Key Highlights
- Aptos said its encrypted mempool is live on mainnet, allowing users to conceal transaction payloads before execution.
- The system uses threshold encryption, with validators collectively decrypting transactions after consensus.
- Sender addresses and transaction fees remain visible, while transaction functions and arguments stay encrypted until finalization.
Aptos, a Layer-1 blockchain platform, has launched an encrypted mempool on its mainnet, aiming to reduce front-running and other forms of maximal extractable value (MEV).
In an October 12 announcement, Aptos said the optional feature encrypts transaction payloads before execution while leaving sender addresses and fees visible. The system uses protocol-level encryption rather than relying on private transaction builders, hardware enclaves, or a separate committee.
How Aptos’ encrypted mempool works
A user’s device encrypts the transaction payload using a key Aptos validators jointly generate for each epoch. The encrypted transaction then enters the mempool, where it can be ordered without revealing its contents.
The block leader can arrange the transaction without seeing the function being called or the arguments included in the request. Once consensus finalizes the block, validators collectively decrypt the transaction so it can execute. Aptos said no single validator can decrypt a pending transaction. Decryption requires the same stake-weighted threshold used to reach consensus on blocks.
The design is specified in Aptos Improvement Proposal AIP-144. According to Aptos, existing Move smart contracts don’t need modification to use the feature.
Encryption targets front-running and MEV
On public blockchains, pending transactions can reveal details such as the assets being traded, transaction amounts, and the contracts involved. Bots can use this information to submit competing transactions or attempt to profit from a user’s planned trade.
By concealing the payload until the block is finalized, Aptos aims to limit the information available to bots before execution. This could make it harder for them to copy a trading strategy or act on a transaction’s contents before it is confirmed.
The feature does not provide complete transaction anonymity. Sender addresses and fees remain visible at submission, and transaction data becomes available on-chain after finalization. It therefore provides pre-execution confidentiality rather than permanent privacy.
Aptos reports 20ms of additional latency
Aptos said its batched threshold-encryption system added approximately 20 milliseconds of latency per block in benchmarks involving 50 nodes, compared with its production AptosBFT consensus implementation.
The company said blocks finalize in less than a second. The full cryptographic construction is described in a research paper submitted to the ACM Conference on Computer and Communications Security (ACM CCS).
Aptos also said the implementation underwent an independent security review by ZKSecurity. The announcement did not detail the review’s findings.
Petra integrates the feature as Decibel plans adoption
Aptos said Petra is the first wallet to integrate the encrypted mempool into its transaction-signing flow. Users can choose to encrypt a transaction when signing it. The company also identified Decibel, a perpetual futures exchange built on Aptos, as a planned user of the feature. Decibel intends to keep orders private until they execute.
The Aptos TypeScript SDK, from version 7.0.0 onward, can retrieve the current epoch’s encryption key, encrypt transaction payloads on the client side, and submit them to the network.
Wider adoption will depend on support from wallets and applications and whether users choose to enable encryption for their transactions.
How Aptos’ approach differs from other privacy methods
Aptos compared its protocol-native design with three other approaches to shielding pending transactions:
- Private builders and RPC services: These can hide transactions from the public mempool, but the service handling the transaction can see its contents.
- Trusted execution environments (TEEs): These rely on hardware enclaves and the systems operating them to preserve confidentiality.
- External threshold committees: These distribute decryption authority among a separate group rather than the blockchain’s own validators.
Aptos said its approach uses the validator set already responsible for consensus, removing the need for a separate decryption committee. The design still relies on the network’s consensus assumptions.
Ethereum also exploring cryptographic privacy
Other blockchain developers are also exploring ways to protect transaction privacy while managing computational costs.
On September 9, Ethereum co-founder Vitalik Buterin proposed recursive STARK mempools through EIP-8288. The proposal explores moving computationally intensive work outside Ethereum’s execution layer to reduce the cost of privacy and post-quantum cryptography.
Ethereum’s proposal takes a different technical approach, but both efforts address how blockchain networks could add cryptographic protections while managing performance costs.
Aptos is also developing faster execution with MonoMove
Aptos is separately working to improve smart-contract execution performance through MonoMove, a new execution engine for its Move virtual machine.
On October 6, Aptos said MonoMove delivered execution improvements of up to 55x in internal tests using real DecibelTrade transactions. A mainnet rollout is planned for 2027, subject to a governance vote.
MonoMove focuses on execution speed, while the encrypted mempool focuses on concealing transaction payloads before execution.
What Aptos users need to know
The encrypted mempool is available on Aptos mainnet through supported applications. Petra offers the first wallet integration, while Decibel plans to use the feature for order privacy.
Its broader impact will depend on support from additional wallets and applications and how frequently users choose to encrypt their transactions.
Also Read: FSB Flags Stablecoin and AI Risks Ahead of G20 Meeting