Uniswap and the Automated Market Maker Model Explained

Every Uniswap pool is funded by liquidity providers who deposit both assets of a pair. In return they receive a claim on the pool and a proportional share of the fees paid by traders. There is no application process and no minimum size beyond what gas costs make sensible.

In Uniswap v2 that claim is an ERC-20 token minted at deposit and burned at withdrawal. Because it is fungible, it can itself be moved, staked elsewhere, or used as collateral, which is how so much of the wider DeFi ecosystem ended up building on top of Uniswap pool shares.

Uniswap v3 replaced that with a position defined by a lower and an upper price bound. Capital sits idle outside those bounds and earns nothing, but inside them it backs far deeper quoted liquidity than the same money would in v2. A tight range around the current price can multiply fee income for the same deposit, at the cost of needing active management.

When price moves outside a v3 range, the position converts entirely into the weaker asset and stops earning fees until price returns or the provider rebalances. This is the practical burden of concentrated liquidity, and it is why passive providers sometimes prefer wide ranges or older Uniswap v2 pools despite lower theoretical efficiency.

The risk that gets discussed most is impermanent loss, also called divergence loss. When the relative price of the two pooled assets changes, the automated rebalancing inside a Uniswap pool leaves the provider holding more of the asset that fell and less of the one that rose, so the position is worth less than simply holding both tokens would have been. Fees may or may not cover the gap.

Impermanent loss is worst for volatile, uncorrelated pairs and mild for assets that track each other, such as two dollar stablecoins. That is the main reason Uniswap v3 introduced several fee tiers: correlated pairs need only a sliver of fee to be viable, while long-tail tokens need much more to compensate providers for the risk they take.

UNI launched in September 2020 as a governance token for the Uniswap Protocol. Its most discussed feature at launch was a retroactive distribution: anyone who had used the protocol before a cutoff date could claim 400 UNI, an unusual move that rewarded prior users rather than buyers.

The initial supply was one billion tokens, allocated across community members, the team, investors, and advisors, with team and investor allocations vesting over four years. After that period the design includes a low perpetual inflation rate so that governance participation can continue to be funded without a fixed cap forcing a hard stop.

Holding UNI is not the same as holding equity in Uniswap Labs and carries no promise of revenue. What it confers is voting weight over the protocol's on-chain parameters and over a treasury of tokens that governance can spend on grants, integrations, and deployments to new chains.

The parameter that generates the most debate is the protocol fee switch. The contracts allow a portion of the swap fee to be diverted from liquidity providers to a destination that governance chooses. Turning it on would create protocol revenue but reduce the return to the people supplying liquidity, and the tradeoff has been argued over repeatedly since Uniswap v2 shipped.

UNI is also used to authorize new deployments of Uniswap contracts to other chains, since the license and the brand attach to deployments blessed by governance. That is why proposals about supporting a particular network appear regularly in the forum before the contracts show up.

For everyday users, none of this is required. You can swap on Uniswap or provide liquidity without ever touching UNI. The token matters if you want influence over how the protocol develops, or if you are evaluating the project as an investment, which is a different question from evaluating it as software.

Uniswap started on Ethereum mainnet and still treats it as the canonical home for governance and the UNI token. But high mainnet gas costs pushed activity toward rollups and other EVM-compatible chains, and governance has approved deployments across a broad set of them, including networks such as Arbitrum, Optimism, Polygon, Base, and BNB Chain.

Each deployment is a separate set of contracts with its own liquidity. A pool for a given pair on one chain has nothing to do with the pool for the same pair on another, and prices can differ until arbitrage closes the gap. Bridged versions of a token are also distinct assets, which is a common source of confusion and of costly mistakes.

Practically, this means checking which network your wallet is connected to before doing anything on Uniswap. Sending funds to a contract address that exists on a different chain, or trading a bridged wrapper you did not intend to hold, are among the most frequent user errors and neither is reversible.

The upside is real. On a low-fee rollup, small trades and modest liquidity positions become viable in a way they never were on mainnet, and rebalancing a concentrated Uniswap v3 range stops being prohibitively expensive. Much of the routine activity in the protocol now happens away from Ethereum layer 1 for exactly this reason.

Smart contract risk sits underneath everything. The core Uniswap contracts have been audited, formally reviewed, and battle-tested by years of high-value use, which is meaningful evidence but not a proof of safety. Peripheral contracts, hooks, and third-party integrations have far less scrutiny behind them.

Token contracts themselves are a distinct hazard. Some tokens implement transfer fees, blocklists, or the ability to mint arbitrarily, and behavior like that can break assumptions a Uniswap pool makes or trap a holder who can buy but not sell. The pool cannot protect you from a malicious asset.

Rug pulls follow a recognizable pattern. A team seeds a pool, promotes the token, and then removes the liquidity, leaving holders with an asset that has no market. Because listing on Uniswap requires no permission, the presence of a pool tells you nothing about the project behind it.

Approval hygiene matters too. Granting an unlimited allowance to a contract you do not understand is a standing risk, since that contract can move the approved token later. Signature-based approvals with expiries, as used by Permit2 in the Uniswap stack, reduce the blast radius, and periodically revoking stale allowances is good practice.

Finally, be skeptical of yield figures. A pool advertising an extreme annualized return on Uniswap is usually extrapolating from a brief spike in volume, or sitting on an asset whose price is about to move against providers. Fee income is real, but it is compensation for risk, not a free return.

Because the contracts are autonomous while the interfaces are operated by a company, regulators have generally focused on Uniswap Labs rather than on the code. That split raises hard questions that different jurisdictions have answered differently, including who, if anyone, is the operator of a permissionless exchange.

Uniswap Labs disclosed in 2024 that it had received a Wells notice from the U.S. Securities and Exchange Commission, indicating that enforcement action was under consideration. The agency later closed the investigation without bringing charges, a resolution reported alongside a broader shift in the SEC's posture toward crypto firms during that period.

Tax treatment is a separate matter and varies widely by country. In many jurisdictions each swap is a disposal that must be reported, and fee income earned from providing liquidity on Uniswap may be taxable as it accrues. This page is not tax or legal advice, and local rules should be checked with a qualified professional.