Automated market makers (AMMs) are the engine behind most decentralized exchanges. They let anyone swap tokens against a pool of liquidity without an order book. But the classic AMM design has a well-known weakness: it is often the last to learn about a price change, and arbitrage traders profit from that delay at the expense of liquidity providers.
A newer approach, often called a propAMM or oracle-priced AMM, tries to close that gap. On October 22, 2026, Curve Finance put a "permissionless propAMM" up for a DAO vote, which makes it a good real-world example. This guide explains the idea in plain English.
How a classic AMM sets prices
In a standard AMM, the price comes from the ratio of the two tokens in the pool. If a pool holds ETH and a stablecoin, buying ETH removes ETH from the pool and adds stablecoins, which pushes the ETH price up inside that pool.
The pool itself does not know what ETH is worth on other exchanges. It only knows its own balances. When the market price moves elsewhere, the pool's price stays where it was until someone trades against it.
Why arbitrage costs liquidity providers money
That someone is usually an arbitrage trader. Suppose ETH rises on major exchanges, but the pool still quotes the old, lower price. An arbitrageur buys cheap ETH from the pool and sells it elsewhere at the higher price, until the pool's price catches up.
The arbitrageur keeps the difference. The liquidity providers are left with less ETH and more stablecoins, sold at a stale price. Researchers often call this cost loss-versus-rebalancing (LVR). It is one reason why providing liquidity can be less profitable than it first appears, even when trading fees look attractive.
Arbitrage is not bad in itself. It keeps prices aligned across markets, which helps every user. The question is who captures the value: the arbitrage trader or the pool.
What a propAMM does differently
A propAMM, or oracle-priced AMM, uses an external price feed (an oracle) to guide its pricing. Instead of waiting for trades to move the price, the pool updates based on fresh market data.
The goal is simple:
- Quote tighter prices to ordinary traders, because the pool is more confident about the current market price.
- Reduce the amount of value that leaks out to arbitrage when the market moves.
- Turn more of that value into revenue for the pool and its liquidity providers.
The trade-off is that the pool now depends on the oracle. If the price feed is wrong, delayed or manipulated, the pool can misprice trades. Good designs add safeguards such as limits on how fast prices can change and the ability to switch the system off in an emergency.
Curve's proposal as a real example
According to Curve's announcement, on-chain proposal #1508 is open for voting on the Curve DAO through October 29, 2026. It would attach a contract called YBOraclizedPolicy to the YieldBasis crvUSD/WETH pool. Swaps would be priced using signed Pyth Pro ETH/USD price reports rather than pool balances alone.
The main parameters described in the proposal:
- Swaps that move toward a fresh price report, without crossing it, pay a fee of 50 to 175 basis points, with the discount fading over five minutes.
- Other operations pay 175 basis points.
- The pool's internal price scale follows trend-gated 80-minute and 160-minute averages of the reports, and can move by at most 50 basis points per block.
- The Yield Basis DAO owns the policy, and Curve's emergency DAO can switch it off. Votes on both sides are needed for the change to take full effect.
Curve says an informed pool can offer much tighter spreads and keep a larger share of mispricing as pool revenue instead of arbitrage profit. According to the forum post, the pool has been earning less than its re-leverage costs in low-volatility conditions, and this change is designed to address that.
What this means for everyday swappers
Most users will never interact with a propAMM directly by name. They will simply see a quote in a DEX or aggregator interface. Still, understanding the mechanics helps you read those quotes better:
- A pool with oracle pricing may offer tighter prices on popular pairs such as ETH and stablecoins.
- Fees can vary depending on the direction of your trade and how recent the price data is.
- Oracle-based designs add a new dependency, so it is worth knowing which price feed a protocol relies on.
- Aggregators and routers increasingly compare many pool types at once, so the best route may change from one swap to the next.
What this means for liquidity providers
If you provide liquidity, the propAMM idea speaks directly to your returns. Before adding funds to any pool, ask a few questions:
- How does the pool set its price: balances only, or an external oracle?
- Who controls key parameters, and is there an emergency shutdown?
- What happens to fees during quiet versus volatile markets?
- Has the code been audited, and how long has it been live?
Key takeaways
- Classic AMMs price trades from pool balances, which lets arbitrage traders profit when outside prices move.
- That arbitrage profit is a cost to liquidity providers, often called loss-versus-rebalancing.
- propAMMs use oracle prices to quote more accurately and keep more value in the pool.
- The main new risk is reliance on the oracle, which good designs limit with caps and emergency controls.
- Curve's vote on proposal #1508 runs through October 29, 2026, and is an early real-world test of the idea.
Whether you use Curve, Uniswap or a multi-network swap service such as AviorSwap, which lists 17+ networks, the same habits apply: compare the final amount you receive, set slippage carefully and double-check the network before you confirm.
This article is for educational purposes only and is not financial advice.