Bots made real money last year by simply watching other people’s trades before they landed. An EigenPhi analysis cited by Cointelegraph put Ethereum sandwich-attack losses near $60 million over the twelve months through October 2025. Blockchain analytics firm Bitquery found a separate $29.7 million drained from Solana traders over 11 months, across roughly 6.9 million attacks hitting about 2.2 million wallets. On October 12, 2026, Aptos turned on a fix for that exact problem: the Aptos Encrypted Mempool. It hides a pending transaction’s contents until the moment validators lock in its place in line, so the bots that used to read your trade and jump in front of it have nothing left to read.
Aptos, a Move-language Layer 1 blockchain, confirmed the rollout in a post on its foundation’s forum and in its developer documentation. The feature, specified under AIP-144, is opt-in. It’s not a new idea industry-wide, Gnosis Chain’s Shutter network got there in 2024, but Aptos says its version works differently, and it’s worth understanding both the mechanics and the limits before deciding it solves the problem for good.
How Aptos’s Encrypted Mempool Hides Your Trade
The whole point of an encrypted mempool is to separate two things that normally travel together: the order transactions get processed in, and what’s actually inside them. Here’s the sequence, based on Aptos’s own documentation:
- A compatible wallet encrypts your transaction’s instructions on your own device before it ever reaches the network.
- You submit it. Validators can see your wallet address and the fee you’re paying, but not which asset you’re trading, how much, or which contract you’re calling.
- Validators agree on the order transactions will execute in, still without seeing what’s inside any of them.
- Once that order is locked in, validators combine partial decryption keys to reveal the transactions together and run them.
No single validator holds the key. Aptos says decryption needs a stake-weighted validator majority, the same trust assumption that already secures its consensus. That’s a deliberate design choice: rather than hand the sensitive job to a separate committee or a private relay, Aptos routes it through validators who are already there doing the ordering work.
What It Costs to Turn Encryption On
Privacy on a blockchain is rarely free, and Aptos’s documentation is specific about the price:
| What changes | Figure | Source |
|---|---|---|
| Minimum gas price for an encrypted transaction | 200 Octas per gas unit, double the network’s base minimum | Aptos developer docs |
| Extra charge per encrypted transaction | 375 gas units added for decryption | Aptos developer docs |
| Added latency per block in a 50-node test | About 20 milliseconds | Aptos Foundation, reported by Cointelegraph |
Aptos’s consensus normally finalizes a block in under a second, so 20 extra milliseconds is a small tax by comparison. The gas costs are the more direct trade: you’re paying a bit more, every time, for every encrypted transaction, to keep bots from reading it. The security firm ZKSecurity reviewed the encryption scheme before launch, which is a reasonable sign of diligence, though an audit reduces risk rather than eliminating it.
Why Aptos Says This Beats Other Privacy Fixes
Aptos isn’t the first chain to try hiding pending trades. Gnosis Chain’s Shutter network runs a threshold-encrypted mempool through an outside decryption committee. Ethereum traders already route sensitive orders through Flashbots Protect, a private channel to block builders that keeps transactions off the public mempool entirely. Ethereum also has its own encrypted-mempool proposal on the table, a draft called EIP-8184, or LUCID, which this site covered back in August; it’s still working through open cryptography and governance questions and has no activation date.
Aptos’s pitch is that it skips the extra layer. Its existing validators handle both ordering and decryption, so there’s no separate committee to trust and no private builder relationship to negotiate. Whether that’s actually safer than Shutter’s approach is a judgment call people will keep arguing over, but it does mean Aptos didn’t have to wait on anyone else to ship it.
The exchange built on Aptos called Decibel plans to use the feature for private pending orders, and Petra is, so far, the only wallet confirmed to support encryption inside its own signing flow. A broader comparison of the MEV problem across chains is worth a look if you want the fuller picture; The Defiant’s coverage of Shutter’s 2024 launch frames the scale of the problem Aptos is now responding to two years later.
Why It Matters to You
Picture a nurse named Elena, an illustrative example and not a real person, who trades a little on the side during slow shifts using an app built on Aptos. Before this update, a bot watching the public mempool could spot her pending buy order, jump in ahead of it, and sell right after, quietly taking a slice of her money as the price moved against her. If Elena’s wallet supports the new feature, and right now that means Petra, she can opt in, pay a bit more in gas, and that bot loses its view entirely. Nothing changes if she skips it: for everyday traders, this protection is something you have to actively choose, not a default that kicks in on its own.
Where the Protection Stops
- It’s opt-in. Unless a trader or their wallet specifically turns it on, transactions stay exactly as exposed as before.
- Your wallet address and the fee you pay are still visible to anyone watching, even on an encrypted transaction.
- The $29.7 million Bitquery figure describes Solana, not Aptos. No independent tally yet exists for how much MEV, if any, Aptos traders have lost historically or will avoid going forward.
- The 20-millisecond latency figure comes from a 50-validator test. Aptos’s full mainnet validator set is larger, and real-world performance could differ.
- Simulation tools don’t yet work directly on encrypted payloads, and only one wallet has confirmed support so far, so adoption will take time.
Watch two things from here: whether other major wallets beyond Petra add support, and whether anyone publishes real usage numbers once traders start opting in. A feature that protects almost nobody because nobody turns it on isn’t much of a fix. If Aptos can show meaningful adoption within a few months, that’s the signal this was more than a press release.
Sources
- Cointelegraph: Aptos rolls out Encrypted Mempool to curb frontrunning, sandwich attacks (Oct 13, 2026)
- crypto.news: Aptos launches Encrypted Mempool: How does it work? (accessed Oct 19, 2026)
- Aptos Foundation Forum: Aptos Introduces Built-in MEV Protection with Encrypted Mempools (accessed Oct 19, 2026)
- Aptos developer documentation: Encrypted Mempool (TypeScript SDK) (accessed Oct 19, 2026)