What an Automated Market Maker Actually Does
An automated market maker, or AMM, is a protocol that lets you trade crypto without waiting for another person to take the opposite side of your trade. It replaces the order book (the list of buy and sell offers you see on Coinbase or Binance) with a pool of tokens and a formula that sets the price automatically.
When you swap ETH for USDC on Uniswap, you are not matched with a seller. You trade directly against a smart contract holding reserves of both tokens. The contract adjusts the price based on how much you take out and how much you put in. No market maker sets the spread. No exchange holds your funds. The math does the work.
The Order Book Problem AMMs Solve
Traditional exchanges, whether crypto or stock, run on order books. You place a limit order to buy 10 ETH at $2,500 each. Someone else offers to sell at that price. The exchange matches you. But order books need liquidity. If nobody wants to sell at your price, your order sits unfilled. If the book is thin, a large trade moves the price sharply before it completes.
Crypto made this worse. Hundreds of token pairs exist, but most have low volume. Maintaining an order book for a niche altcoin costs money. Professional market makers demand fees to provide liquidity on illiquid pairs. Smaller projects cannot afford it, so their tokens trade poorly or not at all.
AMMs flipped the model. Instead of matching buyers and sellers, liquidity providers deposit token pairs into a pool. Anyone can trade against that pool at any time. The protocol guarantees a price for every trade, even on pairs nobody else is actively trading. Slippage still exists, but the market never closes and you never wait for a counterparty.
The Constant Product Formula
Most AMMs, including Uniswap's original version, use the constant product formula: x multiplied by y equals k. The variables x and y represent the quantities of two tokens in the pool. The constant k stays the same after every trade.
Here is a simplified example. A pool holds 10 ETH and 20,000 USDC. Multiply them: 10 times 20,000 equals 200,000. That is k. The implied price is 2,000 USDC per ETH, because 20,000 divided by 10 equals 2,000.
Now you buy 1 ETH. You remove 1 ETH from the pool, leaving 9 ETH. To keep k at 200,000, the USDC side must rise to 22,222. Subtract the original 20,000 USDC, and you paid 2,222 USDC for that 1 ETH. The price increased because you shifted the ratio. The larger your trade relative to the pool, the more the price moves. That movement is slippage.
After your trade, the pool holds 9 ETH and 22,222 USDC. The new implied price is 2,469 USDC per ETH. You moved the market just by trading. The next buyer pays that higher price. If someone sells ETH back into the pool, the price drops again. The formula keeps k constant and lets the price float.
Where Liquidity Comes From
Someone has to supply the tokens in the pool. Those suppliers are called liquidity providers, or LPs. They deposit equal values of both tokens (say, $10,000 worth of ETH and $10,000 worth of USDC) and receive LP tokens representing their share of the pool.
LPs earn a cut of every trade. Uniswap, the largest automated market maker, charges 0.3% per swap, and that fee goes directly to the liquidity providers. High-volume pools generate substantial fee income. In the 30 days ending September 2026, Uniswap collected $208 million in fees. Most of that went to LPs, with $15.79 million captured as protocol revenue.
But LPs face a risk called impermanent loss. When the price ratio between the two tokens changes, the LP's position loses value compared to simply holding the tokens. If ETH doubles in price against USDC, arbitrage traders rebalance the pool by buying the cheap token and selling the expensive one. The LP ends up holding more of the token that went down and less of the token that went up. The loss becomes permanent if they withdraw at that point. Fees offset this loss only when trading volume is high enough.
AMM Variants for Different Assets
The constant product formula works well for volatile pairs like ETH-USDC in most automated market makers. But it performs poorly for stablecoins. Swapping USDC for USDT should cost almost nothing, because both tokens trade near $1. A constant product pool would still charge slippage on large trades.
Curve Finance introduced the stableswap formula, which flattens the price curve for assets that stay close in value. Swapping $100,000 USDC for USDT on Curve incurs minimal slippage because the formula assumes the tokens will remain near parity. When the peg holds, stableswap is far more capital efficient. When it breaks, the formula can leave LPs holding the depegged asset.
Balancer allows pools with more than two tokens and custom weightings. A pool might hold 50% WBTC, 25% ETH, and 25% USDC. The weighting lets LPs build index-like exposure while earning fees. Traders get access to multiple pairs in one transaction.
Uniswap v3 introduced concentrated liquidity. Instead of spreading capital across all possible prices, LPs choose a price range. If ETH trades between $2,000 and $3,000, an LP can concentrate their capital there and earn higher fees per dollar deposited. If the price moves outside that range, the position stops earning until it returns. This increases capital efficiency but adds complexity and requires active management.
Why AMMs Became the Standard for DeFi
AMMs exploded because they removed barriers. You do not need permission to list a token. You do not need a market maker to provide liquidity. You do not need to trust an exchange with custody. You deposit two tokens into a smart contract, and the pair is tradable.
Uniswap alone holds $3.89 billion in total value locked across 49 blockchains as of September 2026. In a single day, it processes $3.42 billion in DEX volume across 5.49 million transactions from 454,174 active addresses. Those numbers reflect real usage, not speculative hype. People trade on AMMs because they work.
The permissionless model matters for automated market makers. Anyone can launch a token and create a liquidity pool in minutes. That openness enables both innovation and scams. Rug pulls thrive on AMMs precisely because listing requires no approval. But the same openness lets legitimate projects bootstrap liquidity without paying listing fees or meeting exchange requirements.
AMMs also enable composability. Because they are smart contracts, other protocols can integrate them directly. A lending platform can use an AMM to liquidate collateral. A yield aggregator can route trades across multiple AMMs to find the best price. Restaking protocols and other DeFi primitives build on top of AMM liquidity without asking permission.
The Security Trade-Offs
AMMs inherit the risks of smart contracts. A bug in the code can drain the pool. The $292 million KelpDAO exploit in 2026 demonstrated that even audited protocols can fail. Once tokens leave the pool, they cannot be recovered unless the attacker chooses to return them.
Front-running is another problem. Because all transactions sit in the public mempool before confirmation, bots can see your trade, calculate the price impact, and insert their own trades before and after yours to profit from the movement. This is called a sandwich attack. MEV extraction costs traders billions annually, and AMMs are the primary target because price movements are predictable and profitable.
Liquidity can also disappear. If LPs withdraw during a market crash, the pool shrinks and slippage explodes. Small pools are especially vulnerable. A token with $50,000 in liquidity might see 20% slippage on a $5,000 trade. Large holders cannot exit without taking severe losses, which traps capital and accelerates declines.
Where AMMs Fit in the Broader DeFi Landscape
AMMs are infrastructure. They do not compete with lending protocols or derivatives platforms. They enable them. When Aave liquidates an undercollateralized loan, it often swaps the collateral through an AMM. When a prediction market settles a bet, it may use an AMM to convert the payout token into stablecoins.
Most AMMs run on Ethereum or its Layer 2 networks, but they have spread to every chain with smart contract support. Robinhood Chain, for example, generated $9.17 million in Uniswap fees over 30 days in 2026, outpacing Ethereum itself at $3.15 million. Base, another Layer 2, contributed $2.18 million. The fragmented L2 economy splits liquidity across many chains, reducing capital efficiency, but AMMs adapt by deploying wherever users trade.
Competition between automated market maker protocols is mostly technical. Uniswap dominates by volume, but Curve owns the stablecoin market. PancakeSwap leads on BNB Chain. SushiSwap carved out a niche with incentivized liquidity mining. Each AMM protocol tweaks the formula, fee structure, or token incentives to attract users and liquidity providers.
What AMMs Cannot Do
AMMs cannot match the capital efficiency of a well-run order book on a high-volume pair. If you want to trade $10 million of BTC-USDT, a centralized exchange with deep liquidity will give you a tighter spread and less slippage than any AMM. Professional traders still use order books for size.
AMMs also cannot eliminate price impact. The constant product formula guarantees that large trades move the price. Concentrated liquidity and larger pools reduce slippage, but they do not remove it. Order books can fill large trades at a single price if enough limit orders sit at that level. AMMs cannot.
They also depend entirely on external price feeds or arbitrage to stay accurate. If the ETH-USDC pool on Uniswap drifts away from the true market price, arbitrage traders buy the cheap side and sell the expensive side until the pool realigns. That arbitrage costs the pool money in the form of adverse selection. LPs effectively pay arbitrageurs to keep prices accurate. When arbitrage is slow or expensive, AMM prices lag.
The Model That Refuses to Die
Every year since 2020, someone declares AMMs obsolete. Order book DEXs will take over. Concentrated liquidity will kill passive LPs. Institutional traders will never use them. And yet AMMs keep growing.
They work because they solve a real problem: how do you let anyone trade anything without permission, intermediaries, or custody risk? The answer is a smart contract, a formula, and a pool. Slippage is the price of decentralization. For millions of users, that trade-off is worth it.