Aptos floats native encrypted mempool to reduce frontrunning on trades

Targeting pre-trade information leakage

Aptos argues that on most blockchains, pending transactions are exposed before they are finalized, giving validators and other participants a chance to copy trades, reorder them, or trade ahead of users. In earlier material published in May, the company framed protocol-level protection as a way to reduce orderflow leakage, censorship, and frontrunning for large onchain transactions.

Under the design described by Aptos, users would be able to choose a protected submission mode with a single click. Validators would only reveal the encrypted transaction instructions after ordering has already been completed, so the transaction could then be executed without exposing its contents during the most sensitive stage.

Aptos has presented this as especially relevant for exchange activity, including platforms such as Decibel. Its position is that the key privacy benefit is confidentiality before execution, not permanent secrecy, because final transaction data would still be recorded publicly onchain.

How the proposed system works

In the proposed setup, validators would share control of the decryption key during each epoch, when the validator set remains fixed. A stake-weighted threshold of validators would need to cooperate to reveal encrypted transaction payloads.

Aptos says the system is designed to decrypt transactions in batches rather than one by one. According to the company, batching reduces processing demands and allows much of the work to happen while validators are already voting, which is meant to limit the extra burden on transaction flow.

The company has said the remaining online phase of decryption takes less than 20 milliseconds per batch. That figure refers to one part of the process rather than the full end-to-end transaction timeline.

Research claims modest latency tradeoff

A later research paper by Rex Fernando, Guru-Vamsi Policharla, Andrei Tonkikh and Zhuolun Xiang describes TrX, an approach that combines encrypted mempools with a high-performance Byzantine fault-tolerant consensus protocol. The preprint was published in November 2025.

The authors argue that visible transaction contents before ordering are what enable frontrunning, sandwich attacks, and other MEV strategies. Their proposed system is intended to preserve confidentiality during ordering while still fitting within a high-throughput consensus design.

In tests cited by Aptos, the researchers reported an added 27 milliseconds from proposal to execution, which they said was equivalent to a 14% increase over their baseline. They characterized that result as comparable with leading consensus systems, although that assessment depends on the model and comparisons presented in the paper.

Part of a wider Aptos trading push

The encrypted mempool proposal comes alongside a broader Aptos spending and product push. Aptos Foundation and Aptos Labs have committed more than $50 million to first-party products, research, protocol infrastructure, and a strategic fund focused on trading and AI partners.

Among the projects highlighted by Aptos is Decibel, an onchain order book and perpetual futures exchange that went live on Aptos mainnet in February. By May, Aptos said cumulative trading volume on Decibel had surpassed $1 billion. The network has described onchain recording of orders, matches, and cancellations as making confidentiality around trading intent particularly important for professional trading activity.

Aptos has also pointed to growth in other areas, saying stablecoin market capitalization reached a peak of $1.93 billion, nearly ten times late-2024 levels, while real-world assets reached $1.2 billion. In separate product efforts, the network has highlighted Shelby for AI-related dataset storage and said it launched Confidential APT on April 24 to conceal transfer data while keeping transactions verifiable.

What happens next

The encrypted mempool is still a proposal rather than a deployed network feature. Aptos has made clear that its claim of becoming the first layer-1 chain to offer a native encrypted mempool depends on governance approval.

If approved, the feature would give Aptos users the option to send transactions through a protected path intended to hide trade details until ordering is complete. Based on the company’s description so far, the next confirmed step is the governance process that would determine whether the design moves from research and proposal into native network functionality.