.