The Complete Guide to Intent-Based Cross-Chain Bridges: From Traditional Bridges to the Solver Competition Era
Cross-chain bridge hacks have caused losses exceeding 624M](https://chain.link/education-hub/cross-chain-bridge-vulnerabilities), Wormhole 190M -- behind these numbers lies a fundamental architectural problem: traditional cross-chain bridges lock massive amounts of assets in smart contracts, creating a super honeypot in hackers' eyes.
The emergence of intent-based architecture has fundamentally changed this landscape.
What Is Intent-Based Architecture?
Three Modes of Traditional Cross-Chain Bridges
Before understanding intents, let's look at how traditional bridges work:
Lock-and-Mint: Users deposit assets into a source chain contract, and the bridge mints an equivalent amount of wrapped tokens (e.g., WETH, WBTC) on the destination chain. When redeeming, the wrapped tokens are burned to unlock the original assets. Representative: Wormhole Portal. The problem is that the source chain contract locks a large amount of assets, making it an attack target. 1
Burn-and-Mint: Tokens are burned on the source chain, and an equivalent amount of new tokens are minted on the destination chain, with no locking required. Representative: Circle CCTP. More secure but requires deep integration with the token issuer. 1
Liquidity Pool: Pre-funded pools are maintained across multiple chains. Users deposit on the source chain and withdraw from the destination chain pool. Representative: Stargate. Capital inefficient, and also carries concentrated TVL risk. 1
Intent: From "How to Do It" to "What You Want"
Intent-based architecture represents a fundamental paradigm shift: users only need to declare their "desired outcome" (what), rather than specifying "how to execute" (how). 2
For example: a user simply expresses "I want to receive 100 USDC on Arbitrum," and solvers in the system competitively fulfill this intent.
Three Key Characteristics:
-
Reversed Execution Order: Traditional bridges move user funds first, waiting for finality before releasing on the destination chain. Intent systems reverse this -- solvers immediately deliver assets on the destination chain, with verification happening after the fact. 2
-
Competitive Execution: Multiple solvers bid on the same intent, ensuring optimal pricing. It's like a ride-hailing platform where multiple drivers compete for rides. 3
-
Risk Transfer: Risk shifts from end users to professional solvers. Solvers bear settlement risk in exchange for fee income. 2
ERC-7683: A Unified Standard for Cross-Chain Intents
In 2024, Uniswap Labs and Across Protocol jointly proposed the ERC-7683 standard, defining the CrossChainOrder struct and ISettlementContract interface, enabling different intent systems to share solver networks and infrastructure. 4
As of 2026, over 50 protocols support this standard, including major L2s such as Arbitrum, Optimism, Polygon, and Scroll. In February 2025, the Ethereum Foundation launched the Open Intents Framework (OIF), backed by 30+ teams.
Historical Context: Why Do We Need Intents?
The evolution of cross-chain technology can be divided into five phases:
Phase 1: Early Cross-Chain (2020-2021) -- Dominated by lock-and-mint, with pioneers like Wormhole and Multichain. Simple but fragile.
Phase 2: Liquidity Pool Model (2021-2022) -- Stargate and Hop Protocol introduced pool models to accelerate cross-chain transfers, but with low capital efficiency.
Phase 3: Security Crisis Catalyzes Change (2022) -- Ronin Bridge 320M, Nomad $190M. Cross-chain bridge hacks accounted for approximately 40% of total Web3 losses, prompting the industry to rethink architectural issues. 5
Phase 4: Rise of Intent-Based (2023-2024) -- Across pioneered large-scale intent model implementation, CowSwap's batch auction proved the viability of competitive execution, and UniswapX introduced Dutch auctions. 1
Phase 5: Standardization and Maturation (2025-2026) -- ERC-7683 standard established, Across V4 introduced ZK proofs, protocols began supporting non-EVM chains, and NEAR Intents surpassed $13 billion in cumulative trading volume. 6
Security Model: Intent vs Traditional Bridges
Attack Surface of Traditional Bridges
Traditional bridges face multiple attack vectors: validator private key theft, validator collusion, malicious smart contract upgrades, contract vulnerability exploitation, and RPC endpoint attacks. In Q1 2025, private key leaks accounted for 88% of stolen funds. 5
Security Advantages of Intent-Based Architecture
- No TVL Honeypot: deBridge DLN adopts a zero-TVL design where solvers use their own funds, and the protocol itself does not lock large amounts of assets. 7
- Risk Diversification: Risk is distributed across multiple solvers rather than concentrated in a single contract.
- Aligned Economic Incentives: Solvers are only compensated after correctly fulfilling an intent.
- Minimal Trust Assumptions: Across only requires "one honest participant" to maintain security. 8
New Risks
Intent-based architecture also introduces new risk categories:
- Liquidity Exhaustion Attacks: Exploiting natural liquidity fluctuations to disrupt system availability. 9
- Solver Centralization: If the solver network is insufficiently decentralized, it may create new centralization risks.
- Oracle Dependency: For example, Across relies on the UMA Optimistic Oracle, making oracle security a critical factor. 8
In-Depth Comparison of Nine Intent-Based Protocols
1. Across Protocol -- Market Leader
Across is currently the largest intent-based cross-chain bridge, accounting for 54% of all daily active bridge users, having processed over $28 billion in cumulative volume, with zero exploit record. 10
Technical Architecture:
- Users declare intents via Request-for-Quote
- A decentralized relayer network competitively bids and immediately fronts funds
- UMA Optimistic Oracle verifies repayment correctness: 1-hour challenge window; if unchallenged, the claim is accepted. 8
V4 Upgrade (July 2025) introduced ZK proofs to trustlessly verify repayments, enabling new chains to be added within hours. 10
| Metric | Value |
|---|---|
| Supported Chains | 18+ |
| Speed | Median 2 seconds |
| Fees | 0.06-0.12% |
| Cumulative Volume | $28B+ |
| Security Record | Zero exploits |
Sources: Across Protocol Review 2026, CoinMarketCap, UMA Case Study
2. NEAR Intents -- Widest Chain Coverage Intent Protocol
NEAR Intents (formerly Defuse Protocol) is a multi-chain trading protocol built on NEAR Protocol, where users declare intents and solvers compete to fulfill them. 11
Three-Phase Architecture: Express, Solve, Settle
- Express: Users create signed messages describing their desired outcome
- Solve: The solver network competes to provide the best quote, with Solver Relay waiting up to 3000ms to collect quotes
- Settle: NEAR's Verifier contract (
intents.near) settles on external blockchains using Chain Signatures (MPC threshold cryptography) 12
Confidential Intents (launched February 2026) use NEAR's private shard and TEE to prevent front-running and sandwich attacks, causing NEAR token to surge 17%.
| Metric | Value |
|---|---|
| Supported Chains | 35+ |
| Speed | 2-3 seconds |
| Protocol Fees | 0.0001% |
| Cumulative Volume | $13B+ |
| Cumulative Swaps | 15.7M+ |
Sources: NEAR Intents Docs, Yahoo Finance, Simply Staking
3. UniswapX -- Deepest Liquidity Intent Swap
UniswapX is an auction-based swap protocol launched by Uniswap, using a Dutch auction mechanism where fillers compete to fulfill swaps. 13
Dutch Auction Mechanism:
- Users sign off-chain orders
- Order prices start slightly above the target and decrease over time
- Fillers fill orders when the price becomes profitable
- Users enjoy a gasless swap experience
Cross-chain bridging was launched in October 2024, deeply integrated with ERC-7683.
| Metric | Value |
|---|---|
| Supported Chains | 36+ |
| Fees | Implicit in Dutch auction spread |
| 30-Day Volume | ~$148.5B (including Uniswap) |
| TVL | ~$4.5B (all Uniswap) |
Sources: UniswapX Docs, Uniswap Blog, CoinLaw Statistics
4. deBridge DLN -- Zero-TVL Security Pioneer
deBridge Liquidity Network adopts an aggressive "zero-TVL" design: Makers (users) create orders, and Takers (solvers) fulfill them on the destination chain with their own funds. No shared liquidity pools, no wrapped tokens. 7
deBridge Bundles (launched December 2025) package multiple cross-chain operations into a single atomic transaction.
| Metric | Value |
|---|---|
| Supported Chains | 27+ (including Tron) |
| Speed | Median 1.96 seconds |
| Fees | 4 bps + gas |
| Cumulative Volume | $9.96B+ |
| TVL | Zero (by design) |
Sources: Phemex, Baltex Review, deBridge Docs
5. CowSwap -- Strongest MEV Protection
CowSwap uses a batch auction + solver competition model. Orders are collected into approximately 30-second batches, solvers submit solutions, and the solver that provides the most surplus to users wins execution rights. 14
Coincidence of Wants (CoW) Matching: Direct P2P matching -- Alice swaps USDC->ETH while Bob swaps ETH->USDC simultaneously, without pool routing. Ring Trade supports multi-party circular swaps.
MEV Protection: Off-chain order placement avoids the public mempool, and Uniform Clearing Price ensures all participants in the same batch receive the same price, making transaction reordering unprofitable. 14
| Metric | Value |
|---|---|
| Supported Chains | 10+ |
| Speed | ~30 seconds (batch cycle) |
| 2025 Volume | $87B |
| MEV Protection | Industry-leading |
Sources: Shoal Research, CoW DAO 2025 Review, CoW Protocol Docs
6. 1inch Fusion+ -- Atomic Cross-Chain Swap
1inch Fusion uses Dutch auction + resolver network, with Fusion+ focusing on cross-chain atomic swaps. Three-phase process: Announcement -> Deposit (resolver deposits assets in escrow on both source and destination chains) -> Withdrawal (secret reveal unlocks funds). 15
Hashlock-Timelock Mechanism ensures fund safety: both escrows are linked via cryptographic hash, with timelock expiration enabling cancellation and refund.
| Metric | Value |
|---|---|
| Supported Chains | 13+ |
| Speed | < 5 minutes |
| Fees | Completely gasless |
| Cumulative Volume | $25B+ |
| Resolver Count | ~10 (capped) |
Sources: 1inch Fusion+, 1inch Blog, 1inch Deep Dive
7. Chainflip -- Native Bitcoin Swap
Chainflip uses a unique JIT (Just-In-Time) AMM model, where market makers compete in real-time to provide the best price for each swap. No wrapped tokens are used -- native assets are delivered directly. 16
Core Innovation: Flipping frontrunning on its head -- liquidity providers compete to serve users rather than attack them. 150 validators jointly control vaults on each chain via the FROST threshold signature scheme.
| Metric | Value |
|---|---|
| Supported Chains | 5 (BTC, ETH, SOL, BNB, Tron) |
| Speed | ~60 seconds |
| Fees | 0.10% + spread |
| Cumulative Volume | $6.45B |
| TVL | $10.6M |
Sources: Chainflip Docs, Chainflip Q4 2025
8. Squid Router / Axelar -- Most Chain Support
Squid evolved from a bridge aggregator to an intent protocol. In March 2026, it launched Squid Intents -- core logic is processed within TEE (Trusted Execution Environment), with blockchain serving only as the settlement layer. 17
| Metric | Value |
|---|---|
| Supported Chains | 100+ |
| Cumulative Volume | $6B+ |
| Security Record | Zero exploits |
Sources: Squid Router, Axelar Blog
9. Everclear -- The Clearing Layer for Bridges
Everclear (formerly Connext) is not a user-facing bridge but rather a backend clearing layer for other bridges. It uses an Arbitrum Orbit rollup as the clearing chain, Hyperlane for cross-chain communication, and EigenLayer for security. 18
| Metric | Value |
|---|---|
| Role | Clearing Layer |
| Tech Stack | Arbitrum Orbit + EigenLayer + Hyperlane |
| TVL | ~$1.13B |
Sources: The Block - Everclear Rebrand, The Block - Mainnet Launch
Comprehensive Comparison
| Protocol | Execution Mechanism | Supported Chains | Speed | Fees | Cumulative Volume | Security Model |
|---|---|---|---|---|---|---|
| Across | Relayer + UMA Oracle | 18+ | ~2s | 0.06-0.12% | $280B+ | Optimistic Oracle |
| NEAR Intents | Solver + Chain Signatures | 35+ | 2-3s | 0.0001% | $13B+ | MPC + Verifier |
| UniswapX | Dutch Auction + Filler | 36+ | Seconds | Implicit spread | - | Filler Competition |
| deBridge | Maker-Taker P2P | 27+ | ~2s | 4bps | $9.96B+ | Zero-TVL |
| CowSwap | Batch Auction + Solver | 10+ | ~30s | Gas only | $87B/yr | Off-chain + UCP |
| 1inch Fusion+ | Dutch Auction + Resolver | 13+ | <5min | Gasless | $25B+ | Hashlock-timelock |
| Chainflip | JIT AMM | 5 | ~60s | 0.10% | $6.45B | FROST Threshold |
| Squid/Axelar | TEE + Auction | 100+ | <1s | Variable | $6B+ | TEE + Verifier |
| Everclear | Clearing Layer | 7+ | <60s | <20bps | - | EigenLayer |
Which One Should You Choose? Use Case Recommendations
| Need | Recommendation | Reason |
|---|---|---|
| General Cross-Chain Transfers | Across | Fastest, most secure, largest market share |
| Cross-Chain Token Swaps | UniswapX | Deepest liquidity, best pricing via Dutch auction |
| Stablecoin Cross-Chain | deBridge DLN | Lowest fees, Zero-TVL, supports Tron |
| Anti-MEV Trading | CowSwap | Batch auction + UCP for strongest MEV protection |
| Gasless Cross-Chain | 1inch Fusion+ | Completely gasless, atomic swap security |
| Native BTC Swaps | Chainflip | The only true native BTC swap |
| Maximum Chain Support | NEAR Intents | 35+ chains + Confidential Intents |
| Developer Integration | Socket/Bungee | MetaMask default solution, zero platform fees |
Conclusion: Intents Are the Future of Cross-Chain
Intent-based architecture is not just a technological advancement but a philosophical shift -- from "telling the system how to do it" to "telling the system what you want."
With the maturation of the ERC-7683 standard, the expansion of solver networks, and the addition of privacy features like NEAR Confidential Intents, we are witnessing cross-chain infrastructure evolve from a "dangerous necessary evil" into a "seamless background service."
For everyday users, the most important change is this: you no longer need to understand how a bridge works -- you just need to state the outcome you want.
References
Footnotes
-
Uniswap Labs and Across Propose Standard for Cross-chain Intents ↩
-
Cross-Chain Bridge Vulnerabilities - Chainlink / Cross-Chain Bridge Security - Hacken ↩ ↩2
-
Exploiting Liquidity Exhaustion Attacks in Intent-Based Bridges - arXiv ↩
-
CoW Swap: Intents, MEV, and Batch Auctions - Shoal Research ↩ ↩2
-
Intents: Executing Cross-Chain Transactions in 1 Second - Axelar Blog ↩

