Uniswap v3 vs v4 Liquidity: What Actually Changed for LPs

Your exposure is the same on both versions. A band of plus or minus 5% on ETH against dollars behaves identically on Uniswap v3 and v4.

What changed is where the pool’s state lives, how tokens move during a transaction, and who is allowed to run code when a swap arrives.

The first two make trading and managing positions cheaper. The third is the part you need to inspect. By the end you should be able to judge a specific v4 pool and decide whether moving a v3 position there is worth the cost.

Key point: Read the hook before the yield. In v4 the hook is part of the pool’s identity, so two pools on the same pair with different hooks are different markets with different code paths and trust assumptions. Before you migrate, identify every callback the hook can run.

One contract instead of thousands

In v3, every pool is a separate contract, identified by its pair and fee tier^[1]. Each one holds its own tokens, so a route through three pools moves tokens three times.

In v4, one contract, the PoolManager, holds every pool as internal state. A pool is identified by its two tokens, its fee, its tick spacing and its hook address^[8]. Creating a pool becomes an entry in that contract rather than a new deployment, which makes launching a market much cheaper^[8].

v4 also drops v3’s fixed fee tiers. A pool can set any fee from 0% to 100%, or let its hook set the fee dynamically^[10].

The routing effect is direct. A three-hop swap in v3 moves tokens at every step. The same route in v4 updates internal balances and moves tokens once, at the end^[8].

Why settling once is so much cheaper

Ethereum added a kind of storage that lasts only for one transaction and then clears itself^[3]. v4 keeps its running tally there.

During a transaction the PoolManager records what each side owes. Before the transaction finishes, every one of those debts must be settled, or the whole thing reverts^[2].

What that buys	Why it matters
Lower gas on multi-hop and multi-pool operations	Intermediate token transfers disappear[8]
Composability	One contract can rebalance across several pools in one transaction, settling once
Internal balances	Frequent traders can hold token claims inside the PoolManager instead of moving tokens back and forth[5]

Those internal claims use ERC-6909 — a lightweight standard for holding many token balances in one contract^[5]. Your liquidity position is still an NFT, minted by the position manager, just as v3 positions were^[13].

None of that changes the price a swap gets. The pricing rule and tick math are inherited from v3^[2].

Hooks: the change that affects you

A hook is a contract attached to a pool when it is created. It can run at defined moments: when the pool is initialised, around each swap, around liquidity going in or out, and around donations^[6]. Its permissions are encoded in its own address, so you can read what it is allowed to do without trusting anybody’s description^[6]. Once the pool exists, its hook cannot be added, removed or swapped^[6].

What hooks make possible	What hooks also introduce
Fees that rise with volatility, charging arbitrage more when it takes most[9]	Code sitting in the swap path
Custom pricing rules that used to need a separate protocol	Conditions on withdrawal that did not exist in v3
Limit orders and automatic range management inside the pool	Upgradeability, if the hook’s logic sits behind a proxy with an admin key
Fee routing to a treasury, insurance fund or reward programme	Behaviour under stress that may not have been tested at scale

The first row on the left matters most to anyone supplying liquidity. A dynamic fee gives the pool a way to charge for loss-versus-rebalancing — the value arbitrageurs extract because the pool’s price lags the wider market^[12][9]. See Loss-Versus-Rebalancing and Uniswap v4 Architecture and Hooks.

What did not change at all

Your concern	v3	v4
The range maths	The shifted curve between your two bounds	Identical
Divergence	Full, amplified by how narrow your band is	Identical
Out of range	You convert and stop earning	Identical
How fees accrue	Pro rata to liquidity in range at each swap	Identical
Being picked off	Arbitrage against a stale quote	Identical, unless a hook prices it

A band of plus or minus 5% behaves the same way on both. The reasons to prefer one are gas, where the volume goes, and whether a hook improves the fee side. The reasons to be careful about a specific v4 pool are mostly about its hook. Because the range maths is inherited, the Uniswap v3 liquidity calculator prices a v4 band exactly as it prices a v3 one.

If the range mechanics are new, start with Concentrated Liquidity Explained and Out-of-Range Liquidity.

The protocol fee is different on each side

One thing that does differ is how much of the swap fee reaches you. Uniswap governance switched on protocol fees in December 2025^[10]. On v3, LPs now keep three-quarters of the fee in the 0.01% and 0.05% tiers and five-sixths in the 0.30% and 1% tiers^[10].

On v4, protocol fees began in July 2026 on a subset of pools: those without hooks, pools launched through Uniswap’s auction hook, and aggregator hook pools^[11]. For hookless pools, the protocol’s share follows a curve set by governance rather than a fixed fraction^[11]. Pools with other hooks were left out of that vote^[11].

So when you compare two pools, compare fee income net of the protocol fee on each. Governance can change these settings, so check the current values before you decide^[10].

Checking a v4 pool before you supply

1. Resolve the hook address from the pool key, and check whether it is verified and audited.
2. Decode its permissions from that address. Specifically, can it act when liquidity is removed^[6]?
3. Check for a proxy. If the hook’s logic can be replaced, the rules can change after you deposit.
4. Read the fee logic. If the fee moves, understand what sets it and how far it can go^[9].
5. Compare routed volume with the equivalent v3 pool. Architecture does not pay fees; volume does.
6. Simulate the whole lifecycle on Tenderly or a fork: mint, swap through your range, collect, withdraw. Confirm the exit returns what you expect with the hook in place.
7. Run the ordinary checks too, from How to Evaluate a Liquidity Pool. Contract risk and out-of-range risk apply to every version^[4].

A migration decision, worked

You hold $50,000 in a v3 ETH and USDC band earning about $18 a day after the protocol fee. The equivalent v4 pool, with no hook, pays about $21 a day after its protocol fee for the same band, because routers now send it more volume.

	Value
Extra income from moving	about $3 a day
Cost of moving: two transactions and a small swap	about $60 on a quiet day
Time to repay the move	about 20 days

That is a reasonable trade if the extra volume is steady rather than a one-week spike. Now suppose the v4 pool has a hook that can change its fee. Add the time it takes to read and verify that hook, and ask for a longer track record before you trust the higher number.

Whether to migrate

Migration has real costs: two gas payments, locking in your current composition, and price impact (the cost of your own swap moving the price) if the ratio has to be adjusted. Compare that against the extra fee income you expect, and set a payback period you would actually accept.

One operational note for anyone running many positions. In v3 every pool is its own address, so tooling tracks a set of contracts. In v4 there is one address and pools are identified by key. That simplifies indexing, but your monitoring has to resolve the hook for each pool rather than assuming pools on the same pair behave alike.

A sensible default: migrate when the destination’s volume per unit of liquidity clearly beats the source, and the hook is either absent or immutable and audited. Otherwise let the market decide where the flow settles, and follow it with new money rather than churning what you have.

The architecture is cheaper to use, and it allows things v3 could not do. Returns are still decided by volume, volatility, who else is in your band, and how actively the position is managed^[7].

Where to go next

The exposure both versions share, impermanent loss (how far a range position falls behind holding the same tokens), is derived in The Impermanent Loss Formula, and the fee decision is in Uniswap Fee Tiers Explained. For how all three generations sit side by side, including v2, read Uniswap Liquidity Pools and the earlier comparison in Uniswap v2 vs v3.

References

1. Uniswap v3 Core (Adams et al., 2021)
2. Uniswap v4 Core (Adams et al., 2024)
3. EIP-1153: Transient storage opcodes (Ethereum Improvement Proposals)
4. What are the risks when providing liquidity? (Uniswap Labs)
5. ERC-6909: Minimal Multi-Token Interface (Ethereum Improvement Proposals)
6. Uniswap v4 Hooks (Uniswap Developer Documentation)
7. Risks and Returns of Uniswap V3 Liquidity Providers (Heimbach et al., 2022)
8. PoolManager (Uniswap Developer Documentation)
9. Dynamic Fees (Uniswap Developer Documentation)
10. Fees (Uniswap Developer Documentation)
11. Activate v4 Protocol Fees, Part 1/2 (Uniswap Governance Proposal 100, 2026)
12. Automated Market Making and Loss-Versus-Rebalancing (Milionis et al., 2022)
13. Position Manager (Uniswap Developer Documentation)