You might want to check out the takeover model of perp DEXs for lending protocols.
Written by: Eric, Foresight News
In the past week, the news around the Kelp DAO attack has been mostly positive.
Arbitrum, despite its centralized reputation, forcibly transferred over 30,000 ETH from hacker addresses. Aave's DeFi United has also received around 102,000 ETH. With the ETH frozen by Arbitrum, the total available to cover the gaps left by the hack now exceeds the amount stolen by the hacker.
Meanwhile, assets like WBTC and USDT0 that were paused for cross-chain transactions via LayerZero right after the hack are now back up and running. The industry's sentiment towards this major incident has gradually shifted from pessimism to a sense of unity in tackling the issue.
Compared to user issues that are likely to be resolved, Aave's own problems seem to be more complicated. Apart from subjective confidence factors and the visible issues of strengthening risk management, the low health positions that may exist on the platform present a higher-priority direct risk.
The origin of this concern comes from a tweet released by aixbt.
aixbt reported that it monitored addresses related to MEXC using $260 million worth of ETH and WBTC as collateral to borrow nearly $260 million in USDC and USDe. Currently, the health score of this position is only around 1.01, and with daily interest of $110,000, the health score is deteriorating at a rate of about 0.04% per day, potentially facing forced liquidation in 6 to 8 days.
However, after my actual verification, the rumored related address 0xcfdCb934657C1FB627Db8F2b690EBbaFD7a46E31 is not associated with MEXC. A user named Leon Crypto also stated that he checked 4 main addresses related to MEXC, and none had positions on Aave. aixbt also warned shortly after the tweet that its monitored addresses had significantly reduced leverage.
Because specific address information is unknown, I can't verify the authenticity of this news. But it does spark another discussion: after many panicked and borrowed all the stablecoins on Aave, which positions are now facing similar liquidation risks?
Fortunately, there are currently no large borrowing positions with low health scores on Aave, and the overall risk is under control.
According to data from Chao Labs, which exited the Aave ecosystem, as of the time of writing, the estimated value of positions facing liquidation (health score below 1.2) on Aave's Ethereum mainnet is under $100 million. The most common high-risk borrowing positions using WBTC and WETH as collateral amount to about $86 million, spread across over 1,100 addresses, presenting no significant risk overall.
Additionally, according to data from DeFi Saver, although there are many low health positions worth over $100 million and even $1 billion, almost all of them use the E-Mode mechanism (where similar assets can have a very low collateral ratio, like using wstETH to borrow ETH, or using USDT to borrow USDC). Unless a black swan event occurs, the risk is manageable.
However, on the flip side, ETH, USDT, USDC, and even USDe on Aave are still either fully borrowed or close to being borrowed out.
Although things have improved compared to the state of being fully borrowed after the Kelp DAO attack, currently, over 95% utilization is still above normal levels, meaning there are still a large number of panic borrowers who haven't repaid their debts.
Even though, as previously analyzed, the data doesn't support the occurrence of 'systemic risk', it still warrants our vigilance.
Currently, the overall crypto market is in a warming phase, so we’ll only discuss Aave's risk management issues, not the overall risk. However, if the Kelp DAO hack happened during a rapid market downturn, that's another story:
Many users are panicking and choosing to borrow stablecoins at the highest ratio. As the prices of quality collateral (like WBTC and WETH) drop, these users have given up on redeeming their collateral, resulting in a flood of Bitcoin and Ethereum hitting the market. During the downturn, some whale users improved their health score by repaying part of their loans, but those stablecoins were quickly borrowed again, triggering further liquidations.
As liquidations progressed, ultimately, whale users also gave up resisting, leading to cascading liquidations. These scenarios, which may seem a bit conspiratorial for some newcomers, actually happened on May 19, 2021. If it weren't for the Ethereum network congestion at that time, even more ETH might have hit the market as it reached liquidation levels.
In extreme cases, spot sell-offs can be more powerful than liquidations in the futures market. On October 11 last year, the total liquidation amount in the futures market approached $20 billion, but if the spot market had a $20 billion market sell-off, Bitcoin wouldn't just stop at around $100,000.
Risk management in the DeFi lending market should always come before returns. Situations like the sudden appearance of over 100,000 rsETH are highly uncontrollable; we don't know what other methods hackers might use to exploit vulnerabilities that we can only think of in hindsight. For a DeFi platform like Aave, besides reflecting on asset admission standards, it should also prepare for those 'thankfully not yet occurred' problems that might happen under extreme situations.
In response to the potential issues mentioned above, the fund established by Aave shouldn't just bear the responsibility of 'compensation', but should also take on the responsibility of 'problem-solving'. For example, when another black swan event causes users to panic and borrow assets, the fund should have special addresses to directly take over these 'panic positions' during liquidation, and after replenishing capital with the fund's resources, gradually offload the liquidated positions as the market recovers.
Hyperliquid's liquidation vault is doing something similar; it's like the government intervening in the market during the 2008 subprime crisis. If things are allowed to worsen, the same rescue methods might be ineffective after a complete market collapse. The 100,000+ ETH raised by Aave would cover bad debts without liquidation, but if a cascading liquidation occurs, who will fund the affected users?