Why Cross-Chain Bridges Keep Getting Hacked

A cross-chain bridge locks your asset on chain A and mints a representation of it on chain B. To do that, it has to hold the locked collateral somewhere — usually behind a small set of signing keys or a single smart contract. That makes bridges concentrated honeypots: enormous pooled value guarded by a narrow attack surface. That concentration, more than any single bug, is why bridges are hacked so often.

That is the short answer. The rest of this page explains how bridges work, walks through the real cases, and shows what you can actually do to reduce your exposure.

---

How a bridge actually works

Blockchains cannot natively read each other's state. Ethereum does not know what happened on Solana, and vice versa. A bridge is the machinery that fakes a connection between them.

The dominant design is lock-and-mint:

1. You send 10 ETH into the bridge contract on Ethereum. It is locked there.
2. The bridge observes this deposit and mints 10 wrapped ETH on the destination chain.
3. To go back, you burn the wrapped ETH, and the bridge unlocks the original 10 ETH.

The wrapped asset on the far side is an IOU. It is only worth something because the bridge promises the locked collateral still exists and can be redeemed. Break that promise — drain the locked pool, or mint tokens that were never backed — and the wrapped asset collapses.

Everything hinges on who is allowed to authorize a mint or an unlock. That is the bridge's trust model, and it comes in a few flavors:

• Validator / multisig bridges rely on a set of external signers. If enough of them sign, the bridge acts. Ronin worked this way.
• Light-client / relay bridges try to verify the other chain's consensus cryptographically, minimizing trusted humans — more secure in principle, but complex and expensive to build correctly.
• Liquidity-network bridges keep pools on both sides and swap rather than mint, shifting the risk toward pool solvency.

The more the model leans on a handful of keys, the more it behaves like a single vault with a shared password.

---

The hacks that defined the category

These are historical figures, widely reported at the time. Values move as funds are recovered or laundered, so verify current numbers before citing them.

• Poly Network (~$610M, August 2021). An attacker exploited a flaw in how the cross-chain contracts verified instructions and effectively told the bridge to hand over assets. Unusually, most funds were later returned.
• Wormhole (~$320M, February 2022). A flaw in signature verification let the attacker forge approval for a mint of 120,000 wrapped ETH on Solana that was never backed by real collateral. The gap was covered by the backers.
• Ronin bridge (~$625M, March 2022). The bridge used a 9-validator multisig requiring 5 signatures. Attackers gained control of 5 keys — four through a compromised node setup and one via a third party — and simply signed fraudulent withdrawals. This is a pure key-compromise story, not a contract bug.
• Nomad (~$190M, August 2022). A routine upgrade set a trusted root to a value that made almost any withdrawal message pass validation. Once one person noticed, it became a free-for-all: hundreds of copycats replayed the same trick, draining the bridge in hours.

Four incidents, four different root causes — and together they account for well over a billion dollars.

---

Trust models and how they fail

The column that matters is the middle one. Notice how little the failures have in common at the technical level — and how much they have in common at the human level: value was concentrated, and one weak point unlocked all of it.

---

Why the same failures keep recurring

The specifics differ, but the root causes cluster into four buckets:

1. Private key compromise. Multisig thresholds are only as strong as the operational security around every key. Phishing an operator or misconfiguring a node bypasses all the clever cryptography (Ronin).
2. Smart-contract bugs. Bridges are among the most complex contracts deployed, handling serialization, signatures, and cross-chain proofs. Complexity breeds bugs (Wormhole).
3. Upgrade and initialization mistakes. Bridges are frequently upgraded. A single bad deployment can silently disable a critical check (Nomad).
4. Sheer value concentration. All of the above exist elsewhere in crypto, but bridges combine them with the largest single pools of collateral. The payoff justifies the effort, so the best attackers show up.

You cannot fix the fourth cause by writing better code. As long as one contract or one key set guards hundreds of millions, it will remain a target.

---

What you can actually do

You usually cannot audit a bridge yourself. The practical goal is to minimize how much value you expose and for how long.

• Minimize time in the bridge. The riskiest moment is while your value sits as locked collateral or wrapped IOUs. Bridge, use the funds, and move on. Do not treat a bridge as a savings account.
• Prefer native issuance over wrapped. Where an asset is issued natively on your destination chain, use that. Native USDC transferred via Circle's CCTP is burned on one chain and minted on the other by the issuer itself — there is no third-party locked pool to drain, unlike bridged or wrapped USDC.
• Know whose security you are trusting. Before you bridge, ask what the trust model is. A 5-of-9 multisig you have never heard of is a different bet than a battle-tested, audited light-client bridge.
• Do not hold large balances of bridged or wrapped assets long-term. If the bridge behind your wrapped token is exploited, the token can go to zero even though you did nothing wrong. Convert to native assets when you intend to hold.

None of this makes bridging safe. It makes your exposure smaller and shorter, which is the only lever most users control.

---

FAQ

Are cross-chain bridges safe? No bridge is fully safe. They concentrate large amounts of value behind a small number of keys or contracts, which makes them prime targets. Some designs (audited light-client bridges) are meaningfully safer than others (small external multisigs), but you are always trusting the bridge's security model on top of the two chains. Treat bridging as a risk to minimize, not a settled solution.

What was the biggest bridge hack? The Ronin bridge exploit in March 2022 was around $625M, and Poly Network in August 2021 was around $610M — the two largest by reported figures. Wormhole ($320M) and Nomad ($190M) also rank among the worst. These are historical figures; check current sources for exact and updated numbers.

What is a wrapped token and is it risky? A wrapped token is an IOU minted on one chain to represent an asset locked on another — for example wrapped ETH or bridged USDC. Its value depends entirely on the collateral behind it remaining safe and redeemable. If the bridge holding that collateral is hacked, the wrapped token can lose its backing and its value, so wrapped assets carry the bridge's risk on top of the underlying asset's.

How can I reduce bridge risk? Keep the amount and the time you bridge as small as possible, prefer native issuance (like native USDC via CCTP) over wrapped versions, understand which trust model you are relying on, and avoid holding large long-term balances of bridged or wrapped tokens. Convert back to native assets once the cross-chain step is done.

---

Bridges are one piece of a larger picture. If you move between networks often, understand the tradeoffs specific to Layer 2 bridges, keep an eye on the latest hacks and how they happened, and learn why a bridge being audited is not the same as being safe in what a smart contract audit actually covers.