Impermanent Loss

Impermanent loss is the mathematical byproduct of how automated market makers keep a trading pair balanced. Most AMMs use a constant product formula (x × y = k): as one token's price moves against the other, the pool's smart contract automatically sells the appreciating asset and buys more of the depreciating one to keep the product constant. A liquidity provider's share therefore ends up holding less of the winning token and more of the losing one than if they had simply held both assets in a wallet.

The size of the loss depends only on how far the price ratio has moved, not on which direction: a 25% price change produces roughly 0.6% impermanent loss, a 2x move about 5.7%, and a 5x move can exceed 25%. Volatile or uncorrelated pairs carry the most risk, while pools of two assets that track each other closely, such as two dollar-pegged stablecoins or ETH and a liquid-staked ETH derivative, see very little.

Concentrated-liquidity designs, popularized by Uniswap v3, sharpen this trade-off: providers earn more fees per dollar by supplying liquidity only within a narrow price band, but a position that trades outside that band stops earning fees and can suffer a larger effective loss than a full-range pool. Research has found that a majority of concentrated positions have historically lost more to impermanent loss than they earned in fees, which is why many providers now use automated vault managers, wider ranges, or correlated pairs like ETH/stETH to reduce exposure. The loss reverses if the price ratio returns to its original level, so it only becomes realized upon withdrawal.