11 — Bridges and Interoperability

How to move value/messages between chains. Bridge hacks = the largest category of loss in crypto. Models: lock-mint, burn-mint, atomic swap, IBC, restaking-based, intent-based.

1. Why bridges?

Crypto is fragmented: BTC on Bitcoin, ETH on Ethereum, USDC on N chains, etc. For the user:

• Move assets between chains.
• Liquidity unified.
• Interoperable apps (DeFi composability cross-chain).

Bridges solve this — but with diverse trust assumptions and a catastrophic history of hacks.

2. Bridge models

Lock-and-Mint

Asset locked on chain A; "wrapped" version minted on chain B.

Chain A: lock 1 ETH in bridge contract
Chain B: mint 1 wETH (representing the locked ETH)
[reverse: burn wETH on B → unlock ETH on A]

Trust: bridge custodian honest + secure.

Burn-and-Mint

Native asset burned on chain A; minted on B. Used for canonical bridges (USDC CCTP, USDT recent).

Requires issuer (Circle, Tether) participation.

Liquidity pool / swap

Two-sided liquidity on both chains. User swaps on A; liquidity provider releases on B. Stargate (LayerZero) model.

Atomic swaps

HTLC-based. User and counter-party reveal H(x) = h; both legs settle. No bridge custody.

Limitations: requires counter-party on both chains; slow; terrible UX.

IBC (Inter-Blockchain Communication)

Cosmos native. A light client of each chain runs on the other chain → trustless via consensus proofs. The relayer can be anyone (no permission).

Optimistic bridges

Like optimistic rollups: assert + challenge window. Across, Connext.

ZK bridges

Validity proofs. Polyhedra, Succinct (formerly Telepathy), zkBridge.

Restaking-based

EigenLayer AVS provides decentralized validators for the bridge. EigenLayer + bridge = pooled security.

3. Major bridges

Wormhole

Multi-chain message protocol, originally Solana ↔ Ethereum.

• Guardian network: 19 entities sign messages.
• Supports 30+ chains.
• Feb/2022 hack: US$ 320M minted via a signature bypass bug. Jump Trading bailed it out.
• Wormhole token (W) airdropped 2024.
• Native tokens: integrated with Circle CCTP, etc.

LayerZero

Bryan Pellegrino 2021. Ultra-light node model:

• Oracle (Chainlink originally) submits block header.
• Relayer (specific to the app) submits proof of message.
• Both must agree → the app accepts.

Adopted by: Stargate (own canonical product), thousands of dApps.

Concerns: oracle + relayer collusion = compromise. Configurable per-app.

ZRO token launched Jun/2024.

Axelar (AXL)

Cosmos SDK chain that serves as a cross-chain hub. Validators sign cross-chain messages. Lighter than running each chain's light client. Used by Squid Router, Lifi.

Chainlink CCIP (Cross-Chain Interoperability Protocol)

Chainlink's offering, 2023. Decentralized Oracle Network + Active Risk Management (anti-fraud). Enterprise positioning.

Stargate (LayerZero canonical)

Composable cross-chain liquidity. Uses unified liquidity pools — not wrapped tokens. Top USDC/USDT cross-chain bridge.

Across

Risk Labs (UMA team). Optimistic bridge with relayers fronting capital. Cheapest ETH bridge for L1 ↔ L2. Across token (ACX).

Hyperlane

Modular interoperability. Permissionless deployment of cross-chain messaging between any chains. ISMs (Interchain Security Modules) configurable.

Connext

Modular. Atomic swaps (CRT-style). Pivot to intent-based.

Synapse Protocol

Multi-chain bridge + AMM. SYN token.

Multichain (defunct)

Earlier formerly "AnyCallV7", Multichain network. Halted Jul/2023 after CEO Zhaojun disappeared (arrested in China). US$ 130M+ stuck/lost.

Hop Protocol

L2 ↔ L2 specialist for the Ethereum stack. HOP token.

Bitcoin bridges

• WBTC (BitGo custody) — biggest BTC representation on Ethereum. ~150k WBTC in circulation. Controversial 2024 management transition to Tron-aligned BiT Global.
• tBTC (Threshold Network) — decentralized.
• renBTC (defunct Ren Protocol).
• cbBTC (Coinbase, 2024) — new contender.

4. IBC (Inter-Blockchain Communication)

Cosmos native, RFC-style spec, native on all Cosmos SDK chains.

How it works

1. Light client of chain A runs on chain B (and vice-versa).
2. State changes on A produce a commitment proof.
3. Relayer (off-chain, anyone) submits the proof to B.
4. Chain B verifies the proof against A's light client → accepts.

Trustless: cryptographic verification, no third-party custody.

Apps

• ICS-20: Fungible token transfer.
• ICS-721: NFT transfer.
• ICS-27: Interchain accounts (control an account on another chain).
• ICS-29: Fee middleware.
• ICS-32: NFT counterpart.

Adoption

100+ chains connected (Osmosis, Cosmos Hub, dYdX, Injective, Celestia, etc.). ~US$ 1B+ flowing monthly via IBC.

Limitations

• Only IBC-enabled chains.
• Ethereum mainnet is not IBC-native (would need a light client of a Cosmos chain on Ethereum — gas-prohibitive). Bridges like Composable, Gravity Bridge attempt it.

5. CCTP (Circle Cross-Chain Transfer Protocol)

Circle (USDC issuer), 2023. Native burn-and-mint for USDC.

• Burn USDC on chain A.
• Circle attests.
• Mint USDC on chain B.

No wrapped USDC (no USDC.e versions). Solves the fragmentation problem.

Supported chains expanding: Ethereum, Arbitrum, Optimism, Base, Polygon, Avalanche, Solana (2025), Sui, Aptos.

Tether doing similar with native USDT migrations 2024-2025.

6. Native bridges (canonical for L2s)

L2 to Ethereum L1 = native rollup bridge.

Optimistic Rollup bridges

Deposit: instant (L1 → L2). Withdraw: 7-day challenge window (or use a fast bridge like Across, paying a premium).

ZK Rollup bridges

Deposit: instant. Withdraw: post-proof verification (~hours).

Liquidity bridges atop native

Across, Hop, Synapse let users skip the 7-day window by paying a liquidity provider fee.

7. Intent-based architectures

New paradigm 2024-2026: the user states an intent (e.g., "I want 1 ETH on Base"), solvers compete to execute. Cross-chain abstracted away.

Examples

• Across V3 (intent settler).
• Anoma (intent-centric L1).
• Brink (intent infrastructure).
• Skip Protocol (Cosmos intent + interchain swaps).
• Bungee, Squid Router, LIfI: cross-chain swap aggregators with intent capabilities.
• CowSwap (intra-chain intent).
• dYdX Chain Abstraction initiatives.

User benefits: 1-click cross-chain, MEV protected, gasless options.

8. EigenLayer-based bridges

Decentralized validation via restaking:

• Omni Network: cross-rollup interop via EigenDA + EigenLayer.
• Hyperlane (optional restaking config).
• Lido CSM (Community Staking).

Trust assumption: 1/3 of restaked ETH validators must be honest.

9. Bridge hacks — the most expensive category in crypto

The Lazarus Group (N. Korean state APT) is credited with the vast majority of bridge hacks 2022-2025.

Why bridges are vulnerable

1. High-value honeypot: holds assets from two chains.
2. Complex multi-chain logic: more attack surface.
3. Trust assumption sold as "decentralized" but frequently a small validator set.
4. Signature verification bugs: state machine errors in complex flows.

Vitalik's view

2022 reddit post: "I think bridges are fundamentally limited compared to L2s..." He argued a multichain world is more vulnerable than a rollup-centric one.

10. Trusted vs trustless tradeoff

11. CCIP, Wormhole, LayerZero — comparison

12. Cross-chain TVL

Total locked in bridges (excluding canonical L2 bridges):

• ~US$ 30B in production bridges (May/2026).
• WBTC alone ~US$ 10B equivalent.
• LayerZero ecosystem ~US$ 5B locked.
• Stargate ~US$ 1B.

Plus canonical L2 bridges: ~US$ 50B in ArbitrumOptimismBase/zkSync L1 bridges.

13. Best practices for using bridges

✅ Use canonical bridges when possible (L2 official, CCTP, native IBC). ✅ Verify the URL (phishing common). ✅ Approve only the required amount (not infinite approval). ✅ Verify the destination chain ID. ✅ Test with a small amount first. ✅ Wait for the recommended confirmations (~6 BTC, ~32 ETH blocks for finality). ❌ Don't use random new bridges without audit + bounty. ❌ Don't bridge highly volatile thin liquidity pairs.

14. Future trends

Native chain abstraction

The user doesn't think in "chains". Wallet + intents handle multi-chain seamlessly. NEAR is pushing the thesis hard.

Cross-rollup interop

Ethereum L2 ↔ L2 today requires a bridge. Pushes for:

• Native rollup standard (EIP-7702, etc.).
• Shared sequencers (Espresso, Astria) → atomic cross-L2.
• OP Superchain (intra-OP-Stack).
• AggLayer (intra-Polygon-CDK).

PQC bridges

Bridge cryptography must migrate to PQC. Long-term assets at HNDL risk.

Unified liquidity

Stargate-style + CCTP-style: a single canonical version of stablecoins across chains.

Restaking interop

EigenLayer AVS-based bridges multiplied 2024+ (e.g., Polyhedra, Omni).

15. WormholeLayerZeroAxelar code samples

Sending a message via LayerZero (OFT — Omnichain Fungible Token)

contract MyOFT is OFT {
    function send(uint16 dstChainId, bytes memory to, uint256 amount) external payable {
        _send(msg.sender, dstChainId, to, amount, payable(msg.sender), address(0), bytes(""));
    }
}

IBC ICS-20 transfer (Cosmos)

msg := &types.MsgTransfer{
    SourcePort:    "transfer",
    SourceChannel: "channel-0",
    Token:         sdk.NewCoin("uatom", math.NewInt(1000000)),
    Sender:        senderAddr.String(),
    Receiver:      receiverAddr.String(),
    TimeoutHeight: clienttypes.NewHeight(0, 100),
}

16. Cross-reference

• L1s connecting via bridges: 04-l1-bitcoin.md, 05-l1-ethereum.md, 06-l1-alt.md.
• L2 native bridges (rollup withdraw): 07-l2-scaling.md.
• Detailed hacks: 14-incidents.md.
• Restaking + EigenLayer: 07-l2-scaling.md §Restaking.
• IBC + Cosmos: 06-l1-alt.md §Cosmos.