Bridges

Cross-chain bridges are the connective tissue of Web3 — and historically its single most exploited component. More than $2.5B has been lost to bridge hacks since 2021, which is why the right question is never simply "which bridge is cheapest" but "which trust model am I comfortable underwriting for the size and duration of this transfer." This hub groups every bridge we cover by architecture so you can reason about that trade-off directly.

There are four architectures worth distinguishing. Canonical (native) L2 bridges are operated by the rollup itself and inherit the security of the underlying chain — the safest option, but withdrawals back to Ethereum can take up to seven days on optimistic rollups. Intent-based bridges like Across use bonded relayers who front you the destination asset in seconds and reconcile later against an optimistic oracle; they are fast and capital-efficient but introduce relayer and oracle assumptions. Liquidity-network bridges such as Stargate move value through pools on each chain and can deliver native assets via burn-and-mint standards like CCTP. Lock-and-mint bridges issue wrapped representations and concentrate risk in large honeypot pools — the model behind most historical exploits.

For everyday transfers, the practical decision tree is short. Moving major assets between EVM L2s? An intent bridge, or a bridge aggregator that can route to one, usually gives the best combination of speed and price. Need native USDC on the destination rather than a wrapped variant? Route through CCTP-backed liquidity bridges. Bridging to or from a non-EVM chain like Solana? Coverage, not cost, becomes the deciding factor, and Wormhole or a solver-auction aggregator is often the only credible path. Making a very large, non-urgent transfer to Ethereum mainnet? The canonical bridge's longer wait buys you the strongest security guarantee available.

Every comparison in this silo applies the same security-first methodology: we describe the trust model in plain terms, name the specific failure mode (relayer compromise, validator collusion, wrapped-asset concentration), and only then weigh fees and speed. We never rank a bridge higher because it pays a larger referral — affiliate relationships are disclosed and never influence ordering.