Why does stopping MEV still require trusting a committee not to peek?

Every production encrypted mempool protects pending transactions behind a threshold decryption committee. If enough members collude they can reconstruct keys and front-run the same transactions the protocol was built to hide. Unlike a consensus fork, early decryption by a corrupt committee leaves no on-chain evidence and cannot be slashed or even detected. A 2026 SoK paper systematising the entire field confirms this trust assumption persists across every deployed scheme. The only published alternative, DecryptChain, replaces the committee with proof-of-work, but introduces new latency tradeoffs and has not shipped on any production chain.