Photo by Pierre Borthiry - Peiobty on Unsplash
Some analyses suggest a significant proportion of liquidity providers on major automated market makers have, at various points, withdrawn less value than they deposited. The cause isn't fraud or exploits. It's a structural feature built into every constant product pool by design. The mechanism is called impermanent loss, a label that does quiet rhetorical work on behalf of the protocols and token projects that depend on deep liquidity without bearing its full cost. Understanding who actually absorbs that cost, and why the current framing obscures it, is the only way to evaluate whether depositing into a pool is a yield strategy or a subsidy.
This isn't about memorizing a formula. It's about understanding a specific agreement you enter when you deposit into a pool, and recognizing exactly which party to that agreement holds the structural advantage.
How the AMM Mechanism Creates Impermanent Loss
An automated market maker like Uniswap v2 doesn't use an order book. It holds two assets in a pool and prices them against each other using a constant product formula: x multiplied by y equals k, where x and y are the quantities of each token and k is a fixed constant. When a trader buys token A using token B, the pool's balance of token A falls and its balance of token B rises. The price shifts accordingly. The pool is always rebalancing itself in response to trades, automatically, with no human judgment about whether the new price is fair.
Impermanent Loss Grows as ETH Price Moves Away from Entry
Impermanent Loss Grows as ETH Price Moves Away from Entry
% loss vs. holding, at various ETH price multiples from entry
Source: Constant product formula (x * y = k) calculations
Source: Constant product formula (x * y = k) calculations
Here's where the liquidity provider loses ground. Say you deposit ETH and USDC in a 50/50 ratio when ETH trades at $2,000. If ETH climbs to $3,000 on Coinbase or Binance, arbitrage bots, many operated by firms like Wintermute or specialized MEV searchers, immediately buy ETH from your pool at the stale price and sell it on the external market at the higher one. They keep doing this until the pool price matches the external market. By the time that process completes, your pool position holds more USDC and less ETH than when you started. You've been automatically sold out of the asset that appreciated.
The loss is the gap between what your pool position is worth now and what you'd hold if you'd simply kept the original ETH and USDC in a wallet. At a 50 percent price increase in ETH, that gap is roughly 2 percent of total position value. At a doubling of ETH price, it approaches 5.7 percent. These figures follow directly from the math of the constant product formula, not from market conditions that might improve. The mechanism guarantees this outcome every time price moves away from the entry ratio and stays there.
The word "impermanent" does real rhetorical work here. If ETH returns exactly to $2,000, the loss disappears entirely. The framing invites the liquidity provider to imagine a world where prices mean revert. In practice, assets that achieve a 100 percent price gain rarely return to their starting point, and the pool has been bleeding value at every step of the upward journey.
Uniswap v3 Fee Tiers by Pair Type
Uniswap v3 Fee Tiers by Pair Type
| Fee Tier | Typical Pair Type | Volatility Level | IL Risk |
|---|---|---|---|
| 0.01% | Stablecoin pairs | Very Low | Minimal |
| 0.05% | Blue chip pairs (ETH/BTC) | Low to Medium | Low |
| 0.30% | General token pairs | Medium to High | Moderate |
| 1.00% | Exotic or volatile pairs | Very High | High |
Source: Uniswap v3 protocol documentation
Source: Uniswap v3 protocol documentation
Fee Revenue and the Real Arithmetic of DeFi Liquidity Provision
Protocols compensate liquidity providers through trading fees. On Uniswap v3, the most common fee tiers are 0.01 percent, 0.05 percent, 0.30 percent, and 1 percent per trade, with the appropriate tier varying by pair volatility. A highly active pool on a volatile pair can generate meaningful fee revenue. Whether that fee revenue actually covers the impermanent loss over the same period is the real question.
The answer depends on three variables:
Photo by Art Rachen on Unsplash
How Arbitrage Bots Drain Value from a Liquidity Pool
How Arbitrage Bots Drain Value from a Liquidity Pool
Source: Article description of AMM arbitrage mechanics
Source: Article description of AMM arbitrage mechanics
- Trading volume relative to total pool liquidity
- Volatility of the underlying assets over the deposit period
- How concentrated the liquidity position is relative to where price actually trades, which matters more than most new depositors expect
All three interact. A provider earning 40 percent annualized fee APR on a concentrated position in a volatile pool may still finish the month behind a simple hold strategy if a single large price move pushes the position out of range and into single asset exposure. The fee clock stops the moment price exits the range. The impermanent loss clock does not.
Uniswap v3 introduced concentrated liquidity in 2021, letting providers deploy capital within a specific price range rather than across the entire price curve. This amplifies fee earnings when price stays inside the range, sometimes dramatically. But it also amplifies impermanent loss when price exits, because the position becomes entirely one asset and earns zero fees until price returns. A provider who sets a tight range on a volatile pair is making a leveraged bet on price staying put. Most don't fully price that risk at entry.
Stablecoin pairs are the one category where the arithmetic consistently favors the liquidity provider. A USDC/USDT pool on Curve carries essentially zero impermanent loss because neither asset appreciates meaningfully against the other, so fee revenue is nearly pure yield. The trade-off is that fee rates on stablecoin pools are razor thin, typically 0.01 to 0.04 percent, and the absolute yield reflects that. Protocols like Curve have channeled governance token incentives toward their stablecoin pools for years, specifically because rational liquidity providers would otherwise find more attractive risk-adjusted returns elsewhere. The practical verdict: stablecoin pools are the only AMM environment where passive depositors hold a genuine structural edge, and even there it's measured in basis points rather than percentage points.
What Determines Whether Fee Revenue Covers Impermanent Loss
What Determines Whether Fee Revenue Covers Impermanent Loss
Relative impact weight of each factor on LP net outcome (illustrative)
Source: Article analysis of LP profitability variables
Source: Article analysis of LP profitability variables
Who Actually Profits From the Impermanent Loss Structure
The clearest beneficiary of impermanent loss as a structural feature is the arbitrageur. Every time an external price diverges from a pool's internal price, there's a near-zero-risk profit available to whoever executes the trade first. In 2025 and into 2026, that market has consolidated significantly around MEV infrastructure: private mempools, block builder relationships, and latency advantages measured in milliseconds. Retail participants who attempt to arbitrage AMM pools are consistently outrun by professional operations that have invested heavily in execution infrastructure.
The second beneficiary is the protocol itself. Deep liquidity makes the protocol useful, reduces slippage for traders, drives volume, and generates protocol revenue and token demand. The cost of that liquidity, the impermanent loss absorbed by depositors, is externalized onto individual providers rather than borne by the protocol treasury. Token incentives offset some of this, but incentive programs have a documented history of attracting mercenary capital that exits the moment emission rates drop, leaving the pool thinner than before the program started. This pattern repeated across multiple protocols in the 2022 to 2024 cycle and hasn't fundamentally changed.
The liquidity provider who actually comes out ahead tends to fit a specific profile: deploying capital into high-volume stablecoin or correlated asset pairs, actively monitoring and adjusting ranges on concentrated liquidity positions, and treating fee APR as a starting point for analysis rather than a final answer. Passive depositors in volatile pools, the most common participant type on protocols marketed to retail audiences, face a structural disadvantage that the AMM design was never really built around.
The Real Cost of Impermanent Loss at Key Price Milestones
The Real Cost of Impermanent Loss at Key Price Milestones
ETH deposited at $2,000 alongside USDC in a 50/50 pool
|
ETH rises 50%
$2,000 to $3,000
2.0%
IL as share of total position
|
ETH doubles
$2,000 to $4,000
5.7%
IL as share of total position
|
|
The word "Impermanent" only holds true if
ETH returns EXACTLY to $2,000
Assets that double rarely mean-revert. The loss is structural.
|
|
Source: Constant product formula (x * y = k) calculations
Source: Constant product formula (x * y = k) calculations
There's a broader design question sitting underneath all of this. AMMs were built to solve the bootstrapping problem: how do you get liquidity into a new market when no market maker wants to take the risk? The constant product formula solved that elegantly. But elegant for whom? The arbitrageur, the protocol, and the active professional extractor all have tools to manage or profit from the price divergence dynamic. The passive depositor doesn't. That asymmetry isn't a flaw someone forgot to fix. It's a design choice whose costs are still being distributed, mostly downward.
This article is for informational and educational purposes only and does not constitute financial, investment, or legal advice. The views expressed are analytical observations and should not be relied upon for personal financial decisions. Always consult a qualified financial advisor before making investment decisions.
Impermanent Loss Is Quietly Draining DeFi Liquidity Providers