What Is an Automated Market Maker (AMM)?

An automated market maker prices trades algorithmically from the reserves in a liquidity pool, rather than matching buyers against sellers. The classic design uses a constant product formula — x × y = k — where the product of the two reserve balances must stay constant through every trade.

The consequence is that price emerges from the ratio of reserves. Buying token X reduces its reserve and increases token Y's, which raises X's price for the next buyer. Large trades move price more than small ones, automatically and without any market maker deciding to widen a quote.

This is elegant because it always quotes a price. There is no need for anyone to be actively making markets, no minimum viable order flow, and no counterparty to find. A token can have a functioning market from the moment someone seeds a pool.

It is also structurally worse than an order book in specific ways. Price impact on large trades is significant, because the formula guarantees it. AMM prices lag centralised markets, so arbitrageurs continuously extract value from liquidity providers to keep them aligned — that extraction is one way of describing impermanent loss. And traders have no ability to place resting limit orders at a chosen price.

Newer designs mitigate some of this: concentrated liquidity lets providers allocate capital within a chosen price range for far better efficiency, and stablecoin-optimised curves reduce slippage between assets expected to trade near parity. Notably, Hyperliquid uses a full on-chain order book rather than an AMM, which is why its execution behaves like a centralised venue.