Types of Decentralized Exchanges

Overview: The DEX Architecture Landscape

Not all decentralized exchanges are built the same way. While the term "DEX" is often used as a monolith, the design space encompasses fundamentally different architectures, each with distinct trade-offs in capital efficiency, execution quality, and user experience. Understanding these architectures is essential for evaluating which DEX is appropriate for a given trading scenario.

The major DEX categories are:

1. Constant Product AMMs (Uniswap V2-style)
2. Concentrated Liquidity AMMs (Uniswap V3-style)
3. StableSwap DEXes (Curve-style)
4. On-Chain Order Book DEXes
5. DEX Aggregators
6. Hybrid Models

1. Constant Product AMM DEXes

The constant product AMM is the foundational DEX design, introduced by Uniswap V1 in 2018. It uses the invariant:

x * y = k

Where x and y are the reserves of two tokens in a pool, and k is a constant that only changes when liquidity is added or removed. This formula distributes liquidity uniformly across the entire price range from zero to infinity.

How It Works

Consider a pool with 10 ETH and 20,000 USDC. The invariant k = 10 * 20,000 = 200,000. The implied price of ETH is 20,000 / 10 = $2,000. If a trader wants to buy 1 ETH, the pool must maintain k = 200,000, so the new ETH reserve = 9 and the new USDC reserve = 200,000 / 9 = 22,222.22. The trader pays 22,222.22 - 20,000 = $2,222.22 for 1 ETH—a price of $2,222.22 per ETH, reflecting the price impact of the trade.

Strengths and Weaknesses

• Strengths: Extreme simplicity, gas-efficient, liquidity always available at any price, no active management required from LPs.
• Weaknesses: Low capital efficiency (most liquidity sits at prices far from the current market), significant price impact for large trades, and LPs suffer impermanent loss on directional moves.

Examples: Uniswap V2, SushiSwap, PancakeSwap (BSC), QuickSwap (Polygon).

2. Concentrated Liquidity AMMs

Uniswap V3 (launched May 2021) introduced concentrated liquidity, allowing LPs to allocate capital within custom price ranges rather than across the entire curve. This was a paradigm shift in AMM design.

How It Works

Instead of providing liquidity from price 0 to infinity, an LP on Uniswap V3 specifies a range—for example, $1,800 to $2,200 for an ETH/USDC pool. Within that range, the LP's capital behaves as if it were a much larger position on a V2-style curve. If the price stays within the LP's range, they earn a proportionally higher share of trading fees relative to their capital deployed.

The trade-off is that if the price moves outside the LP's selected range, their liquidity becomes inactive (earns no fees) and is entirely concentrated in one of the two tokens. LPs must actively manage their positions to remain in range.

Capital Efficiency Gains

Uniswap's research showed that a concentrated liquidity position within a narrow range can be up to 4,000x more capital efficient than a full-range V2 position. In practice, most LPs achieve 2–10x efficiency gains, which translates to higher fee income per dollar of liquidity deployed.

Examples: Uniswap V3 and V4, PancakeSwap V3, Ambient Finance (formerly CrocSwap).

3. StableSwap DEXes

Curve Finance pioneered the StableSwap invariant, a bonding curve specifically designed for assets that should trade near a 1:1 ratio: stablecoin pairs (USDC/USDT/DAI), wrapped token pairs (ETH/stETH), or synthetic asset pairs.

The StableSwap Invariant

Curve's bonding curve is a hybrid between a constant sum formula (x + y = k, which gives zero slippage but can be fully drained) and a constant product formula (x * y = k, which has high slippage near parity). The StableSwap curve behaves like a constant sum near the peg—providing extremely low slippage for balanced trades—and transitions to a constant product curve as reserves become imbalanced, providing resilience against full drainage.

The amplification coefficient A controls how "flat" the curve is near the peg. Higher A values provide lower slippage for balanced swaps but increase the risk of losses if one asset depegs significantly.

Use Cases

• Stablecoin swaps: Trading between USDC, USDT, and DAI with minimal slippage.
• Liquid staking derivatives: Swapping between ETH and stETH, rETH, or cbETH.
• Wrapped asset pairs: Trading between WBTC and renBTC or similar wrapped Bitcoin variants.

Examples: Curve Finance, Ellipsis Finance (BSC), Saddle Finance.

4. On-Chain Order Book DEXes

Order book DEXes replicate the traditional exchange model: users submit limit orders at specific prices, and a matching engine pairs buyers with sellers. On-chain order books provide more granular price discovery and are preferred by professional traders accustomed to the CEX experience.

Challenges

The fundamental challenge is that every order placement, cancellation, and modification requires a blockchain transaction. On Ethereum mainnet, the gas cost and latency make this impractical. Viable on-chain order books require either:

• High-throughput blockchains: Solana (400ms blocks, low fees), Sei (optimized for trading).
• Application-specific chains: dYdX Chain (Cosmos-based, built-in matching engine), Hyperliquid (custom L1).
• Hybrid off-chain/on-chain models: Off-chain order matching with on-chain settlement.

Examples: dYdX (perpetuals), Hyperliquid, OpenBook (Solana), Econia (Aptos).

5. DEX Aggregators

DEX aggregators are not exchanges themselves but routing protocols that search across multiple DEXes to find the optimal execution path for a given trade. They solve the fragmented liquidity problem: rather than manually checking prices across Uniswap, Curve, SushiSwap, and Balancer, a user submits a trade to the aggregator, which splits the order across multiple venues to minimize price impact and maximize output.

How Aggregation Works

1. The aggregator queries on-chain reserves (or off-chain caches) across dozens of DEXes and liquidity sources.
2. A routing algorithm calculates the optimal split—for example, 60% through Uniswap V3, 25% through Curve, 15% through Balancer.
3. The trade is executed in a single transaction through a smart contract that interacts with each DEX in sequence.

Advanced Features

• CowSwap uses a batch auction model where trades are collected over a window and matched peer-to-peer before routing remaining volume to on-chain liquidity. This provides MEV protection and can yield prices better than any single DEX.
• 1inch Fusion uses a Dutch auction model where professional market makers (resolvers) compete to fill user orders at progressively better prices.

Examples: 1inch, Paraswap, CowSwap, 0x API, Jupiter (Solana).

6. Hybrid Models

Some protocols combine elements of multiple architectures:

• Balancer: Generalized AMM supporting multi-token pools with custom weights (not just 50/50), enabling index-fund-like products and complex liquidity structures.
• Maverick Protocol: AMM with directional liquidity provisioning, allowing LPs to express market views by concentrating liquidity in the direction they expect price to move.
• Uniswap V4 (hooks): Introduces a plugin system where developers can attach custom logic (dynamic fees, on-chain limit orders, TWAP execution) to standard AMM pools.

Comparison Table