Aave could face up to $230 million in losses after Kelp DAO bridge exploit triggers DeFi chaos
Aave published a report outlining two possible outcomes: around $123 million in losses if damage is shared across all rsETH, or up to $230 million if confined to Layer 2s, with the final impact depending on how Kelp DAO allocates the shortfall.
Aave could face up to $230 million in losses after Kelp DAO bridge exploit triggers DeFi chaos
Aave published a report outlining two possible outcomes: around $123 million in losses if damage is shared across all rsETH, or up to $230 million if confined to Layer 2s, with the final impact depending on how Kelp DAO allocates the shortfall.
Aave’s incident report found that the rsETH exploit created unbacked collateral used to borrow roughly $190 million, leaving the protocol exposed to potential bad debt despite its systems functioning as designed.
The report outlines two possible outcomes, around $123 million in losses if damage is shared across all rsETH, or up to $230 million if confined to Layer 2s, with the final impact depending on how Kelp DAO allocates the shortfall.
An attacker exploited that setup by forging a transfer message that appeared valid. The system approved the transfer even though the tokens were never taken out of the sending chain, meaning new tokens were effectively created without backing, releasing 116,500 rsETH from the Ethereum-side bridge.
Rather than selling the assets on the open market, the attacker deposited 89,567 rsETH into Aave as collateral and borrowed roughly $190 million in ETH and related assets across Ethereum and Arbitrum, according to the report. This left Aave exposed to collateral whose backing may be significantly impaired.
Aave Labs said it moved quickly to contain the risk. Within hours, the protocol froze rsETH markets across its deployments, set loan-to-value ratios to zero, and halted new borrowing against the asset.
The outcome now depends largely on how Kelp handles the shortfall. If losses are spread across all rsETH holders, the token would face an estimated 15% depegging (meaning the value of the staked tokens would not match the value of actual ETH), resulting in about $124 million in bad debt for Aave. If losses are instead isolated to Layer 2 networks, the impact would be far more severe, with bad debt rising to roughly $230 million and concentrated on networks such as Arbitrum and Mantle.
The exploit stemmed from weaknesses in how Kelp verified cross-chain messages using LayerZero. By manipulating this process, the attacker was able to make certain assets appear fully backed when they were not, allowing them to extract value from the system. LayerZero itself was not directly hacked, but its messaging layer exposed flawed assumptions in how Kelp validated cross-chain data.
The incident raised concerns that some positions on Aave were backed by collateral that was mispriced or no longer fully backed, increasing the risk of undercollateralized loans.
The episode highlighted its indirect exposure to external systems. The impact was felt through increased collateral risk, pressure on lending positions, and a sharp decline in deposits as users reassessed the safety of interconnected DeFi infrastructure.
The report said its DAO treasury holds approximately $181 million in assets and that discussions are underway with ecosystem participants to address potential losses. Kelp has not yet outlined how it plans to allocate losses, leaving Aave’s ultimate exposure uncertain as the situation continues to evolve.
The liquid restaking protocol said the compromised verifier was LayerZero’s own infrastructure, and the setup it was faulted for running was LayerZero’s onboarding default.
What to know:
Kelp DAO is disputing LayerZero’s account of a $290 million rsETH bridge exploit, claiming that the compromised single-verifier setup relied on LayerZero’s own infrastructure and defaults rather than an outlier configuration it chose against advice.
Some security researchers say LayerZero’s public documentation and deployment code promote single-source verification across major…
