Crypto bridges move tokens and data between blockchains that cannot talk to each other on their own. They are also one of the most attacked parts of crypto. This article explains how bridges work and what the numbers show about hacks.
Cumulative losses from bridge exploits have passed $4.3 billion since 2021, according to a 2026 guide to blockchain bridges. Bridges also accounted for two of the three largest DeFi exploits in the first quarter of 2026. Each section below covers one part of the story.
What Is a Crypto Bridge?
Blockchains such as Bitcoin, Ethereum and BNB Smart Chain run on different rules. They use different consensus methods, programming languages and token standards. By default, they cannot send messages or assets to one another.
A crypto bridge is a protocol that links two separate blockchain networks. It lets a user move assets or data from one chain to the other. The ability to do this is called interoperability.
Guides often compare a bridge to a physical bridge between two islands. Each island is its own blockchain ecosystem. Without a bridge, assets and users stay stuck on their own island.
There are practical reasons to cross. Users can reach apps and DeFi services that only exist on another chain. Fees also differ from one network to the next, so some users move assets to a cheaper chain for certain transactions.
Developers gain from bridges too. They can build apps that use the strengths of more than one blockchain. Curve Finance and Aave, for example, use bridges to let users move liquidity between Ethereum and networks such as Polygon and Avalanche.
The other option is to sell the asset and buy a new one. That costs transaction fees and exposes the user to price swings. A bridge lets the user keep the original asset while gaining access to a new network.
Usage data shows people rely on bridges. Chainalysis reported that monthly bridged volume in 2024 ranged from $1.5 billion to $3.2 billion. The table and chart below sum up the main numbers.
| Measure | Figure |
|---|---|
| Monthly bridged volume, 2024 | $1.5 billion to $3.2 billion |
| Bridge transactions in Chainalysis data | 220 million |
| Blockchains covered by Chainalysis | More than 25 |
| Assets covered by Chainalysis | 17 million |
| Assets supported by Trust Wallet | 10 million+ across 100+ blockchains |
How Crypto Bridges Work
Most bridges move value by locking and minting. The user sends the original token to a bridge contract on the source chain. The contract locks the token and registers the deposit.
The bridge then sends a secure message to a matching contract on the destination chain. That contract creates a wrapped token. The wrapped token is pegged to the value of the original.
Bridges do not carry the original coin across. Instead, the user gets a stand-in token, such as Wrapped Bitcoin (WBTC), an ERC-20 token that works on Ethereum. The original BTC stays locked in smart contracts.
Reversing the transfer uses burning and unlocking. The user sends the wrapped token back to the bridge contract, which removes it from circulation. A message then goes to the source chain, and the original asset is released.
Chainalysis gives an example using Ethereum and Solana. An asset is locked on Ethereum, a message goes to Solana, and a wrapped copy is created there. The copy can be used on Solana as if it were native.
For the user, the steps are short. Pick the source chain and amount, send the assets to the bridge, and receive wrapped tokens on the target chain. Behind the scenes, the system relies on messages, proofs, validators, contracts and liquidity pools.
That last point matters for security. If one link in the chain breaks, funds can leave fast. Later sections show how that has played out.
| Step | Lock and mint (forward) | Burn and unlock (reverse) |
|---|---|---|
| 1 | Asset is sent to the bridge contract on the source chain | Wrapped asset is sent to the bridge contract on the destination chain |
| 2 | Contract locks the asset and registers the deposit | Contract burns the wrapped asset and registers the burn |
| 3 | Message is sent to the destination chain | Message is sent back to the source chain |
| 4 | Wrapped asset is minted | Original asset is unlocked |
Types of Crypto Bridges
Bridges can be sorted by who runs them, how they work and which direction they move assets. The most common split is between trusted and trustless designs. Each has trade-offs.
A trusted, or custodial, bridge is run by a central entity. Users must trust the operator to keep assets safe. These bridges may offer a smoother experience, but the operator could misuse funds.
A trustless, or non-custodial, bridge uses smart contracts to lock and release tokens. There is no central operator. Security then depends on the quality of the code, and these bridges can be more complex and slower.
Wrapped asset bridges turn an asset from one chain into a compatible token on another. WBTC is the best-known example. Sidechain bridges link a parent chain to a child chain that uses a different consensus method, such as the link between Ethereum and Gnosis Chain.
Direction is another factor. One-way bridges only move assets to the destination chain, with no way back. Two-way bridges work in both directions and suit general use better.
Some wallets offer cross-chain swaps as another route. The wallet uses an aggregator to search decentralized exchanges for the best rate. It may use bridges behind the scenes without the user handling them.
| Type | How it works | Main trade-off |
|---|---|---|
| Trusted (custodial) | A central operator holds locked assets | Relies on the operator |
| Trustless (non-custodial) | Smart contracts lock and release tokens | Code flaws can be exploited |
| Wrapped asset | Issues a pegged token on another chain | Depends on the peg and redemption |
| Sidechain | Links a parent chain to a child chain | Used when consensus methods differ |
| One-way | Moves assets to the destination only | No return path |
| Two-way | Moves assets in both directions | More flexible, more to secure |
How Much Has Been Lost to Bridge Hacks?
Bridge exploits have cost users billions. A 2026 guide put cumulative losses at more than $4.3 billion since 2021. That makes cross-chain infrastructure one of the most targeted areas in crypto.
An earlier Chainalysis analysis counted $1.4 billion in losses from the start of 2022. By then, the amount stolen from bridges had already passed $1 billion. The total has kept growing since.
In 2024, access control flaws were the leading cause. They were responsible for roughly 75% of the value lost in bridge incidents that year, or $1.7 billion. A later section looks at this in more detail.
The largest single event in the sources is the $615 million Ronin exploit. Ronin was a bridge that supported the game Axie Infinity. Wormhole lost about $320 million after a weakness in digital signature checks.
The Kelp DAO attack, reported at about $290 million, was the largest bridge-related exploit of 2026 as of May. Horizon lost about $100 million in June 2022. The chart ranks these cases by size.
Size alone does not show how serious a flaw is. Hyperbridge lost far less, but only because its liquidity was limited. A bug that drains $2.5 million today could drain much more if the pool grows.
| Incident | Reported loss | Cause as reported |
|---|---|---|
| Ronin | $615 million | Attacker controlled five of nine validators |
| Wormhole | About $320 million | Weakness in digital signature validation |
| Kelp DAO | About $290 million (alleged) | LayerZero setup with a single DVN, per LayerZero; Kelp disputes this |
| Horizon | About $100 million | Cause not detailed in the sources |
| Verus | About $11.6 million | Missing source-amount validation |
| Hyperbridge | About $2.5 million | Forged cross-chain message |
Bridge Hacks in 2026
Several bridge incidents hit in 2026. The largest was Kelp DAO, where attackers allegedly took about $290 million. TechRadar reported that the attack targeted Kelp’s LayerZero setup, and some researchers linked it to the Lazarus Group.
LayerZero reportedly said the problem came from Kelp’s configuration, including its use of a single DVN. Kelp disputed that view. The case shows that security depends on how a project sets up its messaging layer, not only on the protocol itself.
Verus Protocol’s Ethereum bridge was exploited for about $11.6 million. Blockaid and PeckShield flagged it on May 18. A fake cross-chain message tricked the bridge into sending reserves to the attacker’s wallet.
The attacker drained ETH, USDC and tBTC and converted them into roughly 5,402 ETH. Blockaid said the cause was missing source-amount validation in the bridge’s Solidity code. It said the cause was not an ECDSA bypass, a notary key compromise or a parser bug.
Hyperbridge was hit by a forged cross-chain message. The attacker gained control of a bridged DOT token contract, minted 1 billion bridged DOT and sold them into available liquidity. Initial reports said $237,000 was lost, and a later review put realized losses near $2.5 million.
The sources differ on timing. One guide dates the Hyperbridge loss to March 2025, while a 2026 report places it in April.
AFX Trade lost about 24.15 million USDC on July 22. An attacker obtained the five validator signatures needed to approve a withdrawal from its bridge. This matches the validator takeover pattern covered in the next section.
| Incident | Reported loss | Reported cause |
|---|---|---|
| Kelp DAO | About $290 million | LayerZero configuration, single DVN |
| AFX Trade | About $24.15 million | Five validator signatures obtained |
| Verus | About $11.6 million | Fake message, missing validation |
| Hyperbridge | About $2.5 million | Forged message, minted bridged DOT |
Why Bridges Get Hacked
A bridge must answer one hard question. Did something really happen on another chain? If the bridge gets the answer wrong, money can move when it should not.
Bridges are attractive targets for three reasons. They hold large reserves, they connect several ecosystems, and they turn a checking mistake into a direct withdrawal. They also rely on complex smart contract logic spread across more than one chain.
One analysis splits attacks into two groups. Code attacks exploit flaws in smart contracts. Network design attacks are often carried out through social engineering.
Most bridge hacks aim to release tokens on one chain with no matching deposit on the other. Three methods appear again and again. The table lists them with examples.
Access control flaws were the leading attack vector in 2024, at roughly 75% of the value lost, or $1.7 billion. The rest came from other causes. Custodial bridges add another risk, which is that the operator could misuse funds.
DeFi is also more connected than it was in 2022, when Nomad and Wormhole were hit. There are more chains, wrappers and message layers now. That gives a small validation gap more places to turn into a large loss.
| Method | What happens | Example |
|---|---|---|
| False deposit events | The bridge sees a deposit event with no real deposit, or one made with a worthless token | Qubit, where a legacy deposit function allowed fake deposits |
| Validator takeover | Attackers control enough validators to approve fraudulent transfers | Ronin, where five of nine validators were controlled |
| Fake deposits that pass validation | A forged deposit is accepted as real | Wormhole, about $320 million, via a signature validation weakness |
How Stolen Funds Move Through Bridges
Bridges also help move stolen money. They link chains, which makes the path harder for law enforcement and security teams to follow. Chainalysis points to fragmented data and obscured transaction paths as the main problems.
After the FTX collapse, hackers stole about $477 million from the exchange and began to launder it. They used peel chains, CoinJoins and bridges to hide where the funds came from and where they went. The aim was to increase anonymity.
Money laundering usually has three stages. Placement puts dirty money into the financial system. Layering moves it through many transactions to hide the source, and integration is when the funds can no longer be easily traced to crime.
Integration is a key moment for detection, because crypto meets cash. Exchanges that follow anti-money laundering rules check identity and the source of funds. Other platforms care less about these rules.
Bridge transactions leave long trails across several smart contracts and steps. Chainalysis says its Reactor tool can follow funds across hundreds of bridge protocols and decentralized exchanges. Its data covers more than 25 blockchains, 17 million assets and 220 million bridge transactions.
AML rules are one layer of defense. They help detect the illegal use of crypto, but they do not stop an exploit on their own.
| Stage | What it means |
|---|---|
| Placement | Dirty money enters the financial system, often in smaller amounts |
| Layering | Funds move through many transactions, including bridges, to hide the source |
| Integration | Funds can no longer be easily linked to crime; crypto meets cash |
Other Security Incidents From July to October 2026
Bridges are not the only weak spot. A Santiment report covered a run of incidents from July to early October 2026. They involved exchanges, hardware wallets, bridges and software wallets.
Bitget reported a breach on September 24, with losses of about $387.5 million, up from an early estimate of $351.6 million. Mandiant and SlowMist traced it to compromised third-party security products. Bitget said its cold wallets were unaffected and its protection fund would cover users.
The Tectonic exploit followed on August 30. Cronos put affected borrowing at about $120.4 million. Validators rolled back 10,961 blocks and reversed about $111.2 million, while roughly $9.19 million had already left the network.
On September 6, a bug in the Liquid Network let about 4,000 LBTC be created without backing. Blockstream said its reserve fell from about 4,205 BTC to about 197 BTC. The attacker returned 3,400 BTC, and about 600 BTC remained outstanding.
Coldcard’s trouble began on July 30. A firmware error from March 2021 weakened the randomness used to create seeds. TRM Labs estimated that about 1,816 BTC, worth roughly $116 million, was drained from more than 5,200 addresses.
D’CENT’s app wallet lost about 12.4 million XRP across 7,393 wallets between September 15 and 25. Trezor reported vendor breaches at ShipMonk and Brevo but said its wallets were not hacked. About 347,149 contacts were exported from Brevo, and roughly 2,500 people clicked a phishing link.
On October 9, Ledger began investigating drained wallets among customers who bought from the reseller CryptoBilis. On-chain investigator Specter estimated losses above $86 million, but Ledger has not confirmed a total. Two days earlier, Justin Drake urged planning for AI risks to elliptic-curve cryptography, while Coinbase’s Yehuda Lindell said there is no evidence the underlying assumptions have weakened.
| Incident | Date | Reported loss | Reported cause |
|---|---|---|---|
| AFX Trade | Jul 22 | About 24.15 million USDC | Bridge withdrawal approved with five validator signatures |
| Coldcard | From Jul 30 | About 1,816 BTC (about $116 million) | Firmware flaw weakened seed randomness |
| Tectonic | Aug 30 | $120.4 million borrowing affected | Exploit; $111.2 million reversed by rollback |
| Liquid Network | Sep 6 | About 4,000 LBTC created without backing | Software bug |
| Bitget | Sep 24 | About $387.5 million | Compromised third-party security products |
| Ledger resellers | From Oct 9 | Above $86 million (estimate, unconfirmed) | Under investigation |
How to Reduce Bridge Risk and What Comes Next
No bridge is free of risk, but users can lower it. Check that a bridge has passed security audits and has a long track record. Look for red flags before sending funds.
The sources also list basic habits. Keep funds in a secure wallet such as a hardware wallet, and use strong passwords with multi-factor authentication. Keep software updated, turn on transaction alerts and watch your activity for anything unusual.
Phishing is a common route to theft. Check the URL before entering any private key or sensitive detail. Established platforms with long track records can offer more confidence, though none is entirely safe.
Cross-chain swaps are another option. A wallet or service routes the swap for you, which may avoid direct contact with a bridge. They have limits too, such as fewer supported tokens and fees.
Bridges suit advanced needs, such as moving tokens that swap services do not support. Swaps suit basic trades between popular tokens. The table compares the two.
Newer protocols aim to cut risk. Chainlink’s CCIP has processed over $28.6 trillion in transfers without a protocol-level exploit, using decentralized validation. Wormhole added zero-knowledge proof verification and cross-chain governance in 2024.
1inch, which offers cross-chain swaps, says its product carries out a user’s instructions without taking custody of assets. That claim comes from the company itself.
These steps point to stronger validation standards, but exploits remain a risk. Access control flaws, single points of failure and weak configurations have driven many losses so far. Users should weigh each option before moving assets.
| Factor | Bridge | Cross-chain swap |
|---|---|---|
| Security | No central operator with decentralized bridges, but more complexity | User does not deal with a bridge directly, which may cut some risk |
| Fees | Vary by bridge, asset and amount | Can still apply; aggregators aim for the best rate |
| Speed | Can be slower | Can be faster, depending on congestion |
| Ease of use | Needs more understanding | Simpler for beginners |
| Best for | Unsupported tokens and DeFi on other chains | Basic swaps between popular tokens |
Sources
- What Is a Blockchain Bridge? (updated Apr 21, 2026)
- What is a Crypto Bridge? (published Jul 15, 2024, updated Jul 17, 2026)
- Introduction to Cross-Chain Bridges (Chainalysis Team, Dec 5, 2024)
- Justin Drake’s “Bunker Mode” Warning Meets Ledger and Bitget Security Fears (Trader Edge, Oct 13, 2026)
- What is bridging in crypto?
- The biggest bridge hacks in 2026 (Vladimir Kozlov, May 21, 2026)
- How Do Hackers Use Crypto Bridges to Steal?
