Evercrest Technologies, parent of KelpDAO, has sued LayerZero Labs and a co-founder over the April 18 exploit that drained 116,500 rsETH, roughly $292 million, alleging the bridge's single-verifier configuration had been reviewed and endorsed in writing. It is the first time a DeFi protocol has taken its own infrastructure provider to court, and it puts the sector's oldest slogan on trial.
The $292M DeFi Bridge Liability Suit Killed "Code Is Law"
Key Takeaways
- KelpDAO's parent company, Evercrest Technologies, sued LayerZero Labs and co-founder Bryan Pellegrino over the April 18 exploit that drained 116,500 rsETH, roughly $292 million, the largest DeFi exploit of 2026.
- The claim turns on a 1-of-1 verifier setup that LayerZero allegedly "reviewed and endorsed in writing," then publicly attributed to Kelp's own configuration.
- It is the first time a DeFi protocol has taken its own infrastructure provider to court, a direct challenge to the "code is law" principle.
- The attacker is linked to North Korea's Lazarus Group, but the legal dispute is between two solvent crypto companies over who owned the weak link.
One forged message, $292 million, one block
On April 18, 2026, a single forged message convinced a cross-chain bridge to release 116,500 rsETH, about $292 million, in one block. No private key was stolen and no smart contract malfunctioned. The code did exactly what it was told; it was just told a lie. Kelp's bridge balance fell from 116,723 rsETH to 223 rsETH in that block, meaning roughly 18 percent of the circulating rsETH supply left the bridge.
On September 24, Evercrest Technologies, the parent company of KelpDAO, filed a civil claim against LayerZero Labs and co-founder Bryan Pellegrino in the Supreme Court of British Columbia. In a single filing, DeFi's oldest slogan received a court date.
What does "code is law" mean when the code is wrong?
"Code is law" has been crypto's founding premise for a decade: no counterparties, no courts, no human error, only math. A cross-chain bridge is meant to be the purest expression of that idea. Lock a token on one chain, mint an equivalent on another, and let verifiers rather than people enforce the rules. "Trustless" was always a half-truth, because somebody has to operate the verifier that checks whether tokens were actually burned before an equivalent amount is released elsewhere.
In Kelp's case, that somebody was LayerZero, specifically a 1-of-1 decentralized verifier network, or DVN. In practical terms, one verifier stood between an incoming message and the release of $292 million. There was no second opinion and no cross-check. That single point of trust is the entire story: the attack did not break the bridge, it poisoned the one sensor the bridge relied on.
Who wins and who loses when a bridge fails?
The winners are easy to name. Courts and outside advisers, first. Chainlink's CCIP, the alternative Kelp migrated to in May, gained visibility that advertising budgets rarely buy: Solv Protocol moved roughly $700 million of SolvBTC off LayerZero, and Lombard, with $2.1 billion in total value locked, followed. Insurance and audit firms pick up new business after each headline loss. Every protocol already running multi-verifier redundancy became more attractive by comparison.
The losers are equally clear. LayerZero's reputation absorbed the direct hit, and Aave carries an estimated $123.7 million to $230.1 million in bad debt after the attacker deposited counterfeit rsETH and borrowed real assets against it. The quietest casualty is the hardline no-recourse position, the argument that in DeFi losses are final and nobody owes anyone. A crypto company has now shown that even the industry's own builders do not accept that framing.
The second-order effect may matter more than any of this. The phrase "reviewed and endorsed in writing" is now a legal instrument. Documentation, internal chat logs and direct messages from infrastructure providers are potential discovery evidence. A wave of rewritten terms of service across the cross-chain industry within twelve months would be an unremarkable outcome.
Has this happened before?
Not in this form. In 2022 Wormhole lost $325 million and Jump Crypto made users whole within days, with no courtroom and no blame game. In 2023 Multichain lost roughly $1.5 billion, its chief executive was arrested in China, and the funds remained frozen with no counterparty left to sue. In 2022 Ronin lost $625 million and was resolved through coordinated backstops. DeFi's playbook was binary: a quiet bailout or a loud collapse. There was no third option.
Kelp's claim introduces one. Two solvent crypto companies are asking a court to decide who owned the weak link: the protocol that deployed the bridge, or the infrastructure provider whose single verifier signed off on a forged message. LayerZero initially said its protocol "functioned exactly as intended" and that users should not rely on a sole DVN, then later acknowledged it "made a mistake by allowing our DVN to act as a 1-of-1 DVN for high-value transactions." Kelp states that LayerZero warned a separate developer, USDT0, about verifier risks in late 2024 or early 2025 but did not pass the same warning to Kelp. That asymmetry is what litigation exists to examine.
Where does bridge liability head in one to three years?
Expect "accountability infrastructure" to become a product category: legal wrappers, DAO legal entities, on-chain insurance, and explicit liability clauses written into integration agreements. "Who verifies your bridge?" joins "who audits your code?" on institutional due-diligence checklists. The assumption that a bridge operator is released from responsibility the moment a transaction clears is finished.
A second force runs in parallel. The same week Kelp filed, New York sued Polymarket over gambling, a reminder that state-versus-industry disputes are accelerating on their own track. Regulators will read crypto-versus-crypto litigation as a roadmap: where the industry cannot agree on responsibility, courts and agencies fill the vacuum. The era of settling these matters privately is ending because the losses have grown too large and the counterparties are now substantial enough to contest them.
What market participants should take from this
This is not a price story, and it should not be read as one. It changes how risk should be assessed. First, single points of failure deserve reassessment: capital keeps migrating toward multi-verifier, multi-provider setups, and protocols still running a lone DVN carry a discount the market has not fully priced. Second, written endorsements now behave like contingent liabilities; if an infrastructure provider endorsed a configuration, that endorsement is a potential claim against it. Third, the attribution in this case, social engineering of a LayerZero developer followed by poisoning of its RPC infrastructure by a Lazarus-linked cluster known as TraderTraitor, describes a recurring, well-funded attack class rather than a one-off. Sanctions-aware, compliant DeFi is becoming a genuine differentiator rather than a marketing term.
The deeper lesson is that the exploit was not a key compromise. It was infrastructure poisoning, an attack class that a smart contract audit does not surface. A security thesis that begins and ends with audited contracts is examining the wrong layer.
The question nobody wants to answer
The most telling detail is not the $292 million. It is that two crypto teams that both preach trust in code are now asking a court to decide who trusted whom. When DeFi's own builders refuse to accept "code is law" as an answer, the rest of the market should not either. The real question was never whether crypto would develop accountability, but who writes it: a courtroom, or the DAO that was supposed to replace one.
FAQ
What happened in the KelpDAO exploit? On April 18, 2026, attackers fed a forged message to the single verifier securing Kelp's LayerZero bridge, causing it to release 116,500 rsETH, about $292 million, without a corresponding burn on the source chain. Kelp's bridge balance dropped from 116,723 rsETH to 223 rsETH in one block.
Why is KelpDAO suing LayerZero? Kelp alleges LayerZero "reviewed and endorsed in writing" the vulnerable 1-of-1 verifier setup, then publicly blamed Kelp's configuration after the loss, and that LayerZero warned another developer about verifier risks but never warned Kelp. The claim pleads negligent misrepresentation, negligence and defamation.
Does "code is law" mean nobody is liable? In practice, no. It has generally meant that when code fails, the user absorbs the loss. This lawsuit shows that even the industry's own builders now expect recourse: a court to decide who owned the weak link when the code did what it was told but was told a lie.
Who is believed to have taken the funds? LayerZero's incident report, citing Mandiant and CrowdStrike, attributed the intrusion to TraderTraitor, also tracked as UNC4899, a cluster linked to North Korea's Lazarus Group. The funds trail leads to a state actor, while the courtroom dispute remains between Kelp and LayerZero.
What does this mean for other bridges? Expect capital to keep migrating toward multi-verifier, multi-provider setups such as Chainlink CCIP, and expect infrastructure providers to revise terms of service and disclaimers. "Who verifies your bridge?" is now a due-diligence question rather than an afterthought.