The first suspicious transaction hit the blockchain on Saturday, April 18, 2026. Forty-six minutes later, roughly $293 million in rsETH tokens had been drained from a cross-chain bridge connected to Kelp DAO, a restaking protocol built on Ethereum. But the attacker was not finished. Instead of cashing out, they deposited the stolen tokens as collateral on a separate DeFi lending platform and borrowed additional assets against them, extracting a second round of value from the ecosystem before anyone could intervene.
The breach is the largest decentralized finance exploit of 2026 so far and one of the biggest ever recorded. The Wormhole bridge hack in February 2022 cost $320 million. The Ronin Bridge attack a month later topped $625 million. According to the Rekt News exploit leaderboard, which tracks the costliest DeFi losses, the Kelp DAO theft now ranks among the top five bridge hacks in the sector’s history.
How the attack unfolded
The attacker exploited a vulnerability in the cross-chain bridge Kelp DAO used to move rsETH between networks. According to Coindoo, a crypto-focused news outlet that first reported the timeline, the entire drain took approximately 46 minutes from the initial transaction to the final withdrawal. On-chain timestamps for the relevant transactions are publicly visible on block explorers, though no independent forensic firm such as Chainalysis, Elliptic, or PeckShield has published a consolidated analysis confirming the sequence. That absence is notable: in past exploits of comparable scale, at least one of these firms typically released preliminary findings within days. As of late April 2026, none has done so publicly. If the timeline is accurate, the window was far too narrow for any human-led response. Emergency governance votes, manual contract pauses, and coordination among protocol contributors all require time that simply did not exist.
Reporting from UA.News corroborates the scale and method, describing a nine-figure theft routed through a bridge weakness. By the time broader DeFi markets noticed abnormal rsETH flows, the bridge was already empty.
What happened next compounded the damage. Multiple crypto outlets, including Coindoo and exchange-affiliated analysis from Weex, reported that the attacker deposited the drained rsETH as collateral on at least one DeFi lending protocol and borrowed against it. No named security researcher or forensic analyst has publicly attached their identity to this claim, and no on-chain transaction hashes have been published in a form that outside observers can independently verify. If the reporting is accurate, the move turned a theft into leveraged extraction. The hacker created new liabilities inside lending markets, forcing liquidation pressure onto protocols that had no direct connection to the original bridge vulnerability. For liquidity providers and lenders on those platforms, the exposure was sudden and involuntary.
Which lending protocols accepted the tainted collateral has not been publicly confirmed. None of the major DeFi lending platforms, including Aave and Morpho, have issued statements acknowledging or denying involvement as of late April 2026.
Why rsETH made the damage spread
Kelp DAO’s restaking model meant rsETH was not an isolated token sitting in a single vault. It served as a building block across DeFi: collateral in lending pools, a component in yield strategies, and a trading pair on decentralized exchanges. When a large share of the token’s supply vanished from the bridge and then reappeared as collateral under the attacker’s control, the knock-on effects hit multiple layers of the ecosystem at once.
Some liquidity pools experienced severe imbalances as rsETH’s price diverged sharply from its expected peg. Lending protocols that accepted rsETH as collateral faced shortfalls. Traders holding positions denominated in rsETH watched their value erode in real time, with no clear mechanism to halt the slide.
The pattern echoes the Mango Markets exploit in October 2022, where an attacker manipulated token prices to borrow against inflated collateral, draining roughly $114 million. The Kelp DAO breach operated at nearly three times that scale, but the underlying mechanic, using one protocol’s weakness to create cascading liabilities across others, is the same.
What is still unknown
Key details remain unconfirmed weeks after the attack.
The exact dollar figure varies slightly by source. Most outlets report $293 million, while Weex’s analysis places it at $292 million. The gap likely reflects token price fluctuations during the attack itself. No independent on-chain forensic firm has published a consolidated figure. The silence from firms like Chainalysis, Elliptic, and PeckShield, which routinely publish rapid assessments after major exploits, is itself a gap worth noting. It means the figures circulating in media coverage have not been independently validated by any entity with direct blockchain analytics capabilities.
The timing carries a minor discrepancy as well. Some reports place the attack on April 18, while others describe Kelp DAO acknowledging the breach on April 19. Both could be accurate if the hack occurred late Saturday and the protocol’s public response came the following day.
No official statement from Kelp DAO leadership with verbatim language has appeared in available reporting. It is unclear whether the affected smart contracts have been permanently disabled, upgraded, or simply paused. There is no public information about insurance coverage, backstop funds, or any compensation plan for users who held rsETH on the bridge or in connected lending and liquidity positions. It is also unknown whether Kelp DAO has a bug bounty program that could serve as a basis for white-hat negotiation with the attacker, a tactic that has recovered partial funds in past exploits such as the Euler Finance hack in March 2023.
The attacker’s identity remains unknown. No law enforcement agency has issued a public statement, and no blockchain forensics firm has published attribution linking the exploit to a known threat group. Whether the funds can be traced, frozen, or recovered depends on factors that have not been disclosed, including which lending protocols accepted the stolen collateral and whether those platforms can or will freeze the relevant accounts. If the attacker routed funds through privacy tools or layered cross-chain paths, practical recovery odds drop significantly, even if investigators can follow portions of the trail.
No cryptocurrency exchange has publicly confirmed flagging addresses associated with the exploit, and no regulatory body in any jurisdiction has commented on the breach.
Why bridges keep breaking
Cross-chain bridges have been the single most exploited category of DeFi infrastructure since 2022. The Rekt News leaderboard shows that bridge hacks account for four of the ten largest DeFi losses ever recorded, including Ronin ($625 million), Wormhole ($320 million), Nomad ($190 million), and now Kelp DAO. The fundamental problem has not changed: bridges hold large pools of locked assets and rely on complex verification logic to process cross-chain messages. A single flaw in that logic can expose the entire pool.
The Kelp DAO exploit adds a dangerous new dimension. By using stolen tokens as collateral on a lending platform, the attacker demonstrated that bridge vulnerabilities can propagate damage far beyond the bridge itself. If lending protocols do not have mechanisms to detect and reject tainted collateral in near-real time, they become unwitting amplifiers of the original loss. That is a design gap the broader DeFi ecosystem has discussed for years but has not solved.
For users, the practical lesson is blunt: assets parked on cross-chain bridges, or in pools that depend on bridged tokens, carry a layer of risk that no yield can fully compensate. Until bridge architecture undergoes a fundamental redesign, or until automated circuit breakers become standard across lending protocols, exploits of this kind will keep happening. As of May 2026, no protocol, regulator, or forensic firm has publicly outlined a concrete plan to prevent the next one.
