For educational purposes only; not investment advice. Investing may result in loss.
Direct answer
Concentrated liquidity is an AMM design in which liquidity providers (LPs) allocate capital to selected price ranges. Capital can be used more intensively near the market price, but the LP must choose ranges and accept greater inventory, monitoring, and rebalancing risk.
Instead of making markets uniformly from zero to infinity, an LP supplies liquidity only between a lower and an upper price. Concentrating funds near the market price can create greater local depth from the same capital. The trade-off is that the position can go out of range, become entirely one asset, and require active decisions about whether and when to reposition.
In a classic constant-product AMM, reserves satisfy x × y = k, and liquidity theoretically spans the full price interval. Because most trading occurs near the current price, much of that full-range capital may rarely be used.
Concentrated-liquidity protocols divide price space into discrete ticks. An LP selects lower and upper tick boundaries. The position supplies active liquidity and earns its proportional share of swap fees only while the pool price is inside that interval. Where positions overlap, the protocol aggregates their active liquidity at that price.
These positions are not interchangeable pool shares: equal deposits can produce different asset mixes and fees when their ranges or fee tiers differ. Some implementations, including Uniswap v3’s periphery contracts, represent positions as NFTs; other protocols may use different position records.
How it works
Taking the ETH/USDC pool as an example, the price indicates how much USDC 1 ETH is worth. LP sets a range of 1,800 to 2,200 USDC:
-
When the price is in the middle of the range, the position holds both ETH and USDC.
-
When the price rises, traders add USDC and remove ETH, so the position gradually becomes USDC.
-
After the price reaches the upper limit of 2,200, the positions are approximately all USDC and will no longer participate in transactions at higher prices.
-
When the price drops, the process is opposite, and the position gradually becomes ETH; after falling below 1,800, it is almost all ETH.
This is not a stop-loss order. As the price rises through the range, the position progressively sells ETH for USDC; as it falls, it progressively buys ETH with USDC. If price later re-enters the interval, the position becomes active again unless the LP has withdrawn it.
The narrower the range, the more liquidity the same capital can create while in range and the larger its potential fee share. But a narrow position goes out of range more easily and generally earns no swap fees while inactive. Repositioning may require withdrawing, swapping assets, and opening a new range, although simply waiting allows the original position to reactivate if price returns.
Example
Assume that the current price of ETH is 2,000 USDC, and LP A and LP B each invest the equivalent of 10,000 USDC. A chooses 1,900 to 2,100, and B chooses 1,500 to 2,500. If the price is always between 1,950 and 2,050 within a week, A’s funds are more concentrated and may obtain a higher share of local handling fees under other conditions being equal.
ETH then quickly rose to 2,300. A’s position exceeded the upper limit, almost all of it was converted to USDC, and market making stopped at the current price; B remained within its range and continued to trade. If A wants to follow the new price and establish a range of 2,200 to 2,500, the position must first be closed, part of the USDC may need to be exchanged back to ETH, and gas and slippage must be paid. If the price falls immediately afterward, A may bear the cost of switching positions and miss the rebound in the original range.
Income cannot only be based on handling fees. Suppose A gets 120 USDC in handling fees per week, but it is 180 USDC less than simply holding two assets, and it costs another 20 USDC to rebuild the position, then the relative result is 120 - 180 - 20 = -80 USDC. This is just a scenario calculation and does not represent future results; the actual reinvestment of token rewards, taxes and handling fees will also be included.
Looking at the stable currency pool: the price of USDC/USDT is usually close to 1, and LP can be set in a narrow range from 0.995 to 1.005. Funding efficiency is very high under normal circumstances, but once USDC falls to 0.90, all positions may become depreciating USDC. The narrow range does not eliminate the risk of de-anchoring, but instead completes asset conversion faster.
Risks
An LP can split capital among ranges, for example placing 60% in a wider interval for coverage and 40% in a narrower interval for local depth. This does not guarantee a profit; it only changes the exposure and management burden.
Concentrated liquidity is an active market-making position, not a fixed income product. Price out-of-bounds, impermanent losses, token de-anchoring, contract loopholes, and rebalancing costs may all make handling fees insufficient to cover losses.
The protocol requires range boundaries to fall on permitted ticks. Tick spacing may vary by fee tier, and an interface may round entered limits to valid ticks, so the executed range can differ slightly from the displayed input. An extremely narrow range can be crossed by an ordinary price move.
Swap fees are not distributed according to total capital deposited across the whole pool. They are allocated among liquidity active at the traded price. If active liquidity is 1,000,000 units and one position contributes 100,000, its fee share is roughly one-tenth while that liquidity remains active; the share changes when price crosses ticks or other LPs add or remove liquidity.
Different rate pools will compete for orders of the same trading pair. Routers may split large orders into multiple pools, and low-rate pools may not always have more revenue, and high-rate pools may not necessarily compensate for fluctuations. Short-term incentive tokens will attract a large amount of liquidity, causing the share of individual LP fees to decrease; after the incentive ends, TVL and transaction paths may migrate again.
Limit orders and centralized liquidity may seem similar, but they are not exactly the same. The unilateral narrow range can achieve gradual currency exchange when the price crosses, but it will not be automatically canceled after the transaction like traditional orders; when the price reversely crosses back to the range, the asset may be exchanged again. Users who want to realize a one-time swap need to actively withdraw after going out of bounds.
Protocol upgrades or front-end migrations may also change the way positions are managed. Old positions are usually still recorded in the original contract and will not be automatically migrated as the new version of the interface comes online. Users should confirm ownership, receivable fees and approved objects from the on-chain position manager, and compare the new version of the contract address, rate and audit scope before migration. Granting full NFT permissions to a so-called migration assistant may allow the other party to transfer the entire position.
The starting point for review is not the “handling fees earned” shown on the page, but the establishment of a direct holding benchmark for the same period. Record the quantity and price of the two assets when opening a position, and calculate how much it will be worth on the review date if there is no market making at all; then calculate the total value of the current withdrawable assets, unclaimed handling fees and rewards, and deduct the costs of opening the position, adjustments, exchanges and exits. The difference between the two can show whether market making is effective compared to holding.
For example, if you initially hold 2 ETH and 4,000 USDC, the ETH is $2,000 at the time, and the total value is $8,000. At the time of review, ETH was $2,500, and the direct holding was worth $9,000. The withdrawable asset value of the LP position is US$8,720, the handling fee is US$240, and the accumulated gas and swap losses are US$60. The net value is US$8,900, which is US$100 behind the holding. There is nothing wrong with saying “earned $240 in fees,” but it misses the more important opportunity cost.
It is also necessary to calculate the proportion of time within the range, how long each out-of-bounds period lasts, whether the price is reversed after adjustment, and the fluctuation days on which handling fees are concentrated. When the sample is only one week, the annualized figures are easily amplified by single-day transactions. The goal of review is to identify whether the interval matches the management rules, rather than to use one result to prove that the strategy is permanently effective.
Common misconceptions
Myth 1: Narrow ranges will definitely earn more fees
Narrow ranges only improve capital efficiency when the price is within the range. Actual fees may be less when there are frequent out-of-bounds or excessive competition for liquidity.
Myth 2: Assets are locked after going out of bounds
Usually it can still be withdrawn. The problem is that the position has become unilateral and active market making has stopped, rather than the agreement automatically freezing the assets.
Myth 3: There is no impermanent loss in the stablecoin pool
The difference is small when there are small fluctuations, but deanchoring will cause LPs to hold more vulnerable assets, and extreme results may be large.
Myth 4: Automatic rebalancing can eliminate risks
Automatic strategies only adjust the range according to rules, and will also increase the risk of fees, contracts and strategy failure.
Related topics
Sources
- Uniswap v3 Core - Uniswap Labs (accessed: 2026-08-20)
- Concentrated Liquidity - Uniswap Developers (accessed: 2026-08-20)