Hello friends, today we are talking about Aave.
Aave's predecessor was ETHlend, founded by Stani Kulechov in 2017 through an ICO. The name 'lend' reflects that this is a lending protocol, and the prefix ETH indicates that it is primarily based on Ethereum assets, allowing users to borrow assets through collateral. There was a simple introduction to the decentralized lending business model in previous posts; you can check out the question 'Can you still get loans without banks? What is decentralized lending?'
In 2019, ETHlend officially changed its name to Aave. Aave means 'ghost' in Dutch. It implies that although the medium is transparent, it is powerful, corresponding to a lending protocol that can provide transparent, safe, and efficient services.
As of now (October 2024), Aave has accumulated $18.5 billion in deposits. Due to the existence of over-collateralization, we estimate that the total loan amount is around $10 billion.
What scale does a hundred billion dollars in loans represent?
I found the amounts of loans from national financial institutions. China's is about $35 trillion, and the U.S. is $12 trillion. Compared to centralized lending, decentralized lending is still in a very early stage, two orders of magnitude behind a single large country.
However, a scale of a hundred billion dollars is not very small. It is roughly comparable to Latvia or Côte d'Ivoire, allowing it to compete with small countries.
For those interested, if you want to know more about the lending amounts in different countries, you can check it in CEIC https://www.ceicdata.com/en/indicator/united-states/total-loans
The development of Aave was not achieved overnight. From the fluctuations in token prices, we can look back at its development history over the past few years.
The chart below shows the fluctuations in Aave's token price (the price of Aave before 2020 is filled with the price of lend, calculated based on the exchange rate at that time = 100:1, data source Dune).
After the release of V1 in 2019, there was no significant fluctuation in token prices. The first noticeable increase occurred in the second half of 2020, likely coinciding with the hot defisummer market. This was followed by a brief decline in October with the exchange of two tokens.
In December 2020, the release of V2 coincided with a hot market, and token prices soared past $500.
In March 2022, V3 was released. Even though the market was weak at that time, with BTC hovering around $30,000, the market still highly recognized the release of the new version, showing a slight increase against the trend.
Looking back from 2024, it's like what the elder said: 'A person's destiny relies on self-effort, but historical processes must also be considered.' The fluctuations in Aave's token price are similar, with a surge driven by defisummer and the entire market's helplessness in a downturn.
Market assistance and pressures are indiscriminately spread, but self-effort under external factors is the main reason why it can go far.
Next, let's take a look at what self-improvement Aave has made in efficiency enhancement during the iterations of these three important versions.
From P2P to P2P
During the ETHLEND era, the protocol provided peer-to-peer lending capabilities, but overly dispersed demand was not efficient for both borrowers and lenders. In version V1, this feature was quickly iterated to peer-to-pool with the concept of a liquidity pool, where contributors could inject funds into the pool for staking, while borrowers could borrow assets from the pool. The shift from peer-to-peer to peer-to-pool marked the first step in enhancing matching efficiency and lending efficiency.
From unified pricing to asset classification operations
Before the V3 version, all staked assets needed to be converted to ETH for valuation before participating in lending calculations. This model is consistent with the initial name of the protocol, ETHLend, which is simple and easy to understand when uniformly priced in ETH. However, converting all kinds of assets into ETH valuation also has its drawbacks.
First of all, not all assets have a positive correlation with ETH prices. Some assets are uncorrelated or even negatively correlated with ETH prices. Pricing in ETH can lead to undervaluation of collateral assets, resulting in a reduced borrowing amount and subsequently lowering the overall asset utilization efficiency.
Secondly, unifying all mortgages and loans in ETH ignores the relationship between the price fluctuations of collateral assets and loaned assets. For example, if both the collateral and loan are stablecoins, the prices of the two are highly correlated. Even in a highly volatile market, the risk of liquidation is relatively low. In contrast, pledging high-volatility assets to borrow stablecoins carries a much higher risk of liquidation.
The so-called E-mode refers to classifying assets, where assets belong to highly correlated collaterals and borrowings. For example, if both the collateral and borrowed assets are stablecoins or related to ETH, or BTC-related assets, the capital utilization rate will be higher.
From a mixed stew to isolating high risks
The so-called Isolation Mode is also aimed at improving lending efficiency.
In the past, without asset classification, high-risk assets could also enter the pool and enjoy the same lending ratios and liquidation thresholds. This led to two problems: first, new asset entry into the pool needed to be discussed and approved by the community, which was not efficient; second, since the entire fund pool had to accommodate high-risk assets, for safety reasons, the overall lending ratios and liquidation thresholds would compromise for high-risk assets.
Isolation mode involves separately isolating high-risk assets and establishing safety-related coefficients for them. Other assets do not need to accommodate them, allowing for a higher lending ratio and greater improvements in lending efficiency.
Asset classification and refined operations
Whether it's E-Mode or Isolation Mode, this classification operation method is actually familiar. In internet companies, this approach is called 'refined operation.'
Compared to a mixed stew, refined operations mean classifying merchants and users and matching different resources for different tiers.
For example, providing different resources and rights to merchants at different stages can help them grow better.
The same goes for users. Newspapers and TV provide the same information for everyone, while platforms like TikTok continuously tag you based on your preferences, providing you with content you want to see. So, compared to watching TV, are you more willing to scroll through TikTok? It provides you with 'more suitable' content, which broadly speaking is also TikTok's 'refined operation' for you.
For Aave, the logic is also similar. Distinguishing high-correlation assets can match higher asset utilization efficiency. For high-risk assets, the safety coefficient can be moderately raised.
Classifying assets and matching appropriate lending ratios and liquidation thresholds actually allows for different efficiencies and safety coefficients for different assets.
Using a single logic for all assets will inevitably compromise overall lending efficiency for the safety of certain high-volatility assets. By iterating to apply different coefficients for different types of assets, some assets can first enhance efficiency, thus improving the overall lending efficiency of the fund pool.
Today we'll stop here, friends, see you next time~
