Prop AMMs move on-chain order flow toward venues that quote well at the moment a trade lands, and toward chains where a router can check every quote inside the trade itself. On those chains, best execution becomes provable for each trade. Wallets and aggregators then choose which venues a router calls, so they decide where the flow goes.
What is a prop AMM?
A prop AMM (proprietary automated market maker) is an on-chain venue where one operator quotes from its own inventory and reprices with cheap update transactions, without waiting for a trade. Its pricing logic is private, and it can reprice before a trade arrives, which a passive pool cannot do.
The model is spreading fast. On Monad, the prop AMM share of MON/USDC volume against AMMs rose from 12.5% in the first three months after the first prop AMM launched in April 2026 to 45.7% in the three months to August 2026, according to a study by researchers at ETH Zurich and Category Labs published on 29 September 2026 and summarised in a Category Labs post on 14 October 2026.
Why are prop AMMs hard to route on some chains?
Their quotes change often, and a router that fixes the route before the trade lands carries that risk. Most aggregators read venue prices from the state at the end of the previous block, simulate the trade, and send a transaction with the route already chosen. If a venue has repriced between the read and the trade, one of two things happens. The venue refuses the stale fill and reverts, or it fills at a worse price than it showed.
On Solana, a revert fails everything. The Solana docs state that if any instruction fails, the entire transaction fails and all state changes revert (Solana documentation). A router cannot try one venue, catch the failure and move to the next inside the same transaction. A venue that reverts often turns into failed swaps for the router's users, so routers rank it down or stop calling it. Solana does have on-chain routers that resolve price at execution (Jump Crypto, April 2026). Their limit sits in failure handling: one failed leg sinks the whole swap.
The second outcome costs the user directly. The same study decoded the largest prop AMM on Base. When its default fee changed between the quote an aggregator read and the trade, traders received less than quoted on 94% of the affected volume (arXiv, September 2026). The fee rose early in the block and fell at the end, so routers read the low fee and filled at the high one.
How do partial reverts make best execution provable?
A partial revert lets one call inside a transaction fail while the rest of the transaction keeps running. On the EVM, a contract can call another contract and catch the failure: the failed call's state changes revert and the caller continues (Solidity documentation). Monad has full EVM bytecode compatibility (Monad documentation). So a router on Monad can do this inside a single swap:
- Ask each integrated venue for its live quote at execution.
- Skip any venue that reverts or quotes below the user's limit.
- Settle against the best live quote.
- Log every venue's quote as an event in the same transaction.
Step 4 creates the record. Anyone can read which venues quoted, at what price, and which one filled. When a quote has gone stale, the router skips that venue and the user does not pay for the gap. The study's authors reach the same conclusion: a router that queries each venue inside the swap transaction would route each trade on the price it actually receives.
The SEC proposed rescinding Rule 611, the trade-through rule, on 11 June 2026, and kept broker best-execution duties in place (Sidley Austin, June 2026). The X account SolanaLegend argues that the end state is a global best bid and offer, where brokers must show the best executable quote across venues, on-chain ones included (X post, 30 August 2026). Under that standard, the deciding test for a chain is evidence. The chain that can show, at the moment of each fill, that no integrated venue offered a better price has the strongest claim on routed flow. Deep liquidity helps a chain pass that test, and the proof is what a broker can hand to a regulator.
Which venues win the flow?
Venues that quote tight at execution win it. When a router compares live quotes at settlement, the end-of-block number no longer sets the route, so padding quotes or moving fees inside the block stops paying. In practice:
- Prop AMMs with cheap repricing and fills at the quoted price take flow at their real spread.
- Passive AMM pools keep long-tail pairs and the sizes prop AMMs do not quote.
- Venues that change fees or spreads between read and fill lose the comparison on every trade.
- Venues that no router calls get no flow, whatever their price.
Why do wallets and aggregators become the gatekeepers?
A venue only competes in trades whose router calls it. The aggregator decides which venues have adapters. The wallet or trading app decides which aggregator fills its users. A user who taps "swap" never picks a venue, so one routing decision by a wallet sets the venue list for every user of that wallet.
That gives both groups pricing power over venues and a measurable claim to sell. A wallet can pick the router with the best measured fills and show users the proof per trade. An aggregator that settles on live quotes can publish quote-against-fill data that a router working from the previous block cannot match. Market makers and chains will compete for a place in their venue lists.
The same logic shows up across infrastructure sales: the builder above you decides once, early, and rarely revisits it. We cover that pattern in if blockchain lets anyone run a bank, who wins distribution.
What should a team do to win routing?
Use this checklist. Each line has a test you can run this quarter.
| Team | Action | How to check it |
|---|---|---|
| Market maker | Expose a standard on-chain quote function any router can call inside a transaction | A third-party router calls it in a test swap |
| Market maker | Fill at the quoted price within the block | Publish quote-against-fill data per pair each month |
| Market maker | Get an adapter merged into each major router on your chain | Count routers that call you; target the ones wallets use |
| Market maker | Keep update transactions much cheaper than swaps | Compare median update gas with median swap gas |
| Aggregator | Choose the venue at execution and log every quote | Each swap transaction shows all venue quotes |
| Wallet or app | Pick the router by measured fills | Track your users' fills against the best logged quote |
| Chain | Confirm routers can catch a failed venue call and continue | A test router survives a reverting venue |
| Chain | Fund adapters so new venues reach routers in weeks | Days from venue launch to first routed trade |
| Chain | Introduce wallets and trading apps to routers that settle on live quotes | Share of chain swap volume settled on live quotes |
The list of buyers is short: the routers and wallets on your chain, each of which makes its venue decision with a small team. Winning routing is a sales job aimed at those teams, timed to their integration cycles. Our notes on how blockchain infrastructure protocols grow market share cover the sequence.
To map the routers, wallets and venues that decide your flow, start with The Diagnostic at $2,500.