Kelp DAO absorbed $400 million in bad debt, but it cost Aave $12 billion

Original author: Sanqing, Foresight News

On May 26, Kelp DAO transferred the final batch of 20,373.72 rsETH to the LayerZero OFT Adapter, and Aave simultaneously announced that rsETH and all affected markets had returned to normal. Over 37 days, the full replenishment of 116,500 rsETH was completed.

However, this only restored the 1:1 backing of rsETH; it does not mean Aave's books are completely cleared. The 30,766 ETH frozen by the Arbitrum Security Council remain stuck in the U.S. District Court for the Southern District of New York, with ownership still undetermined. The TVL lost by Aave also cannot simply return along with the restored rsETH.

The Bill Extends Beyond the TVL Column

According to DefiLlama data, on the day of the incident, April 18, Aave's TVL stood at $26.396 billion; by May 25, it had fallen to $14.181 billion. The amount of capital that has not returned after a month exceeds $12 billion.

What lies ahead is even more challenging. On June 5, the U.S. District Court for the Southern District of New York will hold a hearing on the ownership of the 30,766 ETH frozen by the Arbitrum Security Council. Both Aave LLC and Gerstein Harrow submitted supplementary briefs before May 22. While the judge modified the restraining notice on May 8 to allow fund transfers, the substantive ruling is still pending the June 5 hearing.

Gerstein Harrow represents the families of victims of North Korean terrorism, holding an unenforced judgment of $877 million. Regardless of the outcome of this lawsuit, it constitutes a drain on the Aave brand.

DeFi United was able to assemble this time because multiple parties were willing to backstop the situation: Stani Kulechov contributed 5,000 ETH personally, Consensys and Joseph Lubin committed up to 30,000 ETH, the Aave treasury allocated up to 25,000 ETH, alongside a credit line of up to 30,000 ETH from Mantle and support from Lido, Ether.fi, and others.

The scale of community mobilization was unprecedented, but Aave has used up this one-time-only recourse. Should another upstream contamination event occur, it may not be possible to assemble the same list of backers again.

For instance, after the incident, Justin Sun withdrew approximately $174 million (including 65,854 ETH and various stablecoins) from Aave to Spark, accumulating over $1.3 billion in deposits on Spark. Whales are voting with their feet, and capital has already migrated.

V4's Openness is Being Slowed by Governance

Aave's counterpunch cards are not limited to V4, but V4 is the most crucial one.

V4 went live on Ethereum mainnet on March 30, featuring a Hub-and-Spoke architecture with three initial Liquidity Hubs. Aave Labs committed to "security-first growth", gradually increasing deposit caps. It surpassed $10 million in deposits on April 8, crossed $50 million on May 9, and by May 26, total deposits stood at $86.13 million, with active borrowing positions at $27.77 million.

This pacing was a responsible design choice before the rsETH incident, but it turned into a stress test afterward. Aave is simultaneously managing a $200 million bad debt on V3 while slowly expanding the caps on V4.

More problematic is that V4 also faces internal friction within its own governance layer. In February 2026, Aave Labs submitted a strategic proposal bundling product revenue, service provider incentives, V4 growth engines, and brand legal custody, requiring delegates to vote on four different risk dimensions in a single package.

Marc Zeller, founder of the Aave Chan Initiative, publicly questioned the appropriateness of bundling a massive fund request with strategic approvals. This governance dispute has been simmering around the V4 launch, and each delay allows competitors to eat away at Aave's market share.

V4's strength lies in the openness of its Spoke design; anyone can build a Spoke, and meeting conditions grants access to a Liquidity Hub as a credit line. This is why Babylon Labs chose to integrate its Trustless Bitcoin Vaults into V4 rather than elsewhere. However, the speed at which this openness materializes depends on whether the governance layer can keep pace.

More Than Just V4: Aave Fights Three Battles

Aave V3 remains the cash cow, generating over $100 million in annualized revenue, with the bulk of its $14.1 billion TVL residing on V3. The "Aave will win" proposal positions V3 in a "stable maintenance" phase, and Stani has publicly committed to no forced migration and no deadlines.

V4 and V3 will run in parallel for at least 24 to 36 months. V4 acts as an additive layer, capturing heterogeneous use cases that V3 cannot accommodate. Horizon, on the other hand, is a permissioned fork of V3, designed specifically to service institutional RWA.

These three layers each target different growth increments. V4 pursues new scenarios beyond V3's risk framework, and after rsETH, it has an additional task: providing a reason for capital that migrated to Morpho and Spark to return to Aave. Horizon targets RWA traffic from traditional finance, operating in a completely separate liquidity pool from V3 and V4.

The Horizon Market officially launched in August 2025. It is a permissioned V3 instance deployed by Aave, specifically designed for institutions to use tokenized Treasuries, corporate bonds, and money market funds as collateral to borrow stablecoins like USDC, GHO, and RLUSD.

As of May 26, it has accumulated over $500 million in net deposits, aiming to break $1 billion by the end of 2026. Partners include BlackRock, Franklin Templeton, Circle, Ripple, and VanEck.

This path diverges from Morpho's vault management model. Morpho, through third-party curators like Steakhouse and Gauntlet, manages vaults to capture lending flow from retail institutions like Coinbase. Aave, via Horizon, directly connects with asset managers from traditional finance to capture RWA.

These two paths serve different institutional client profiles. Morpho caters to fintech companies that treat on-chain lending as a tool. Aave serves asset managers who view the blockchain as an issuance venue.

The capital flight following the rsETH incident primarily affected the first type of client. The second type faces higher migration costs and reacts more slowly. The compliance framework, KYC processes, and asset onboarding audits Aave has accumulated on Horizon are difficult for Morpho to replicate in the short term after this event.

This is the only incremental growth line for Aave not directly impacted by the rsETH incident, but its growth depends on the pace at which traditional finance integrates with DeFi.

There Will Be No Second DeFi United

Aave remains the largest protocol in the lending market. Its TVL of $14.1 billion is still nearly double that of Morpho, and the deployment depth accumulated over years is something no one can catch up with in the short term.

However, the bill from rsETH isn't on the balance sheet; it's in the column of institutional default preference for lending protocols. Spark's TVL grew from $3.727 billion to $5.3 billion in one month. Morpho's TVL, after hitting bottom on April 21, is slowly recovering to pre-incident levels. These numbers won't automatically reverse course just because Aave's markets have recovered.

The speed of V4's delivery in heterogeneous scenarios, combined with the progress of Horizon in institutional RWA, will determine whether Aave can win back its lost market share. But the former is stuck on governance infighting, while the latter depends on traditional finance's own pace of integration. And Aave can only wait on both fronts.

DeFi United is not a permanent institution; it was a one-time assembly.