[TEMP CHECK] Pure-Collateral Only Product

Governance General
1

Title: [TEMP CHECK] Pure-Collateral Only Product

Author: HB156

Date: 2025-12-29

Summary

Introduce the ability provide collateral without being forced to lend it. Collateral will earn 0% interest.

Motivation

Providing the ability to post only collateral without lending it introduces two key features that can improve Aave growth: 1) growing user adoption and 2) mitigating extreme risks

  1. Grow User Adoption

In order to borrow, you must provide collateral. And in order to provide collateral you are forced to lend.

This effectively bars users who want to borrow, but do not want to take on credit counterparty risks from lending their collateral. Not all borrowers want to take on systemic counterparty risk. That risk aversion to lend out can be high then even with a deficit offset and umbrella module. Similar to traditional finance, users only take on their own borrowing risk without also lending out their collateral to the broader ecosystem.

This can improve utilization in the protocol. Currently there are much more lenders than there are borrowers.To put this in perspective, lending USDC is currently earning a rate of 3.1%, which is lower than short term U.S. treasuries at ~3.5% and arguably much higher risk.

This imbalance can become even more acute as Aave labs is working to allow users to deposit up to $1 million insurance and earn a savings rate on it. That could burden Aave with a deposit heavy model. Those are liabilities to the DAO. Collateral only products could encourage borrowing and improve revenues for the DAO.

Future use cases naturally extend to RWAs. Users will be able to post RWAs as collateral to take loans. This can be an effective tax offset for users who don’t want to sell their holdings, or an ability to add leverage. A positive for DAO revenues.

Mitigating Extreme Risk

Providing the potential to post only collateral without lending could also improve risk frameworks/management for the protocol in the event of extreme left tail risks.

In the current model, when a user borrows they are exposed to 2 primary risks. 1) Their health factor falling if the value of their collateral falls. 2) Their health factor falling if their collateral is slashed. The user really only has control over point #1, they are completely exposed to market risks in point #2.

Borrowing risks are currently correlated to lending risks.

For example, take the scenario of a catastrophic event that depletes the deficit offset and umbrella module. Lender funds could get slashed as bad debt rises. That would catalyze a forced unwind to liquidate more collateral as health factors fall. By providing a collateral only product that is divorced from lending risk, you could stem the contagion risks especially if individual collateral/balance sheets are healthy.

Disclaimer

I am not compensated by any Aave Labs, ACI, BGD Labs or any other related Aave/DeFi organization.

Next Steps

Gather community feedback on this TEMP CHECK.

If consensus is achieved, move this proposal to the TEMP CHECK snapshot stage.

Copyright

Copyright and related rights concerning the transparency act are waived via CC0

2

In the current model, the borrowers are paying the lenders. borrowers’ APY exceeds the APY of lenders of that asset (difference goes to the protocol) and all the borrowings are overcollateralized (you have to lend more in value than you can borrow, see LTV %)

In your model where collateral only would be deposited with no lending allowed, where would the APY come from? Or do you mean deposit as collateral only with no APY, only to allow borrowing ?

Also there must be a source of assets for borrowing. If many people think like you suggest here and deposit as collateral only, the protocol would quickly become short in liquidity for borrowers, how would you tackle this issue ?

2 Likes
3

Hey,

If this post is meant to be a proposal, then please follow the guidelines for a TEMP CHECK.
I will move this post into off topic for now.

4

In that scenario, as utilisation approaches 100%, the interest rate model would automatically push the borrowing APY higher, which creates a strong economic incentive for new users to deposit tokens on the platform. In parallel, the rising borrow rate also naturally discourages additional borrowing and encourages repayment.
So the protocol’s rate curve is designed to balance itself: if liquidity tightens → rates spike → supply increases → utilisation falls back into equilibrium.

5

Yes. Collateral will not earn any interest.

To add to this, essentially the protocol does not operate at 100% utilization as a majority of people lend and don’t borrow. There is a strong imbalance towards lenders vs borrowers.

To put this in perspective, lending USDC last time I checked earned a rate of 3.1%, which is lower than short term U.S. treasuries at ~3.5% and arguably much higher risk.

I would also think this imbalance will rise if the consumer app Aave labs is working on allows users to deposit up to $1m insurance and earn a savings rate on it. Could make Aave a deposit heavy model? Those are liabilities to the DAO. How can we reinvest those within the ecosystem.

I do agree though that the protocol benefits from adding liquidity and from having a buffer of more money lent vs borrowed to ensure stability if you have large redemptions.

Apart from natural market dynamics that will entice users to supply if borrow rates rise. In a pilot program maybe liquidity could be enhanced if the $$ amount of a collateral-only product had a variable ceiling max (ie. the total amount of collateral-only ETH would be capped at x% of total borrowable capacity). Just a thought.

Or you can add a spread to the borrow rate for someone using only collateral since they are not providing liquidity.

6

I would say lenders and borrowers are generally two different user profiles. Someone lending USDC is usually just looking for yield, while someone borrowing stables is often doing it against volatile assets like ETH or BTC to gain leverage or manage liquidity. Because of that split, we don’t end up in a scenario where “everyone user lends and then borrows to the max” pushing utilization close to 100% across the entire market.

If the AAVE consumer app ends up bringing a large inflow of stablecoins, utilization on those markets would drop, lending APYs would go down, and that naturally pulls in more borrowing until things balance again. The buffer you mentioned (more supplied than borrowed) is already handled by the interest-rate model: each asset market has an optimal utilisation target, and once it passes that kink, borrowing becomes a lot more expensive, which slows demand and keeps liquidity healthy.

On the “liability to the DAO” point, the interest paid to suppliers doesn’t work like a bank deposit liability. AFAIK, AAVE always captures a spread between the borrow and supply rates on the same asset, so there’s no scenario where the DAO would ever pay out more than it earns.

7

Essentially then adding a pure-collateral product fits well within the protocol. It can add users that aren’t willing to take on lending risks. And can dampen extreme market movements in the situation deficits overwhelm the offsets and umbrella module.

A big use case will likely be as more RWAs are onboarded, you can take out loans backed by your assets that are separated from the broader risk pool.

Can you explain why the consumer app wont be a liability to the DAO? From my understanding, deposits will be in gho that earn a base 5% rate that will be presumably paid by DAO funds.

8

This could be possible, yet as user deposits will not generate any revenue for the protocol, I expect the users who deposit to this section/wrapper to pay a premium. As EzR3aL said, please follow guidelines for a TEMP CHECK if you expect this to become a proposal.

1 Like
9

Please, like I already said, follow the guidelines and create a proper TEMP CHECK and then you can surely discuss the details of this.
Im going to close this thread till you change it. After the change im happy to open it again.

Edit: I deleted the other post and moved it here.

Thanks

3 Likes
12

Will move this to snapshot voting soon.

13

Hi @HB156, thanks for opening the topic. Indeed, the ability to borrow against your asset without lending out your collateral is something some users are requesting, even if it’s not the main flow. For example, ETH-derived collateral earns significant yield that reduces the net yield for users, thus lowering the cost of capital, and most users are fine with lending out their collateral.

That being said, there have been quite a lot of requests for this feature, either because users don’t receive enough yield on a particular collateral, don’t want to associate it with lending risk, or have some kind of favorable tax positioning.

One way we are thinking of enabling this feature is by providing a spoke for Aave V4 that offers collateral isolation in a segregated vault. This way, users who do not want to lend out their collateral for the reasons above can simply opt into a vault. Separately, if the DAO wishes, it can also apply risk premiums to these collateral vaults to compensate for the reduction in protocol revenue from not lending out the collateral. This is ultimately a choice for the DAO between favoring user acquisition and a better product versus capturing fees, which needs to be balanced carefully.

Regarding the RWAs comment, I agree that static collateral fits this vision well. However, there is also a broader opportunity to capture the securities lending market in the future by enabling collateral to be lent out.

I would personally like to see a more fleshed-out proposal with technical details before going to a snapshot, and I would be happy to work on this together.

14

Isn’t the model you described exactly what our competitor Morpho uses? The collateral is only used as collateral and is never lent out. That’s exactly Morpho’s architecture.

15

I mean, you could technically build different models as v4 spokes: Fluid with LPs as collateral, Morpho-style where collateral only secures borrowing and is never lent out, or even Euler-style designs for cross-collateralized, monolithic, or isolated markets. The difference here feels more about how these are implemented within Aave’s modular framework rather than introducing a new economic model.

Framing this as spoke could make sense without impacting v3 core markets. Either way, v3 is here to stay so this could also be a cool low-risk feature for @bgdlabs could explore.

16

Hello, this is already an option and AAVE asset is a good example of this.

With Aave 3.6 this makes it even easier to implement.

That being said, it’s not desirable for all kinds of assets as a large perk of Aave, and what we did at the @ACI for the past year was to create mutualistic synergies between assets, creating liquidity flywheels.

A good example of this was the LRT → LST → wETH loops that allowed the protocol to earn dozens of millions of ARR and increased our liquidity by dozens of billions of deposits.

Best usecase is for “governance” tokens that present very little interest to be borrowed, such as UNI, 1INCH, ENS, and so on, 3.6 allows 0% LTV emodes with locked-in pairs of debt assets.

UNI → stablecoins emode with 0% LTV outside of it would prevent unwanted risk/reward behaviour such as shorting wBTC with UNI as collateral.

1 Like
17

This is older than Morpho. This concept has been around for some time, for example in MakerDAO you deposited ETH to borrow DAI, and your ETH collateral was not being lent out. Same in Synthetix and some other protocols.

I think the idea makes sense if some users want it, and it should be possible to do in v4.

1 Like