There is a conversation that keeps happening in DeFi developer circles. You can find versions of it on Farcaster threads, in private Discord channels, in the corners of crypto Twitter where people talk about gas optimization without scaring off a retail audience. It always circles the same claim: Uniswap v4 hooks are the biggest architectural change in onchain trading since the original AMM paper. And every time, the same question keeps getting dodged.
Who actually ends up with the money the hooks put in motion?
Not in the abstract. In the specific. When a hook rebates MEV, who signs the rebate check. When a hook lets an institution gate liquidity by jurisdiction, who pays the compliance cost and who captures the premium. When dynamic fees finally work the way the whitepaper suggests they could, which wallet address collects the delta.
This is a "follow the incentives" piece, not a whitepaper summary. You can read the Uniswap Labs docs for the whitepaper summary. What I am interested in is the pattern — and there is a pattern — of how a supposedly neutral upgrade to the third-largest DeFi protocol by TVL (5.8 billion dollars per public measurement, audited by Trail of Bits, ABDK, and Consensys Diligence, zero cumulative exploit losses since its 2018 launch) quietly rearranges who gets to charge rent on a block of code that used to charge nobody.
The Permissionless Story That Requires Permission
Here is the first pattern. Every time someone explains hooks to a general audience, they open with "anyone can now customize a pool". Technically true. Operationally misleading.
A hook is a smart contract. A smart contract that runs at pool lifecycle points — before swap, after swap, before add liquidity, after add liquidity, and so on — with enough power to reject the action or change its parameters. That is not a configuration file. That is live code in the critical path of a pool that could be sitting on seven or eight figures of someone else's money. Writing it badly breaks the pool. Writing it maliciously drains the pool.
Which means the actual set of entities that can ship a hook LPs will trust is small. It is teams with audit budgets. It is firms with legal wrappers. It is the same narrow group that was already writing DeFi protocol code before v4 existed. The "anyone can customize" framing collapses the moment you ask the second question: anyone can customize, yes, but whose pool will LPs actually put capital into?
The answer, in practice, is the hooks shipped by people who can afford to pay the same auditors listed on Uniswap's own audit trail — Trail of Bits, ABDK, Consensys Diligence. The cost of a serious audit for a nontrivial contract sits in the high five figures to low six figures. A solo dev with a weekend idea is not paying that. A VC-backed hook startup is.
So hooks do not democratize pool creation. Hooks re-centralize pool economics around teams that can credibly guarantee the code is safe. Which, if you were already in the business of writing DeFi middleware, is excellent news. If you were hoping v4 would let you spin up a niche pool over a weekend and collect fees on it, read that sentence again.
The Dynamic Fee Promise Where The Fee Goes Up
Second pattern. "Dynamic fees" is the most-repeated selling point of v4. The pitch is that hooks let pools charge more during volatile periods and less during calm ones, which — the argument goes — protects LPs from toxic flow and tightens spreads when the market is quiet.
Set aside whether that is mechanically true for a second. Ask instead: who decides the fee curve, and who captures the delta between the old static fee and the new dynamic one?
In a v3 pool, the fee was a constant. LPs knew the number. Toxic flow still got a piece of them, but the piece was predictable. In a v4 pool with a dynamic fee hook, the fee is a function. Somebody wrote that function. Somebody can change it, or deploy a new version, or — depending on how the hook is built — front-run the parameter update. The LP sees the output. The LP does not see the math unless the hook is open source and verified, and even then, the LP is now depending on three things: the hook author's honesty, the oracle the hook reads from, and the upgrade key (if any).
Compare the alternative that retail actually uses today. On a centralized exchange — pick any of the five I can pull clean fee data on — the schedule is published. Binance takes 10 basis points maker and 10 taker. Bybit, same. Bitget, same. OKX is 8 maker and 10 taker. MEXC is the outlier at 0 maker and 2 basis points taker, which is a fight-for-listings move and also worth noting because MEXC's reserve disclosures are partial rather than fully verified, unlike the other four. Those numbers are boring and fixed. Boring and fixed is also very easy to reason about.
A dynamic fee on a v4 pool is not necessarily worse than any of that. It is just no longer a number on a schedule. It is a variable controlled by whoever wrote the hook. Which, again, is a business model — not a feature.
The thing being sold as "programmable liquidity" is more honestly described as a new layer of intermediation on top of the base AMM, and new layers of intermediation have never, in any market, stopped existing because users were told the layer was optional.
The Compliance Layer Hiding In A Decentralization Pitch
Third pattern, and this is the one that gets the least airtime in the public discussion. Hooks make it trivial to gate liquidity. A before-swap hook can check the caller's address against a whitelist, a KYC attestation, a jurisdiction proof, an on-chain identity credential. It can reject the swap in one line. This is not a theoretical use case. This is what an institutional LP wants.
Here is the thing worth saying out loud. The biggest unsolved problem for institutions holding real money is not "will my Uniswap trades settle". Uniswap has never been exploited. It sits at 5.8 billion dollars of TVL, it is the third-largest DeFi protocol by that metric, and it has been audited by three of the most expensive security firms in the space. The problem for institutions is that the pool is permissionless, which means their compliance desk cannot sign off on the counterparty risk. Everyone in that pool is, from a compliance standpoint, a stranger.
Hooks solve that. A permissioned pool with a KYC hook is no longer a permissionless liquidity venue. It is a private pool that settles on a public blockchain. The distinction matters. And the incentive alignment is obvious — the institution gets the onchain settlement rail they like, the jurisdictional gate they need, and the plausibility of "well, it's still DeFi, we're using Uniswap". The hook author gets a fee for running the compliance middleware. The retail LP in the permissionless version of the same pool is not invited to the gated version, and if the gated version routes most of the real order flow, the permissionless version slowly becomes a ghost town.
This is not a prediction. It is the same playbook that produced the CEX licensing map — Binance with VARA in Dubai and limited tickets from AMF in France and OAM in Italy, Bybit with CySEC and VARA, Bitget with Lithuanian FCIS and Polish KNF, OKX with provisional VARA and a full SCB license in the Bahamas. Every serious crypto venue in the last three years has walked toward jurisdictional licensing, not away from it. Hooks give DeFi the same option. Whoever ships the best compliance hook captures the institutional flow.
The Singleton Tax Nobody Names
Fourth pattern. The v4 architecture collapses all pools into a single contract — the "singleton" design. The marketing around this is about gas savings, which are real. Deploying a new pool is cheaper. Multi-hop swaps through several pools are cheaper. Those are not small wins.
But I want to point at what the singleton also does, which is concentrate the attack surface and the control surface of the entire v4 deployment into one contract. Uniswap has a clean history — zero cumulative exploit losses since 2018, which is genuinely remarkable for a protocol that has moved the volume it has. The audits are from Trail of Bits, ABDK, and Consensys Diligence. That track record is the reason I am willing to take the singleton design seriously at all.
The risk shift is this. In v3, a bug in a pool was a bug in a pool. In v4, a bug in the singleton is a bug in all pools at once. A bug in a hook is still scoped to pools that use that hook, but a bug in the accounting layer the singleton exposes to every hook is a bug in the entire system. The trade-off is not "v4 is less safe than v3". The trade-off is that v4 concentrates the dependency on the core team's security process even harder than v3 already did. Which — follow the incentives again — is excellent for Uniswap Labs. A singleton gives them a tighter choke point on the protocol's continued relevance. It is a codebase moat. Competing forks have to rebuild the singleton, re-audit it, and convince LPs to migrate.
So What Do You Actually Do
If you are a passive LP, the only question you should be asking about v4 hooks is who wrote the hook, are they public, what is their audit history, and is the fee logic open source. Nothing else matters. A hook is a custody-adjacent object. You are not just depositing into a pool, you are depositing into a pool that runs somebody's logic on every swap. Treat it like you would treat a yield strategy, not like you would treat a v2 pair. If the hook is anonymous or unaudited, the fact that the underlying pool sits on Uniswap is close to irrelevant — you have underwritten the hook, not the protocol.
If you are a builder, the interesting question is the inverse. The hooks that will make money are the ones that solve a compliance or routing problem for an entity that has real capital and a real reason to pay a premium for a better execution surface. Dynamic fee hooks for retail LPs are a crowded space. KYC-gated institutional pools, MEV-capture hooks that actually share revenue with LPs on a verifiable basis, and liquidity-sourcing hooks for aggregator routes are not. The incentive is in the plumbing, not in the landing page.
And if you are a retail trader who just wants to swap tokens, the honest answer is that v4 will show up in your life as slightly cheaper gas and slightly better prices on whatever aggregator you already use. That is a real improvement. It is also not the revolution the announcement post sold. The revolution is happening inside the hook market, and whether that market looks like a permissionless playground or a handful of compliance-gated venues with tight fee curves and tighter allowlists is a question of who ends up writing the hooks that matter. I would not bet against the teams with the audit budgets.