Cross-chain bridges have lost $340.7 million across 14 exploits in 2026, according to PeckShield data through June 1. This figure sits within a broader $840 million-plus total DeFi loss tally for the year. Bridge protocols — the infrastructure connecting disparate blockchains — remain the single ...
"We made a mistake." — Bryan Pellegrino, CEO, LayerZero Labs (May 2026, on the KelpDAO incident report)
Cross-chain bridges have lost $340.7 million across 14 exploits in 2026, according to PeckShield data through June 1. This figure sits within a broader $840 million-plus total DeFi loss tally for the year. Bridge protocols — the infrastructure connecting disparate blockchains — remain the single most exploited category in decentralized finance, a pattern now in its fifth consecutive year.
The attack surface has shifted. Code-level smart contract bugs, once the dominant vector, now account for a minority of bridge losses. In 2026, compromised accounts and social engineering represent more than 50% of all DeFi attacks by incident count. Two exploits attributed to DPRK-linked threat actor TraderTraitor — the $285 million Drift Protocol breach and the $292 million KelpDAO drain — together account for 76% of all crypto hack value through April 2026, per TRM Labs. Both attacks targeted human infrastructure, not contract logic.
This report examines the 14 bridge exploits of 2026, the structural vulnerabilities they exposed, the emerging security responses, and what cumulative losses exceeding $3 billion since 2021 mean for bridge architecture going forward.
PeckShield's June 1, 2026 alert documented 14 cross-chain bridge exploits totaling $340.7 million in losses year-to-date. May 2026 alone saw eight major incidents draining a combined $328.6 million, making it the most active month for bridge attacks on record.
The five-year cumulative toll is stark. Since 2021, cross-chain bridge hacks have caused more than $3 billion in losses. In 2022, five bridge attacks accounted for $1.32 billion — 57% of all Web3 losses that year. The 2026 figures represent a continuation rather than an aberration.
Major 2026 Bridge Exploits:
| Date | Protocol | Loss | Vector | |------|----------|------|--------| | Feb 21 | IoTeX ioTube Bridge | $8.8M | Private key compromise | | Apr 1 | Drift Protocol | $285M | Social engineering / multisig takeover | | Apr 18 | KelpDAO rsETH Bridge | $292M | RPC poisoning / single-DVN bypass | | May 11 | THORChain | $10M | GG20 threshold-signature flaw | | May 18 | Verus-Ethereum Bridge | $11.5M | Forged Merkle proof |
The remainder of the 14 incidents involved smaller protocols, but the pattern is consistent: bridge infrastructure presents a concentrated, high-value target.
KelpDAO ($292 million, April 18)
The largest bridge-specific exploit of 2026 targeted KelpDAO's rsETH token bridge, built on LayerZero's Omnichain Fungible Token (OFT) standard. Attackers drained 116,500 rsETH across 20 chains.
The root cause was a single-verifier configuration. KelpDAO operated a 1-of-1 Decentralized Verifier Network (DVN) setup — meaning one compromised verification node could approve fraudulent cross-chain messages unilaterally. The breach began on March 6, when an attacker socially engineered a LayerZero Labs developer to harvest session keys, then pivoted into LayerZero's RPC cloud environment and poisoned internal RPC nodes. External nodes were DDoS'd to force reliance on compromised infrastructure.
Mandiant, CrowdStrike, and independent researchers attributed the attack to DPRK threat actor TraderTraitor (UNC4899). At the time of the exploit, 47% of approximately 2,665 LayerZero applications were running the same single-verifier configuration, according to LayerZero's own incident report.
KelpDAO subsequently migrated its rsETH infrastructure from LayerZero to Chainlink's Cross-Chain Interoperability Protocol (CCIP).
Drift Protocol ($285 million, April 1)
Drift, the largest decentralized perpetual futures exchange on Solana, lost $285 million in user assets. The attack was not a bridge exploit in the traditional sense but exploited the same class of vulnerability: privileged access to cross-chain-capable infrastructure.
Per TRM Labs and Chainalysis, DPRK operatives spent six months conducting a social engineering campaign targeting Drift's multisig signers. The attackers tricked signers into pre-signing hidden authorizations, then executed a zero-timelock Security Council migration that eliminated the protocol's governance safeguards. The entire vault system was drained in a single transaction.
This was the second-largest exploit in Solana history, behind only the $326 million Wormhole bridge hack of February 2022.
THORChain ($10 million, May 11)
A malicious node operator exploited a vulnerability in THORChain's GG20 threshold-signature scheme. GG20 distributes private key control across multiple node operators to prevent single points of failure, but the attacker reconstructed the full private key through the flaw.
The exploit hit four blockchains — Bitcoin, Ethereum, BNB Chain, and Base — affecting 12,847 wallets. THORChain's automated security system halted further outflows within eight minutes. The protocol opened a recovery portal and proposed covering losses from protocol liquidity without issuing new tokens.
The 2026 data confirms a structural shift in DeFi attack methodology. According to multiple security firms, compromised accounts now account for more than 50% of all DeFi attacks by incident count, overtaking smart contract exploits as the primary loss category.
The pattern is consistent across the year's major incidents:
Impersonation scams surged 1,400% year-over-year in 2026, according to industry tracking data, making social engineering the fastest-growing crypto threat vector.
The logic is straightforward. As smart contract auditing has matured — with firms like Trail of Bits, OpenZeppelin, and Halborn making code-level bugs harder to find — attackers have moved upstream to the human layer. The cost-benefit calculation favors social engineering: one compromised developer session key can bypass months of audit work.
North Korea-linked actors dominate the loss statistics. Per TRM Labs, DPRK-attributed thefts accounted for 76% of all crypto hack value through April 2026, achieved through just two major operations: Drift ($285M) and KelpDAO ($292M).
The cumulative scale is significant. DPRK-linked actors stole $2.02 billion in 2025 (a 51% year-on-year increase), pushing their all-time total past $6.75 billion since 2017, according to TRM Labs. The February 2025 Bybit theft alone — $1.5 billion in Ethereum, attributed by the FBI to TraderTraitor — was the largest single crypto heist in history.
In Q1 2026, $309 million was stolen across 12 DPRK-attributed incidents. The operational playbook is well-documented: fake recruiter pitches, malware-laced pre-employment tests, compromise of wallet software vendors, and infiltration of signing infrastructure. TraderTraitor operatives have been documented maintaining cover identities within DeFi teams for periods exceeding six months before executing attacks.
Cross-chain bridges are a preferred target because they aggregate large pools of locked assets secured by a relatively small set of privileged keys or verification nodes. A single point of compromise can yield nine-figure payouts.
The KelpDAO exploit triggered measurable changes in bridge security architecture.
LayerZero's Policy Shift
LayerZero Labs announced that its DVN "will not sign or attest messages from any applications that utilize a 1/1 configuration." The company began outreach to projects still running single-verifier setups, effectively conceding that configuration flexibility without enforced safety rails was too permissive. This was a direct reversal: LayerZero had previously positioned configurability as a feature, not a risk.
Migration to Chainlink CCIP
KelpDAO migrated its rsETH infrastructure to Chainlink's CCIP. Kraken also adopted CCIP as its cross-chain standard, replacing LayerZero, in May 2026. CCIP's architecture features 16 independent, audited node operators providing redundant validation, plus a separate risk management network for rate limiting and anomaly detection. Chainlink reports CCIP has secured over $28 trillion in cumulative value with $90 million-plus in weekly volume.
Intent-Based Architecture
The broader market is shifting toward intent-based bridges (Across, deBridge) that operate with near-zero TVL because operators front funds without custody. This model reduces the attack surface by eliminating large pools of locked assets. The trade-off is reduced capital efficiency and reliance on solver networks.
Wormhole Settlement
Wormhole launched its Settlement product in 2026 as an institutional-grade, intent-based cross-chain protocol using auction mechanisms to reduce MEV extraction and improve capital efficiency.
The economic math is unfavorable. Cross-chain bridges hold approximately $388 million in total value locked as of June 2026, per DefiLlama. Against $340.7 million in exploit losses this year alone, the loss-to-TVL ratio approaches 88% — meaning the sector has lost in 2026 nearly as much as it currently holds.
This ratio understates the true economic cost. Protocol teams absorb recovery expenses, user compensation commitments, and reputational damage. IoTeX pledged 100% compensation from its treasury. THORChain proposed covering losses from protocol liquidity. Drift outlined a recovery plan for affected users. These commitments further reduce protocol treasuries, compounding the financial impact.
The fee revenue bridges generate does not justify the risk exposure. Bridge protocols collectively generate modest revenue relative to the capital they secure — and certainly relative to the capital they have lost. This confirms the broader pattern identified in foundational economic analyses of the blockchain sector: much of the ecosystem's infrastructure operates at a persistent economic deficit, sustained by subsidies and external capital rather than self-sustaining fee revenue.
Insurance coverage for bridge risk remains underdeveloped. Few protocols carry meaningful exploit insurance, and the DeFi insurance sector (Nexus Mutual, InsurAce) lacks the capitalization to cover nine-figure losses.
Cross-chain bridges remain structurally vulnerable, and the nature of that vulnerability has changed. The 2026 data shows that the problem is no longer primarily one of code quality — it is one of operational security, configuration management, and human-factor resilience.
The industry's response has been reactive. LayerZero's ban on single-verifier configurations came after $292 million in losses, not before. The migration to multi-verifier and intent-based architectures addresses the specific failure modes of 2026 but does not resolve the fundamental tension: bridges aggregate value in ways that create high-yield targets for sophisticated actors, particularly state-sponsored groups.
The economic sustainability of bridge infrastructure is questionable. Protocols that lose nearly as much to exploits as they hold in locked value are not operating viable businesses. The shift toward intent-based models — which reduce locked value to near zero — represents a pragmatic architectural response, but comes with its own trade-offs in capital efficiency and solver reliability.
Until the industry develops bridge architectures that are economically sustainable and resistant to the social engineering campaigns that now dominate the threat landscape, cross-chain interoperability will remain DeFi's most expensive unsolved problem.