From Aave to Ether.fi: Who has captured the most value in the on-chain credit system?
- 核心观点:借贷协议是DeFi信用价值链中价值捕获最强的环节。
- 关键要素:
- 金库支付给借贷协议的利息超过其自身收入。
- 借贷协议收入取决于借款规模,相对稳定且可观。
- 借贷层创造的经济价值超过上游资产发行方。
- 市场影响:凸显借贷协议核心地位,挑战“分发为王”叙事。
- 时效性标注:中期影响。
Author | @SilvioBusonero
Compiled by Odaily Planet Daily ( @OdailyChina )
Translator | Dingdang ( @XiaMiPP )

As vaults and curators gain market share in the DeFi world, questions are emerging: Are lending protocols seeing their profit margins squeezed? Is lending no longer a good business?
However, if we shift our perspective back to the entire on-chain credit value chain , the conclusion is quite the opposite. Lending protocols still constitute the strongest moat in this value chain. We can quantify this with data.
On Aave and SparkLend, the interest fees paid by the vault to the lending protocols actually exceed the revenue generated by the vault itself. This fact directly challenges the mainstream narrative that "distribution is king."

At least in the lending sector, distribution is not king.
In short: Aave not only earns more than the various vaults built on top of it, but also surpasses the asset issuers used for lending, such as Lido and Ether.fi.
To understand why, we need to dissect the entire value chain of DeFi lending and re-examine the value capture capabilities of each player by following the flow of funds and fees.
Lending Value Chain Deconstruction

The annualized revenue of the entire lending market has exceeded $100 million . This value is not generated by a single link, but by a complex stack of components: the underlying settlement blockchain, asset issuers, lenders, the lending protocols themselves, and the vault responsible for distribution and strategy execution.
As we mentioned in previous articles, many current lending market applications stem from basis trading and liquidity mining opportunities, and we have broken down the main strategic logic behind them.
So, who is actually "in need" of capital in the lending market?
I analyzed the top 50 wallet addresses on Aave and SparkLend and labeled the major borrowers.

- The largest borrowers are various vault and strategy platforms such as Fluid, Treehouse, Mellow, Ether.fi, and Lido (who are also asset issuers). They control the distribution capabilities to end users, helping them obtain higher returns without having to manage complex cycles and risks themselves.
- There are also large institutional investors, such as Abraxas Capital , who deploy external capital into similar strategies, whose economic models are essentially very similar to those of a vault.

But the vault isn't the whole story. This chain includes at least the following types of participants:
- User : Depositing assets, hoping to obtain additional returns through a vault or strategy manager.
- Lending agreements : These provide infrastructure and liquidity matching, charging interest to borrowers and taking a percentage as agreement revenue.
- Lenders : The providers of capital, who may be ordinary users or other financial institutions.
- Asset issuers: Most on-chain lending assets have underlying backing assets that generate returns, a portion of which are captured by the issuer.
- Blockchain Network: The Underlying "Track" Through which All Activities Occur
The lending agreement earned more than the downstream vault.
Take Ether.fi's ETH liquidity staking vault as an example. It is the second-largest borrower on Aave, with approximately $ 1.5 billion in outstanding loans. The strategy itself is quite typical:
- Deposit weETH (approximately +2.9%)
- Lend out wETH (approximately -2%)
- Vault charges TVL a 0.5% platform management fee.
Of Ether.fi's total TVL, approximately $215 million is net liquidity actually deployed on Aave. This portion of TVL generates approximately $1.07 million in platform fee revenue annually for the vault.
However, at the same time, this strategy requires Aave to pay approximately $4.5 million in interest annually (calculated as: $1.5 billion in borrowing × 2% borrowing APY × 15% reserve factor).
Even in one of the largest and most successful circular strategies in DeFi, the value captured by lending protocols is still several times that of a vault.

Of course, Ether.fi is also the issuer of weETH, and this vault itself is directly creating demand for weETH.

Even when considering both the returns from the vault strategy and the returns from asset issuers , the economic value created by the lending layer (Aave) is still higher.
In other words, the lending agreement is the link with the greatest value increment in the entire stack.
We can perform the same analysis on other commonly used vaults:
Fluid Lite ETH: 20% performance fee + 0.05% exit fee, no platform management fee. It borrowed $1.7 billion in wETH from Aave, paying approximately $33 million in interest, of which about $5 million went to Aave, and Fluid itself earned nearly $4 million.

The Mellow protocol (strETH) charges a 10% performance fee, with a loan size of $165 million and a TVL of only about $37 million. We see again that, in terms of TVL, the value captured by Aave exceeds the vault itself.

Let's look at another example. In SparkLend, the second-largest lending protocol on Ethereum, Treehouse is a key player, running an ETH revolving strategy:

- TVL approximately US$34 million
- Borrowed $133 million
- Performance fees are charged only on marginal revenue exceeding 2.6%.

SparkLend, as a lending protocol, has a higher value capture capability in the TVL dimension than Vault.
The pricing structure of a vault has a significant impact on the value it can capture; however, for lending agreements, revenue depends more on the nominal size of the loan and is relatively stable.
Even if a strategy is shifted to dollar-denominated pricing, the higher interest rates will often offset the effect of lower leverage. I don't think the conclusion will fundamentally change.
In relatively closed markets, more value may flow to curators, such as through Stakehouse Prime Vault (26% performance fee, incentivized by Morpho). However, this is not the final state of Morpho's pricing mechanism; curators themselves are also collaborating with other platforms for distribution.
Lending Agreements vs. Asset Issuers
So the question is: is it better to use Aave or Lido?
This problem is more complex than comparing vaults because the pledged assets not only generate returns themselves, but also indirectly create stablecoin interest income for the protocol through the lending market. We can only make approximate estimates.
Lido has approximately $4.42 billion in assets in the Ethereum Core market, which is used to support lending positions, generating approximately $11 million in annualized performance fees.
These positions roughly balance ETH and stablecoin lending. Based on the current net interest margin (NIM) of approximately 0.4%, the corresponding lending revenue is around $17 million , significantly higher than Lido's direct revenue (and this is a historically low NIM level).

The real moat of lending agreements
If compared solely to the profit model of traditional finance deposits , DeFi lending protocols may appear to be a low-profit industry. However, this comparison overlooks the true location of the competitive advantage.
In an on-chain credit system, the value captured by lending protocols exceeds that of the downstream distribution layer and also surpasses that of the upstream asset issuers on an overall basis.
On its own, lending may seem like a low-profit business; but within the complete credit stack, it is the layer with the strongest value capture capability relative to all other participants—the treasury, the issuer, and the distribution channels.


