Layer-1 blockchain Aptos has launched an optional Encrypted Mempool feature on its mainnet to protect users from maximal extractable value (MEV) attacks, such as frontrunning and sandwich attacks. Announced on Monday, the system allows users to opt in to keep pending transactions encrypted until block ordering is finalized. The feature targets institutions submitting large transactions, traders seeking to avoid MEV risk, AI agents, and developers.
When enabled, a compatible wallet or application encrypts transaction details—including assets and amounts—on the user's device before submission. While the sender’s address and gas fee remain visible, validators agree on the order of transactions without seeing their contents. Once ordering is complete, validators collectively decrypt the transactions for execution and recording on-chain. Aptos stated that no single validator can decrypt a pending transaction independently.
The implementation introduces specific costs and latency. According to Aptos’ developer documentation, encrypted transactions require a minimum gas price at twice the network’s base minimum. Additionally, tests involving 50 nodes indicated the system adds approximately 20 milliseconds per block compared to the existing consensus implementation. This launch follows broader industry efforts to mitigate MEV, with data showing significant losses on other networks: Cointelegraph analysis of EigenPhi data indicated Ethereum traders lost about $60 million to sandwich attacks over the 12 months through October 2025, while Bitquery identified nearly 6.9 million attacks generating $29.7 million for bots on Solana over an 11-month period.
Aptos' introduction of an encrypted mempool represents a structural attempt to neutralize information asymmetry in transaction ordering, a primary vector for MEV extraction. By hiding transaction payloads until after consensus on ordering is reached, the protocol removes the visibility that bots exploit to front-run or sandwich trades. However, this security comes at a tangible economic cost, doubling the minimum gas price for participating users and introducing measurable latency. This trade-off suggests the feature is initially optimized for high-value institutional flows or sensitive AI agent operations where the cost of being exploited exceeds the premium paid for privacy, rather than for general retail micro-transactions.
The move highlights a growing divergence in Layer-1 strategies regarding MEV mitigation. While some networks focus on redistributing MEV profits to validators or users, Aptos is prioritizing prevention through cryptographic opacity during the ordering phase. The success of this approach will depend on whether the added latency and fees deter adoption among smaller traders, potentially creating a two-tier market where only sophisticated actors can afford protection. Furthermore, the reliance on collective decryption by validators introduces new operational risks; if the threshold scheme fails or validators collude, the integrity of the encryption could be compromised, shifting the trust model from purely algorithmic to partially social.
