Aptos Activates Encrypted Mempool to Reduce MEV

Aptos has activated its Encrypted Mempool on mainnet, introducing an optional protocol-level mechanism that conceals transaction intent until ordering has been finalized. According to the official Aptos mainnet announcement, AIP-144 encrypts transaction payloads before they reach validators, limiting the information available for frontrunning and sandwich attacks while a transaction remains pending. The feature went live on October 12, 2026.

Client software encrypts the Move function and its arguments using a key generated collectively by the validator set for each epoch. Validators then order the ciphertext without knowing the underlying operation, before reconstructing the decryption key after consensus fixes the block. No individual validator or validator group below the required stake threshold can independently reveal the pending payload, keeping the trust boundary within the same validator set already responsible for Aptos consensus.

Encrypted Mempool Protects Intent, Not Identity

The first implementation deliberately leaves some information public. The sending address and gas-related fields remain visible, while the transaction’s executable payload stays encrypted until ordering is complete. After decryption, execution proceeds normally and the transaction becomes part of the public ledger. Encrypted Mempool therefore provides pre-execution confidentiality rather than anonymity or permanent transaction privacy.

Petra is the first wallet to integrate the option directly into its signing flow, while developers can submit encrypted transactions through the Aptos TypeScript SDK version 7.0.0 or later. Existing Move contracts require no modification. Decibel also plans to use the system for perpetual-futures orders, where exposing order parameters before execution can create particularly direct MEV opportunities. Mainnet availability nevertheless remains separate from adoption, because users and applications must explicitly opt into encryption on individual transactions.

The system introduces a measurable performance cost. Aptos reports roughly 20 milliseconds of additional latency per block in 50-node benchmarks against its production AptosBFT implementation. AIP-144 also notes that encrypted transactions have a lower maximum throughput ceiling than ordinary transactions, although blocks without encrypted payloads skip the additional decryption work. The trade-off is therefore targeted MEV protection in exchange for additional cryptographic processing rather than a free improvement to baseline network performance.

Aptos Builds Privacy Across Separate Transaction Layers

Encrypted Mempool joins Confidential APT, which had already introduced a different form of selective privacy. Confidential APT hides balances and transfer amounts while leaving wallet addresses visible, using zero-knowledge proofs to preserve transaction validity. The two mechanisms protect different information at different stages: Confidential APT conceals financial values after settlement, while Encrypted Mempool temporarily conceals transaction intent before execution.

That distinction also separates privacy from performance infrastructure. Aptos has simultaneously pursued shorter block times through its Archon execution roadmap, but faster block production and encrypted transaction ordering solve different problems. Lower latency can shorten the execution window, whereas encrypted mempool changes what validators and observers can learn during that window. Neither capability by itself establishes institutional adoption.

The security model also retains limitations. AIP-144 identifies transaction simulation as one potential privacy boundary because simulating a transaction in plaintext through the same full node later receiving its encrypted version could permit timing correlation. The specification proposes safer defaults and separate simulation infrastructure as possible mitigations. Protocol-native encryption reduces one important source of information leakage without making every stage of transaction creation private.

Aptos now has an operational stack covering concealed amounts, optional pre-execution payload encryption and alternative account-authentication mechanisms. The relevant next metric is usage: how many transactions adopt encrypted submission, which trading applications integrate it and whether measurable frontrunning or sandwich extraction declines as adoption grows. Until those data emerge, AIP-144 establishes functioning MEV-protection infrastructure rather than proof that MEV has been eliminated from Aptos.