Kelp DAO Vulnerability Triggers Exodus of Hundreds of Billions; Two Major DeFi Lending Pathologies Clash Head-On

marsbit發佈於 2026-05-29更新於 2026-05-29

文章摘要

Title: Kelp DAO Exploit Triggers $15 Billion Exodus, Exposing a Clash Between Two DeFi Lending Models. In April 2026, a hacker exploited a LayerZero bridge vulnerability in the Kelp DAO project, minting $292 million in fake rsETH tokens. These were deposited into Aave as collateral to borrow real Ethereum, draining the protocol's liquidity. Within three and a half days, Aave saw $15 billion in deposits flee, forcing a costly $160 million bailout. The root cause was identified as Aave's governance, which had previously voted to set rsETH's loan-to-value ratio to a risky 93%, leaving minimal safety margin. This incident starkly contrasts with the experience of Morpho, the second-largest DeFi lending protocol. Some fake rsETH also flowed into Morpho, but the exposure was limited to $1 million across isolated, pre-configured markets, preventing systemic contagion. The event highlights a fundamental divergence in DeFi lending architectures. Aave employs a shared liquidity pool model, where all deposits back all approved collateral assets, governed by DAO vote. This creates systemic risk, as seen when even users who never interacted with rsETH faced frozen funds. Furthermore, Aave's governance, influenced by leveraged borrowers, prioritized their interests during the crisis, even lowering borrowing rates for frozen markets at the expense of safer depositors. Its supplemental insurance mechanism, Umbrella, also failed as providers withdrew capital when needed. Morpho operates on...

Author: Vaidik Mandloi

Compilation: Saoirse, Foresight News

The underlying principles of all lending protocols in DeFi are largely similar: users deposit stablecoins or Ethereum into a shared liquidity pool, and borrowers draw funds from it after pledging assets; decentralized autonomous organizations (DAOs) vote to decide which assets can serve as collateral and their corresponding loan-to-value (LTV) ratios. Aave has developed a deposit scale of $500 billion precisely by relying on this model. For most of DeFi's development, this has been the industry's sole operating model, and its rationality has never been truly questioned.

However, on April 18, 2026, a hacker exploited a vulnerability in the LayerZero cross-chain bridge of the Kelp DAO project to forge rsETH tokens worth $292 million. The hacker deposited these counterfeit tokens into Aave as collateral to borrow real Ethereum. Within hours, the utilization rates of Aave's major mainstream lending markets reached 100%, meaning all available funds within the protocol had been fully borrowed. Over the next three and a half days, the platform lost $15 billion in deposits. Ultimately, Aave had to collaborate with various ecosystem parties to conduct a rescue, raising $160 million to cover the losses.

Although this vulnerability originated from the Kelp DAO project, the root cause of such massive losses lies in Aave's governance mechanism. As early as January of this year, a community vote decided to raise the collateral factor for rsETH to 93%, leaving only a 7% safety margin for such assets. It was this single decision that brewed one of the largest bank runs in the history of DeFi lending.

On the same day, some of the forged rsETH tokens also flowed into Morpho, the second-largest DeFi lending protocol. However, the risk exposure was only $1 million and dispersed across two independent, small isolated markets, failing to trigger a chain-reaction crisis.

Upon conducting in-depth research into this incident, I discovered that behind this event lies far more than a simple security attack.

Core Differences Between the Two Models

To understand why Aave hemorrhaged billions while Morpho remained largely unscathed, we must first clarify the fund placement and operational logic of the two protocol types.

When you deposit USDC into Aave, the funds flow into a single master liquidity pool, supporting lending activities for all community-approved assets like Ethereum and staked tokens. Depositors cannot choose the type of collateral asset their funds correspond to; all related rules are set by DAO votes. Therefore, when rsETH faced collapse risk, even ordinary users who had only deposited USDC and never touched rsETH found their assets frozen—everyone's funds were in the same risk pool, suffering collective losses.

Source: BingX

More critically, while the market was halted and users couldn't withdraw, Aave's governance layer actually lowered the borrowing rates for the frozen Ethereum markets, aiming to protect borrowers who had leveraged rsETH. Since deposit rates are directly linked to borrowing rates, depositors with the lowest risk and principal security saw their deposit yields shrink further.

In traditional credit systems, lenders with the lowest risk enjoy priority in repayment. However, Aave completely inverted this rule. The reason is that borrowers engaged in rsETH leveraged trading are also the most active voting group in community governance. When risk erupts, high-risk participants holding governance power naturally prioritize protecting their own interests.

Aave launched an insurance mechanism called Umbrella in late 2025, attempting to address such bad debt risks. Users could stake Ethereum; if the protocol incurred bad debts, the staked assets would be used for compensation. However, after the Kelp DAO crisis erupted, 18,922 out of 23,507 staked aWETH positions entered an unstaking waiting period, with nearly 80% of the insurance pool's funds withdrawing collectively.

This mechanism ultimately failed completely. On-chain insurance relies on voluntary user participation, and capital providers inevitably choose to exit when real risk materializes—after all, their assets only face substantial loss when a crisis occurs. This leads to such insurance often existing during peaceful times but becoming ineffective precisely when protection is needed.

Morpho's operational model is entirely different. It abandons the unified shared liquidity pool. Anyone can create an independent, isolated lending market, pre-setting the loan asset, collateral asset, price oracle, and interest rate model. Once parameters are deployed, they cannot be modified. To adjust risk levels, one can only create a new market.

Differences in underlying architecture between the traditional DeFi lending model (represented by Aave) and Morpho's "Morphological" model.

Furthermore, Morpho introduces independent risk management institutions (Stewards), such as Gauntlet and Steakhouse Financial. These entities establish vaults, allocate funds to different markets based on their own analysis, and charge performance fees; if losses occur, they are confined within their own vaults. Gauntlet also provided risk advice for Aave, but in Aave's system, its professional opinions were often overruled by token holders seeking high yields through voting, a situation Morpho prevents at its root.

The Overlooked Hidden Cost

Aave and Morpho are currently the two most widely applied lending models in the crypto space: Aave uses the shared liquidity pool model where all deposits are aggregated, with risk rules set by community votes; Morpho advocates the isolated market model, where each lending pair is independent, with risks managed autonomously by professional institutions.

The Kelp DAO vulnerability exposed the flaws and weaknesses of the shared pool model. But even during stable periods without security incidents, this model harbors a long-overlooked hidden cost. Aave's three core markets on Ethereum (Ethereum, USDT, USDC) contribute 89% of the platform's lending volume. In these three markets, deposit rates are consistently 25% to 35% lower than borrowing rates. This spread essentially represents idle funds lying dormant in the liquidity pool; depositors cannot profit from them, yet borrowers still bear the full borrowing cost.

The interest rate mechanism adjusted based on utilization rates can push rates higher when risk increases but cannot activate idle funds when lending demand is low, leaving large amounts of assets stranded in the pool generating no yield. In these three markets alone, the annual value erosion due to idle funds amounts to approximately $52 million, close to a quarter of Aave's annualized revenue for one quarter. Even zeroing out the reserve ratio and canceling platform fees cannot solve the idle fund issue—it's an inherent shortcoming of the shared pool architecture.

Morpho's interest rate model aims to maintain a utilization rate of around 90%, significantly higher than Aave's 60% to 80% range. This model can sustain high utilization because deposits within the platform are not re-used as collateral for other loans, avoiding chain-liquidation risks at the source and thus eliminating the need to reserve large amounts of capital as a risk buffer. When lending demand is strong and funds are heavily borrowed, rates automatically increase, attracting more depositors; when lending demand is weak, rates decrease, stimulating borrowing. The entire system achieves dynamic balance without requiring community votes.

Source: Gate.com

Actual data confirms its advantage: even after deducting Steward fees, the yield offered to depositors by Morpho's top USDC vaults still exceeds that of Aave and Compound. Currently, Morpho's deposit-to-loan ratio is 41%, while Aave's is 39%, and the former's scale reaches tens of billions of dollars, meaning the yield advantage benefits all depositors on the platform day after day.

Institutional Choice: Which is More Trustworthy?

Surprisingly, all of Coinbase's crypto asset lending services are built on Morpho. The related loan scale has now surpassed $2 billion, and over 100 million platform users are indirectly enjoying the returns provided by Morpho.

Most users aren't even aware they are using DeFi services. Coinbase did not develop its own lending system nor choose another platform. The core reason is that Morpho's underlying architecture allows the platform to independently set risk parameters, select partner risk institutions, and maintain full control over the entire product experience.

Apollo Global Management, a global asset manager with over $1 trillion in assets under management and 30 years of experience in private credit, recently signed a four-year cooperation agreement, planning to acquire up to 90 million MORPHO tokens, accounting for 9% of the total token supply. The institution is connecting its tokenized fund assets to Morpho as collateral, with Gauntlet responsible for vault management and market stress testing.

Beyond that, Anchorage Digital, the first federally chartered native crypto bank in the US, has connected its institutional clients managing hundreds of billions to Morpho vaults; SG-FORGE, the compliant arm of French banking giant Société Générale, is the first licensed bank to implement DeFi lending business through Morpho.

These heavily regulated traditional financial institutions collectively chose Morpho, with a highly consistent core demand: the isolated market model allows them to meet their own compliance and risk control requirements without relying on DAO decisions. In contrast, all market rules in Aave inevitably involve community voting, completely incompatible with institutions' need for autonomous control.

Changes in the regulatory environment have further amplified this trend. The US "GENIUS Act" stipulates that stablecoin issuers cannot directly distribute investment returns, meaning stablecoin institutions require neutral underlying infrastructure to activate vast amounts of idle assets. US-related projections show that by 2028, the scale of stablecoin reserves invested in US Treasury bonds will surge from the current $120 billion to over $1 trillion. This massive pool of capital urgently needs a lending foundation that allows asset custodians to control their own risks, and Morpho is currently the most fitting choice.

熱門幣種推薦

相關問答

QWhat was the root cause of the massive $15 billion deposit outflow from Aave following the Kelp DAO exploit?

AThe root cause was Aave's governance mechanism. Earlier, the community had voted to increase the loan-to-value (LTV) ratio for rsETH to 93%, leaving a very thin safety margin of 7%. When fake rsETH tokens were deposited as collateral, it quickly drained the shared liquidity pool, causing the massive withdrawal.

QWhat is the fundamental difference in the operational models between Aave and Morpho as described in the article?

AAave uses a shared liquidity pool model where all deposits are pooled together to back all approved assets, with rules set by DAO governance. Morpho uses an isolated market model, where each lending market is separate with its own parameters set at creation, managed independently by professional risk stewards.

QWhy did Aave's Umbrella insurance mechanism fail during the crisis, according to the article?

AThe mechanism failed because it relies on voluntary participation. When the crisis hit, nearly 80% of the staked insurance funds (aWETH) entered the unstaking period and were withdrawn. Participants were incentivized to leave to avoid losses, rendering the insurance pool ineffective when it was needed most.

QWhat 'hidden cost' does the shared pool model like Aave's incur during normal market conditions?

AA significant hidden cost is idle capital due to low capital efficiency. In Aave's top three markets, the spread between deposit and borrow rates (25-35%) represents idle funds that earn no yield for depositors but whose cost is still borne by borrowers, leading to an estimated annual value drain of $52 million.

QWhy are regulated traditional financial institutions like Coinbase and Apollo Global Management choosing Morpho over Aave for their DeFi lending operations?

AThey choose Morpho because its isolated market model allows them to set their own risk parameters and choose their risk stewards, giving them full control and compliance. This is essential for meeting regulatory requirements, unlike Aave's model where all market rules are subject to community DAO votes.

你可能也喜歡

崔泰源离婚案落槌:揭秘SK海力士万亿帝国背后的继承暗线

2024年底,SK集团会长崔泰源在家族活动上向子女强调“饮水思源”与继承责任。此时,旗下SK海力士市值已突破1000万亿韩元,成为韩国最值钱资产,但集团第三代接班格局却与传统财阀剧本迥异。 崔泰源与前总统卢泰愚之女卢素英育有三名子女。长女崔允贞被视为最明显接班候选,她拥有生物学背景和咨询经历,现任SK生物制药高管及集团“成长支援部”主管,主导精准医疗等新业务,其婚姻也联姻AI领域创业者。 次女崔敏贞路径独特,曾自愿服役韩国海军并参与亚丁湾护航,退役后曾在SK海力士美国部门处理国际政策,后离职在硅谷创立AI医疗公司。她与曾服役美国海军陆战队的华裔企业家结婚,连接军旅与地缘政治网络。 长子崔仁根最符合传统继承人形象,毕业于布朗大学物理系,曾任职SK旗下能源公司,后转入麦肯锡首尔办公室。他公开表现极为低调,未持有集团股份,也未公开表态。 子女们的成长与父母旷日持久的离婚诉讼交织。2025年,最高法院将涉及1.38万亿韩元财产分割的判决发回重审,期间三名子女曾向法院递交未公开内容的请愿书。 随着SK海力士在AI时代成为全球核心地缘政治资产,崔家第三代继承的已非简单的企业控制权。他们被置于AI科研、华盛顿政策圈与全球投资前沿,必须证明自己有能力应对新时代的产业博弈,而非自动承接旧式家族剧本。

marsbit昨天 09:06

崔泰源离婚案落槌:揭秘SK海力士万亿帝国背后的继承暗线

marsbit昨天 09:06

交易

現貨

熱門文章

如何購買DAO

歡迎來到HTX.com!在這裡,購買DAO Maker (DAO)變得簡單而便捷。跟隨我們的逐步指南,放心開始您的加密貨幣之旅。第一步:創建您的HTX帳戶使用您的 Email、手機號碼在HTX註冊一個免費帳戶。體驗無憂的註冊過程並解鎖所有平台功能。立即註冊第二步:前往買幣頁面,選擇您的支付方式信用卡/金融卡購買:使用您的Visa或Mastercard即時購買DAO Maker (DAO)。餘額購買:使用您HTX帳戶餘額中的資金進行無縫交易。第三方購買:探索諸如Google Pay或Apple Pay等流行支付方式以增加便利性。C2C購買:在HTX平台上直接與其他用戶交易。HTX 場外交易 (OTC) 購買:為大量交易者提供個性化服務和競爭性匯率。第三步:存儲您的DAO Maker (DAO)購買DAO Maker (DAO)後,將其存儲在您的HTX帳戶中。您也可以透過區塊鏈轉帳將其發送到其他地址或者用於交易其他加密貨幣。第四步:交易DAO Maker (DAO)在HTX的現貨市場輕鬆交易DAO Maker (DAO)。前往您的帳戶,選擇交易對,執行交易,並即時監控。HTX為初學者和經驗豐富的交易者提供了友好的用戶體驗。

520 人學過發佈於 2024.12.13更新於 2026.06.02

如何購買DAO

相關討論

歡迎來到 HTX 社群。在這裡,您可以了解最新的平台發展動態並獲得專業的市場意見。 以下是用戶對 DAO (DAO)幣價的意見。

活动图片