DeFi Intents Explained

How intent-based trading works: solvers, RFQs, MEV protection, and the evolution of decentralized exchange execution

15 min read Intermediate Free
Key Insight

Intents flip the traditional transaction model: instead of specifying how to execute a trade (which contract, which route, what slippage), users specify what they want (swap 1 ETH for at least 3,000 USDC). Solvers compete to find the best execution, protecting users from MEV and improving prices. This shift from imperative to declarative transactions is reshaping how DeFi works.

The Problem with Traditional DEX Trades

When you swap tokens on a traditional AMM like Uniswap, you construct a detailed transaction:

  • Which contract to call
  • Exact route through pools
  • Maximum slippage tolerance
  • Gas price and priority fee
  • Deadline for execution

This creates several problems:

MEV Exposure

Your transaction sits in the public mempool before execution. MEV searchers see your trade, front-run it (pushing the price up), then your trade executes at a worse price, then they back-run (selling at the higher price). This "sandwich attack" extracts value from every swap.

  • Information Asymmetry: Sophisticated actors have better data, faster infrastructure, and optimized routes. Regular users get worse execution.
  • Complexity: Choosing optimal routes, estimating slippage, and setting gas requires expertise most users lack.
  • Fragmented Liquidity: Liquidity is spread across hundreds of venues (DEXs, CEXs, OTC). Users typically access only a fraction.

What Are Intents?

An intent is a signed message expressing what a user wants to achieve, without specifying how to achieve it.

Intent
A set of declarative constraints signed by a user. Example: "I want to swap 1 ETH for at least 3,000 USDC within the next 10 minutes." The intent doesn't specify which DEX, route, or execution strategy—only the desired outcome.

Instead of the user figuring out execution, specialized parties called solvers compete to fulfill the intent with the best possible outcome.

Intents vs. Transactions

Aspect Traditional Transaction Intent
Approach Imperative (specify exact steps) Declarative (specify desired outcome)
Execution User-defined route Solver-optimized route
MEV Exposure Public mempool visible to searchers Private orderflow, batch auctions
Liquidity Access Single venue or aggregator On-chain, off-chain, CEX, OTC
Gas Payment User pays in ETH Can be abstracted (solver pays)
Complexity User must understand execution User states goal, solver handles details

How Intent-Based Trading Works

Intent Execution Flow
1
User Signs Intent: Specifies desired outcome (swap X for at least Y) via wallet interface
2
Intent Broadcast: Sent to solver network (not public mempool)
3
Solver Competition: Solvers find optimal execution paths across all liquidity sources
4
Winner Selection: Best execution (auction or batch) wins the right to fill
5
On-Chain Settlement: Winning solver executes, user receives tokens

The Solver Ecosystem

Solvers are the specialized parties that fulfill intents. They're essentially professional trade executors who compete to provide the best prices.

What Solvers Do

  • Aggregate Liquidity: Access DEXs, CEXs, OTC desks, private inventory, even other solver networks
  • Optimize Routes: Find the best path across fragmented liquidity
  • Manage Inventory: Hold tokens to provide instant fills without hitting AMMs
  • Hedge Risk: Maintain delta-neutral positions to minimize exposure

Solver Incentives

Solvers profit from the spread between what the user accepts (minimum output) and what they can actually achieve. Competition drives this spread lower, benefiting users.

Solver Competition Benefits Users

If you submit an intent to swap 1 ETH for at least 3,000 USDC, and a solver can fill it for 3,050 USDC, they keep 50 USDC as profit. But if another solver can fill for 3,040 USDC and offers you 3,030 USDC, they win the auction. This competition pushes execution quality toward the theoretical best.

Request for Quote (RFQ) vs. Intents

Intent-based systems often use RFQ (Request for Quote) mechanisms where users request price quotes from market makers.

Feature Pure Intent RFQ-Based
Liquidity Sources All on-chain and off-chain Whitelisted market makers
Quote Process Solver finds best route Market makers stream quotes
Settlement Can be complex (multiple hops) Direct bilateral transfer
Slippage Route-dependent Zero (fixed quote)
Trust Model Open solver competition Trust whitelisted entities

Most modern systems are hybrids: they use RFQ quotes from professional market makers while also routing through DEXs when that provides better execution.

Major Intent Protocols

CoW Protocol (CoWSwap)

The pioneer of intent-based trading. CoW Protocol batches intents and uses batch auctions to find optimal execution:

  • Coincidence of Wants (CoW): Match opposing trades directly (user A sells ETH, user B buys ETH) without touching AMMs
  • Batch Auctions: All intents in a batch settle at the same price, eliminating MEV
  • Solver Network: Open competition among solvers to fill batches
  • MEV Protection: Private orderflow plus uniform clearing price

UniswapX

Uniswap's intent-based system combining Dutch auctions with solver competition:

  • Dutch Auction: Price starts favorable for solvers and decays toward market price over time
  • Filler Network: Solvers compete to fill orders as prices become attractive
  • Cross-Chain: Designed to support cross-chain swaps via intent settlement
  • Integration: Accessible through Uniswap interface alongside traditional AMM swaps

1inch Fusion

1inch's intent mode combining aggregation with solver execution:

  • Resolver Network: Professional fillers compete for orders
  • Dutch Auction: Price decay mechanism similar to UniswapX
  • Gasless: Resolvers pay gas fees, users sign off-chain
  • Aggregation: Leverages 1inch's existing DEX aggregation infrastructure

Across Protocol

Intent-based bridging for cross-chain transfers:

  • Relayer Network: Professional fillers provide instant cross-chain liquidity
  • Optimistic Verification: Fast fills with delayed settlement verification
  • Capital Efficiency: Relayers front capital, reimbursed from canonical bridge

MEV Protection in Intent Systems

Intent-based systems protect users from MEV through several mechanisms:

1. Private Orderflow

Intents don't go to the public mempool. Solvers see them privately, compete in auctions, and only the winning execution hits the chain. Searchers can't front-run what they can't see.

2. Batch Auctions

CoW Protocol batches multiple intents and settles them at a single clearing price. There's no "before" and "after" within a batch—everyone gets the same price, eliminating sandwich attacks.

3. Dutch Auctions

Price decay mechanisms ensure that solvers who fill earlier (faster execution) pay better prices. This aligns solver incentives with user outcomes.

4. Coincidence of Wants

When two users want opposite trades (A sells ETH, B buys ETH), they can be matched directly at a fair price without touching AMMs. No AMM means no MEV extraction from AMM price impact.

MEV Redistribution vs. Elimination

Intent systems don't eliminate MEV—they redirect it. The value that would go to sandwich attackers instead goes to: (1) solvers as profit margin, (2) users as better execution, or (3) protocol as fees. Competition among solvers determines the split.

Cross-Chain Intents

Intents are particularly powerful for cross-chain operations. Instead of manually bridging, swapping, and managing multiple transactions:

Cross-Chain Intent
"I want to swap 1 ETH on Ethereum for at least 3,000 USDC on Arbitrum within 10 minutes." The user signs one message; solvers handle bridging, routing, and settlement across chains.

ERC-7683: Cross-Chain Intent Standard

Developed by Uniswap and Across, ERC-7683 standardizes how cross-chain intents are expressed:

  • Common format for intent messages across protocols
  • Standardized settlement contracts on each chain
  • Enables solver network interoperability
  • Reduces fragmentation in cross-chain execution

Advanced Intent Types

Conditional Intents

"Swap my ETH for USDC if ETH price drops below $3,000." Intents can include conditions based on price, time, or other on-chain events.

Continuous Intents

"Dollar-cost average $100 of USDC into ETH every day for 30 days." A single signed intent can authorize repeated executions.

Multi-Step Intents

"Unstake my ETH, swap to USDC, bridge to Arbitrum, deposit to Aave." Complex multi-step operations expressed as a single intent, executed atomically or sequentially.

Challenges and Risks

Centralization Risk

If only a few solvers dominate, they could extract monopoly rents or collude. Healthy intent systems need many competitive solvers.

Solver Trust

Users trust solvers to execute honestly. Malicious solvers could provide worse execution within allowed slippage bounds. Reputation systems and bonding mechanisms mitigate this.

Opacity

Intent execution is less transparent than direct AMM swaps. Users may not know why they got a particular price or which venues were used.

Scalability

As intent volume grows, solver infrastructure must scale. Batch auctions add latency compared to instant AMM swaps.

User Education Needed

Intent systems abstract complexity but can hide risks. Users should understand: (1) they're trusting solvers for execution quality, (2) slippage tolerances still matter, (3) gasless doesn't mean free—costs are embedded in execution. The UX is simpler, but the underlying system is complex.

The Future of Intents

AI-Powered Solvers

Solvers increasingly use machine learning to predict optimal routes, manage inventory, and adapt to market conditions. AI agents may eventually express intents on behalf of users.

Account Abstraction Integration

ERC-4337 (account abstraction) combined with intents enables sophisticated automation: conditional trading, recurring transactions, and multi-step strategies—all from smart contract wallets.

Cross-Domain Composability

Intents could eventually span not just chains but entire ecosystems: "Swap my ETH for an NFT on OpenSea" or "Deposit the optimal yield across all lending protocols."

Investment Considerations

Protocol Tokens to Watch

  • COW: CoW Protocol governance and utility token
  • 1INCH: 1inch aggregator and Fusion mode
  • ACX: Across Protocol for cross-chain intents

Bull Case

  • Intent-based trading is strictly better UX than manual routing
  • MEV protection becomes table stakes; protocols without it lose users
  • Cross-chain intents solve the fragmented liquidity problem
  • First-mover protocols (CoW, UniswapX) could dominate solver networks

Bear Case

  • Solver networks may centralize around well-capitalized players
  • Latency from batch auctions may deter time-sensitive traders
  • Traditional AMMs continue improving (Uniswap V4 hooks, Curve ng)
  • Regulatory uncertainty around solver activities (market making, order routing)

Key Takeaways

  • Intents are declarative statements of desired outcomes, letting solvers handle execution complexity
  • Solvers compete to fill intents with the best execution, accessing liquidity across DEXs, CEXs, OTC, and private inventory
  • MEV protection comes from private orderflow, batch auctions, and coincidence of wants matching
  • Cross-chain intents (ERC-7683) enable seamless multi-chain operations with single user signatures
  • Major protocols include CoW Protocol (batch auctions), UniswapX (Dutch auctions), and 1inch Fusion
  • Intent systems are more user-friendly but introduce new trust assumptions around solver behavior