Are Crypto Bridges Safe? Risks, Hacks, and Safer Alternatives
Are Crypto Bridges Safe? Risks and Alternatives
Mar 9, 2026
Crypto bridges have lost users over $2 billion to hacks and exploits. Most of that wasn't bad luck — it came from structural weaknesses in how lock-and-mint bridges are designed. Understanding what makes a bridge risky helps you identify what to avoid — and what a safer architecture looks like.
Why Crypto Bridges Get Hacked
The most common attack surface is the bridge contract itself — a smart contract that holds large amounts of locked tokens on behalf of users. When a bridge accumulates hundreds of millions in a single contract, it becomes an attractive target.
Most major bridge hacks followed one of three patterns:
Lock-and-mint vulnerability
In a lock-and-mint bridge, your tokens are locked on the source chain and a synthetic "wrapped" version is minted on the destination. The synthetic token is only as valuable as the bridge's ability to back it. If an attacker can trigger minting without actually locking the underlying assets — by exploiting the verification logic — they can drain the backing reserves.
This is what caused the Wormhole hack ($320M) and the Nomad hack ($190M). In both cases, attackers found a way to generate unsupported mint events — essentially creating tokens backed by nothing.
Centralised key compromise
Some bridges rely on a small set of trusted validators or multisig keys to confirm crosschain messages. If those keys are compromised — through phishing, insider access, or poor key management — the entire bridge can be drained in a single transaction. The Ronin Network hack ($625M) followed this pattern: attackers gained control of five of the nine validator keys and used them to authorise fraudulent withdrawals.
Smart contract bugs
Bridge contracts handle complex edge cases across chains with different finality rules and block times. The verification logic is difficult to audit completely, and a single misconfiguration can be exploited. The Nomad hack exploited a misconfigured initialisation that accidentally allowed any message to be treated as valid.
The Numbers
Over $2 billion has been lost to bridge exploits since 2021. The four largest incidents:
| Bridge | Year | Amount Lost | Attack Vector |
|---|---|---|---|
| Ronin Network | 2022 | $625M | Validator key compromise |
| Wormhole | 2022 | $320M | Signature verification exploit |
| Nomad | 2022 | $190M | Misconfigured message validation |
| Harmony Horizon | 2022 | $100M | Multisig key compromise |
The concentration in 2022 reflects how quickly cross-chain activity scaled — and how the lock-and-mint architectures of that era weren't designed for the attack surface they created.
What Makes a Bridge Risky
Three structural factors drive most bridge risk:
Custodial pools. If a smart contract holds a large pool of locked assets, that pool is a target. The bigger the total value locked (TVL), the more attractive the exploit. Every dollar locked is an incentive for attackers.
Synthetic asset exposure. When you receive a wrapped or bridge-specific token, your asset's value depends entirely on the bridge's solvency. If the bridge is exploited, the synthetic token can become worthless. Users who received bridged assets from Ronin or Nomad discovered their tokens had no backing after the exploits.
Complex verification logic. Cross-chain message passing requires verifying that an action on Chain A actually occurred before minting or releasing funds on Chain B. Getting this right across dozens of chains with different consensus mechanisms is genuinely hard — and every edge case in the verification path is a potential attack surface.
How Intent-Based Protocols Avoid These Risks
Relay isn't a bridge in the traditional sense — it's an intent-based protocol. The architecture is fundamentally different.
When you submit a crosschain request on Relay:
You specify what you want to receive — the exact output token and amount, on the destination chain
Competitive liquidity providers (fillers) see your request and race to fill it
The winning filler sends the output directly to your wallet on the destination chain
Settlement happens asynchronously — Relay verifies the fill and reimburses the filler from the source chain
What this means for safety:
No large custodial pool. There's no giant contract holding user funds. Fillers use their own working capital, not a shared locked pool. There's no single contract to drain.
No synthetic assets. You receive native tokens — real USDC, real wETH — not a bridge-specific wrapped version that depends on a backing reserve.
No centralised validator set. No validator keys to compromise. The protocol's security doesn't depend on a small group of key holders.
Non-custodial. Relay coordinates execution through smart contracts, but no centralised party ever takes custody of your assets.
The attack surface of an intent-based protocol is structurally smaller than a lock-and-mint bridge.
What to Check Before Using Any Crosschain Protocol
Audit history. Has the protocol been audited by reputable security firms? Are the reports public and recent?
Fill rate. What percentage of orders complete successfully? High fill rate over an extended period is evidence the system works as designed.
Asset type. Will you receive native tokens or synthetic wrapped versions? Native is safer — your asset doesn't depend on a bridge's backing.
Custody model. Does the protocol hold funds in a centralised contract, or does execution happen at the filler level with the protocol only coordinating?
Volume and operating history. $20B+ in volume over multiple years is meaningful signal. New protocols with unproven code carry higher risk by default.
Frequently Asked Questions
Is intent-based crosschain completely without risk?
No system is completely without risk. Smart contracts can contain bugs. Fillers could fail to fill orders — though this triggers automatic reverts, not stuck funds. But the structural risks that caused the largest bridge hacks — custodial pools, synthetic asset exposure, centralised validator sets — are not present in Relay's architecture.
Are all bridges equally risky?
No. Canonical bridges — operated directly by the chain (like the official Base Bridge or Arbitrum Bridge) — inherit the chain's own security model and have a different risk profile than third-party lock-and-mint bridges. Their tradeoff is speed: canonical bridges can take 7 days for withdrawals from optimistic rollups. For everyday use, intent-based protocols like Relay offer the better combination of safety and speed.