How a bridge works
The mechanism
Lock and mint, in four steps.
You send the asset to a bridge contract on the origin chain, which locks it. The bridge confirms this happened. It then mints an equivalent token on the destination chain and sends it to your address there. Going back reverses the sequence: the destination token is burned and the original unlocks.
Nothing crosses. There are two ledgers, and the bridge keeps them consistent. What arrives on the far side is a wrapped token, which is a claim on the locked original rather than the original itself.
The two designs
Trusted bridges rely on an operator holding the locked assets and issuing the claims. Simple, fast, and it works exactly as well as the operator does. If they lose the reserve, the wrapped tokens back onto nothing.
Trustless bridges replace the operator with code and a validator set. No single party can take the reserve, and the guarantee now depends on the code being correct, which brings us to the record.
Where the risk sits
Why bridges are the most attacked infrastructure in crypto
They combine the two properties an attacker wants: large balances and complex code.
Individual bridge exploits have run into hundreds of millions of dollars each, and the category total exceeds two billion. The reasons repeat: a flaw in signature verification, a compromised validator key, a logic error in how deposits are counted. The immutability that makes a smart contract trustworthy also means a bug in a bridge stays exploitable until somebody notices.
This is not an argument against ever using one. It is an argument for using established bridges, for moving amounts you can afford to lose while you learn how one behaves, and for not leaving balances parked in wrapped form longer than needed.
The alternatives worth knowing
Two, and both are usually better.
A centralised exchange
Deposit on one network, withdraw on another. The exchange functions as a bridge with a business behind it, and for most users this is both cheaper and less risky than a bridge contract.
Native issuance
Where the asset exists natively on both chains, no bridge is needed at all. USDT on TRON is not bridged USDT: it is a separate contract issued by Tether, and moving between the two is a matter of selling and buying rather than wrapping. The same applies to USDC across its supported chains.
Why the distinction matters
That distinction matters more than it looks. Native issuance carries issuer risk you already accepted. Bridged versions add a second layer of risk on top.
What a merchant needs
Usually nothing.
A payment gateway supporting multiple networks accepts natively issued assets on each, so a customer paying USDT on TRON and a customer paying it on Ethereum both go through without a bridge existing anywhere in the flow. The setup guide covers how networks are chosen.
Where bridges enter is treasury movement: a business holding funds on one chain and needing them on another. Even then an exchange is usually the simpler route.