DECLASSIFIED // INTELLIGENCE BRIEFING // FOR EDUCATIONAL PURPOSES ONLY
This content is informational only and does not constitute financial, legal, or investment advice. Always do your own research before making any trading decisions.
AMMs and Liquidity Pools: How Decentralized Exchanges Actually Work
Automated Market Makers explained. How Uniswap, Curve, and Balancer work, the constant product formula, impermanent loss, concentrated liquidity, and how LPs earn fees.
Updated May 14, 2026· CRYPTINT.IO Intelligence
Key Takeaways
- +Automated Market Makers (AMMs) replaced order books with liquidity pools. Instead of matching buyers and sellers, AMMs let traders swap against a pool of assets whose prices adjust based on a fixed mathematical formula.
- +The constant product formula (x * y = k) is the core AMM innovation, introduced by Uniswap V1 in 2018. Token amounts in the pool satisfy the formula; trades move prices along the curve.
- +Liquidity providers (LPs) deposit tokens into pools and earn a share of trading fees. They also bear impermanent loss: the opportunity cost of the pool value diverging from simple holding when prices move.
- +Modern AMMs (Uniswap V3, Uniswap V4 with hooks, Curve V2, Balancer) add concentrated liquidity, dynamic fees, and customizable curves. These extend the basic AMM model to specific use cases.
- +Curve's stable-asset formula is optimized for trading pairs that should trade near 1:1 (USDC/USDT, stETH/ETH). Balancer allows pools with multiple tokens and custom weights. Each AMM design trades flexibility for efficiency in specific scenarios.
Why AMMs Exist
Traditional exchanges (Binance, Coinbase, NYSE) use order books. Buyers post bids at specific prices. Sellers post offers. Trades execute when bids meet offers. The system requires market makers who post both sides, provide liquidity, and absorb risk.
On-chain order books work, but they're expensive. Every order book update (placing, modifying, cancelling an order) consumes gas. For low-volume markets, the gas costs exceed trading revenue. Market makers don't show up. Liquidity starves. The design fails.
AMMs solve this with a different approach. Instead of matching orders, AMMs let users swap against a shared pool. Prices are set algorithmically based on pool balances. Market making becomes passive: LPs deposit assets once, and the smart contract handles everything else.
The Constant Product Formula
Uniswap V1 and V2 use the formula:
x * y = k
Where:
- x = amount of token A in the pool
- y = amount of token B in the pool
- k = constant (doesn't change except when liquidity is added or removed)
Price is determined by the ratio: price of A in terms of B = y/x.
Example
A pool has 10 ETH and 30,000 USDC.
- Price of ETH = 30,000 / 10 = $3,000
- k = 10 * 30,000 = 300,000
A trader wants to buy 1 ETH. After the trade:
- ETH in pool drops to 9
- To maintain k = 300,000, USDC must be: 300,000 / 9 = 33,333.33
- Trader pays 33,333.33 - 30,000 = $3,333.33 for 1 ETH
- Effective price of the trade: $3,333.33 (slippage because buying moved the price)
- New pool price: 33,333.33 / 9 = $3,703.70 (significant slippage for a large trade in a small pool)
Slippage gets worse as trades get larger relative to pool size. This is why deeper pools (more total liquidity) produce tighter execution.
Fees
AMMs charge a trading fee, typically 0.3% on Uniswap V2 pools. The fee is added to the pool rather than sent elsewhere, which means LPs capture the fee proportional to their share of the pool.
Fee tiers on Uniswap V3:
- 0.01% (for near-pegged pairs like USDC/DAI)
- 0.05% (for stable pairs like USDC/USDT)
- 0.3% (standard volatile pairs)
- 1% (exotic or low-volume pairs)
Higher fees compensate LPs for higher impermanent loss risk on more volatile pairs.
Liquidity Providers and Impermanent Loss
When you provide liquidity to an AMM, you deposit both sides of the pair (e.g., equal dollar values of ETH and USDC). You receive LP tokens representing your share. When trades happen, the pool rebalances: ETH gets bought, pool holds less ETH and more USDC, your LP share reflects the rebalanced amounts.
Impermanent Loss Explained
If ETH price rises from $3,000 to $4,000:
- The pool rebalances as traders buy ETH. Your share now holds less ETH and more USDC than when you entered.
- Compared to just holding ETH and USDC passively (no AMM), you're worse off. The pool sold some ETH as it appreciated.
- You would have more total value if you'd never provided liquidity.
This is impermanent loss. It's "impermanent" because if the price returns to the original ratio, the loss disappears. If the price stays moved, the loss crystallizes.
When LP Is Profitable
LP earnings depend on trading fees collected minus impermanent loss incurred. LPing works when:
- Trading volume is high (fees accumulate)
- The pair is relatively stable (low impermanent loss)
- Fee tier is appropriate for the volatility
Stablecoin pairs (USDC/DAI, USDT/USDC) have minimal impermanent loss and can be profitable at very low fees if volume is high.
Volatile pairs (ETH/USDC) require significant volume to offset impermanent loss, especially during trending markets.
Concentrated Liquidity (Uniswap V3)
Uniswap V3 introduced concentrated liquidity in 2021. Instead of spreading LP capital across the entire price range (0 to infinity), LPs choose a specific range (e.g., $2,800-$3,200 for ETH) where their capital is active.
Benefits:
- Higher capital efficiency. Capital concentrated in active ranges earns more fees.
- Better execution for traders within the concentrated range.
Tradeoffs:
- Range management becomes active work. If price moves outside your range, your LP stops earning fees.
- Impermanent loss can be amplified within narrow ranges.
- Passive strategies (set once and forget) work better with full-range V2-style pools.
Uniswap V3 LPing has become professionalized. Tools like Gamma, Arrakis, and Bunni offer managed V3 vaults that automate range management.
Curve: Stable Asset AMM
Curve's innovation is a bonding curve optimized for pegged pairs. For assets that should trade near 1:1 (USDC/USDT, stETH/ETH), Curve's formula keeps prices nearly flat across a wide range of pool composition, producing much better execution than x*y=k for those pairs.
Most Curve pools are stablecoin-stablecoin, stablecoin-basket (3pool, Frax-3CRV), or staked-asset pairs. Curve dominates stablecoin DEX volume because its design is better for that specific use case.
Curve's veCRV system (vote-escrow CRV) governs which pools get CRV emissions, creating the "Curve wars" of 2021-2022 where projects (Convex, Yearn, Olympus) competed to accumulate CRV voting power.
Balancer: Flexible Pools
Balancer extends AMMs to pools with:
- More than two tokens (up to 8 in a single pool)
- Custom weights (an 80/20 pool holds 80% of one asset and 20% of another)
- Various mathematical curves beyond constant product
Balancer's Liquidity Bootstrapping Pools (LBPs) are used for token launches with controlled price discovery.
Uniswap V4 and Hooks
Uniswap V4 (launched 2024) introduced "hooks": custom code that runs at specific points in a pool's lifecycle (before swap, after swap, before LP add, etc.). This enables:
- Dynamic fees based on volatility
- On-chain limit orders
- TWAP oracles built into pools
- Custom logic for specific asset types
V4 essentially turns Uniswap into a platform for AMM variants rather than a single AMM design. Any developer can deploy custom-logic pools.
Related Intelligence
- DeFi TVL: Aggregate liquidity across AMMs and other DeFi protocols.
- Uniswap: Full intelligence brief on the largest AMM protocol.
- Smart Contracts: The code AMMs run on.
Frequently Asked Questions
Not financial advice. Educational purposes only. Do your own research.
Cryptint provides data and analysis for educational purposes only. Nothing on this site is financial advice. Past signals do not guarantee future results. Do your own research. Consult a licensed financial advisor before acting on any information presented here.