Before Uniswap launched in November 2018, swapping one crypto token for another without a centralized exchange was a hassle. You either had to track down a willing counterparty or wrestle with a clunky order book on an early DEX that had almost no volume.
Uniswap changed that with one remarkably simple idea. Ditch the order book entirely, and replace it with a math formula and a shared pool of tokens that anyone could deposit into.
That idea is now called an automated market maker, or AMM. It's since settled more than $2 trillion in trades and spawned an entire category of decentralized finance.
Understanding how AMMs actually work isn't just academic.
If you've ever swapped tokens on a DEX, provided liquidity, or wondered why your trade came back slightly worse than the quoted price, you were already interacting with this mechanism. This explainer breaks down the math, the incentives, and the tradeoffs from the ground up.
TL;DR
- An AMM replaces a traditional order book with a liquidity pool and a pricing formula, so trades execute automatically without a matching counterpart.
- The core formula, x × y = k, keeps the product of two token reserves constant, which means buying one token automatically raises its price within the pool.
- Liquidity providers earn a share of trading fees but face impermanent loss when token prices diverge significantly.
- Concentrated liquidity (introduced by Uniswap v3) lets providers focus capital in a chosen price range, improving capital efficiency by up to 4,000x versus the original model.
- Understanding AMM mechanics helps traders estimate slippage, pick better swap routes, and evaluate the real risk of providing liquidity.
What An Automated Market Maker Actually Is
A traditional exchange matches buyers and sellers. Someone posts an offer to sell Bitcoin (BTC) at $65,000, someone else bids $65,000, and the exchange pairs them. An AMM eliminates the matching step entirely. Instead of waiting for a counterparty, a trader swaps against a pool of tokens that sits in a smart contract. The smart contract sets the price algorithmically, executes the swap immediately, and adjusts the pool's composition to reflect the trade.
The pool holds two assets, typically in pairs like Ethereum (ETH) and USD Coin (USD Coin (USDC)). When you swap ETH for USDC, you are adding ETH to the pool and removing USDC from it. The more ETH you push in relative to what is already there, the worse your exchange rate gets. This self-adjusting pricing is what replaces the order book.
Key idea: An AMM is a smart contract that always quotes a price, always accepts a trade, and recalculates the next price after every transaction. It never sleeps, never refuses, and never requires a matching seller.
The term "automated market maker" borrows from traditional finance, where a market maker is a firm that continuously quotes buy and sell prices to keep markets liquid. AMMs achieve the same result through code and collective liquidity instead of institutional capital.
Also Read: https://yellow.com/news/record-207-crypto-hacks-expose-weak-spot-in-digital-asset-security
The Constant Product Formula, x Times y Equals k
Every original AMM design, including Uniswap v1 and v2, is built on one equation: x × y = k.
Here x is the reserve of token A in the pool, y is the reserve of token B, and k is a constant that must stay the same after every trade.
When you buy token B, you add token A to the pool and remove token B. That means x goes up and y goes down. For the product to remain k, the price you pay per unit of token B increases the more of it you try to take out.
A concrete example makes this intuitive. Imagine a pool with 100 ETH and 200,000 USDC, so k = 100 × 200,000 = 20,000,000. You want to buy 10 ETH.
After your trade, the pool will have 90 ETH. For the product to stay at 20,000,000, the USDC reserve must become 20,000,000 / 90 = 222,222 USDC. You had to add 22,222 USDC to take out 10 ETH, giving you an effective price of $2,222 per ETH. If the pool had been thinner or if you had tried to buy 50 ETH, the price would have been drastically worse. This is slippage, and it comes directly from the curve.
Why k stays constant: k only changes when liquidity is added or removed, not during normal swaps. Fees are the one subtle exception. In practice, a small fee (0.30% on Uniswap v2) is taken from each trade and added back to the reserves, which means k grows very slightly over time as fees accumulate.
Also Read: https://yellow.com/news/xrp-tests-107-bulls
How Liquidity Providers Fund The Pool
The pool does not appear from nothing. It is funded by liquidity providers (LPs), who deposit equal values of both tokens into the pool in exchange for LP tokens. Those LP tokens represent a proportional share of the pool. When an LP wants their funds back, they burn the LP tokens and receive their share of the current reserves, including any fees that have accumulated.
The fee revenue is the incentive.
On Uniswap v2, a 0.30% fee is charged on every swap and distributed proportionally to all LPs. If a pool processes $10 million in daily volume, it generates $30,000 in fees per day, split among all depositors according to their pool share. High-volume pairs like ETH/USDC generate enough fee income to make LP positions genuinely attractive.
The risk LPs face is called impermanent loss. When token prices diverge from the ratio at the time of deposit, the pool's automatic rebalancing works against the LP relative to simply holding both tokens. If ETH doubles in price, arbitrageurs will buy ETH from the pool until the price reflects the wider market, leaving the pool with less ETH and more USDC than when the LP entered. The LP ends up with a portfolio worth less than if they had just held the tokens outright. The loss is "impermanent" because if prices return to the original ratio, it disappears, but if they do not, it becomes permanent when the LP withdraws.
Also Read: https://yellow.com/news/crowdstrike-claude-mythos-zero-day
Arbitrage Keeps Prices Honest
The AMM formula sets prices algorithmically, but those prices must stay close to prices on other exchanges or the pool becomes a free-money machine. That alignment is maintained entirely by arbitrageurs.
If ETH is trading at $2,000 on Coinbase but the Uniswap pool implies $1,950 because of a large recent sell, arbitrageurs will buy ETH from the Uniswap pool and sell it on Coinbase.
They keep doing this until the pool price rises to match the market. At that point, the profit opportunity disappears and the prices are in sync again.
This constant arbitrage activity means AMM pools act as price-discovery mechanisms even though the formula itself has no awareness of external markets. The arbitrageurs are the feedback loop. They profit from the price discrepancy but in doing so they restore accuracy. LPs, however, bear the cost of this correction in the form of impermanent loss. Every time an arbitrageur corrects the price, they are effectively buying low from the pool and leaving the LPs with less of the appreciating asset.
The arbitrage loop: External price moves → AMM price lags → arbitrageurs buy cheap → pool price corrects → LPs absorb the difference. This cycle runs continuously across every pool on every DEX.
Also Read: https://yellow.com/news/meta-watermelon-catches-gpt55
Concentrated Liquidity And The Uniswap V3 Breakthrough
The original constant product formula spreads liquidity across every possible price from zero to infinity. In practice, most trading happens in a narrow band around the current market price. Liquidity sitting at price points far from the current rate earns almost no fees because almost no trades happen there.
Uniswap v3, launched in May 2021, introduced concentrated liquidity to fix this. Instead of depositing across the full price curve, LPs now choose a specific price range. All of their capital works only within that range and earns fees only for trades that happen within it.
When the price moves outside their chosen range, their position stops earning fees and becomes entirely one asset.
The capital efficiency gains are dramatic. Uniswap's own analysis showed that an LP providing liquidity in a 0.10% price range on an ETH/USDC pool could achieve up to 4,000 times the fee revenue of a v2 position of the same size, assuming price stays in range. This made concentrated liquidity positions resemble limit orders with fee income, turning passive LP positions into active strategies.
The tradeoff is increased complexity. An LP in v2 can deposit and ignore the position for months. A v3 LP must choose a range, monitor whether price stays in range, and rebalance when it drifts out. This complexity opened the door for active liquidity management protocols that automatically rebalance positions on behalf of passive LPs.
Also Read: Hollywood Director Gets 30 Months After Crypto And Stock Bets
Slippage, Price Impact, And Why Pool Depth Matters
Every AMM trader encounters slippage. It is the difference between the price shown when you initiate a swap and the price at which the trade actually executes. On AMMs, slippage has two sources: price impact and front-running.
Price impact is structural. The constant product formula guarantees that larger trades move the price more. On a deep pool like ETH/USDC with $500 million in liquidity, a $10,000 swap causes barely perceptible price movement. On a thin pool with $50,000 in liquidity, the same $10,000 swap could move the price by 10% or more. Traders use slippage tolerance settings, typically 0.5% to 1.0%, to set the maximum acceptable deviation. If the price moves more than that tolerance between submission and execution, the transaction reverts.
Front-running is the second source and comes from how blockchains work.
All pending transactions on Ethereum (ETH) sit in a public mempool before miners or validators include them in a block. Bots monitor this mempool for large swap transactions and insert their own trades ahead of them, moving the price slightly before the original trade executes, and then unwinding immediately after. This is a form of MEV (maximal extractable value) called a sandwich attack. Slippage tolerance protects against this to a degree, since a sandwich attack that moves the price beyond the tolerance will cause the original transaction to revert.
Pool depth directly determines how vulnerable a pair is to both forms of slippage. Deep pools, funded by many LPs attracted by high fee income, protect traders. Shallow pools, common on new or low-volume pairs, require traders to accept wider slippage or trade in smaller chunks.
Also Read: Bitget Becomes First Crypto Exchange To Offer US Stock Options
Who Benefits Most From Understanding AMM Mechanics
Casual swappers can ignore most of this detail most of the time on major pairs with deep liquidity. But several reader types gain a real edge from understanding the mechanics.
Active traders who use DEX aggregators like 1inch or Uniswap's own smart order router benefit from knowing why routing across multiple pools reduces price impact. A $100,000 swap split across three pools each moves the price less than a single $100,000 swap hits one pool. Aggregators exploit this automatically, but knowing why helps traders evaluate whether a quoted route is actually optimal.
Liquidity providers who want to earn fee income need to understand impermanent loss before depositing. On correlated pairs like ETH/stETH, price rarely diverges much, so impermanent loss is small and fees accumulate safely. On uncorrelated or volatile pairs, impermanent loss can exceed fee income. Modeling both before entering a position is essential.
DeFi developers building on top of AMMs, whether yield optimizers, leveraged farming protocols, or structured products, need a precise understanding of pool math to price their products correctly and avoid edge cases that drain user funds.
Token launchers using AMMs to create initial liquidity for new tokens need to understand how pool initialization price and initial deposit size determine the vulnerability of their launch to early arbitrage and sniping.
Also Read: GLM-5.2 Tests OpenAI And Anthropic With Cheaper AI Power
Final Thoughts
The automated market maker ranks among the most consequential inventions in the history of financial infrastructure.
A single equation, x × y = k, replaced the entire market-making apparatus that once required licensed institutions, dedicated capital desks, and regulatory frameworks.
Anyone with two tokens can now create a market for them. Anyone with capital can earn fees by supplying liquidity to that market. And anyone in the world with an internet connection can trade against it.
Uniswap's ongoing volume tells the story. Now consistently above $300 million per day, it shows that the tradeoffs are acceptable to enough participants to sustain a functional market.
AMM designs keep evolving — better fee tier structures, dynamic ranges, MEV-resistant execution.
But the underlying constant product logic that started it all remains the clearest lens through which to understand how decentralized trading actually works.
Read Next: Tron Processes 385.77M Transactions In A Record-Breaking June