# private mempool adoption trades frontrunning protection for centralization and opacity Approximately 10% of Ethereum transactions now route through private mempools — double the 2022 share. Private transactions bypass the public mempool by sending directly to validators or block proposers via services like Flashbots Protect, MEV Blocker, or private RPC endpoints. The protection is genuine: transactions not visible in the public mempool cannot be frontrun by bots monitoring it. The tradeoff is structural. Since [[frontrunning exploits public mempool visibility to insert competing transactions before profitable pending operations]], the public mempool is the attack vector — removing transactions from it removes the attack surface. But the mechanism of protection is trust concentration: transactions are now visible to a small number of privileged intermediaries (relays, builders, block proposers) rather than the entire network. These intermediaries can still frontrun, censor, or extract value from the transactions they see — they just do so without public visibility. Information asymmetry compounds over time. Select searchers and builders gain privileged access to transaction flow, compounding their structural advantage in MEV extraction. The aggregate effect is that MEV extraction becomes less visible (fewer public mempool bots) but not eliminated — it migrates to where the private flows are controlled. Some chains have extended this to entirely private mempools with centralized validator sets, creating single points of trust that undermine DeFi's open-access model. The censorship risk is concrete. Private mempool operators can selectively exclude transactions from specific addresses — a censorship vector that was previously constrained by the public mempool's permissionless visibility. OFAC-compliant relays already demonstrate this in practice, with some operators choosing not to relay transactions to sanctioned addresses. The opacity dimension is distinct from the security dimension. Even if private mempools reduce harmful MEV for individual users, reduced network transparency undermines the ability to measure, audit, or regulate MEV at the system level. Flashbots Protect reports 98.5% success rates in MEV protection — but the protection metric measures user outcomes, not system-level health. --- Relevant Notes: - [[frontrunning exploits public mempool visibility to insert competing transactions before profitable pending operations]] — the attack vector that private mempools address by removing public visibility - [[PBS relay trust is the critical unresolved centralization point in out-of-protocol MEV-Boost architecture]] — the relay infrastructure through which private mempool transactions flow, creating the trust dependency - [[sandwich attacks exploit AMM deterministic pricing to front-run and back-run victim trades within a single block]] — the specific attack class private mempools most effectively prevent Topics: - [[vulnerability-patterns]] - [[protocol-mechanics]]