[ARFC] Deploy an Ethena Aave v3 Instance

Governance
1

[ARFC] Deploy an Ethena Aave v3 Instance

Author: ACI

Date: 2024-08-07

ARFC updated with Risk Parameters provided by Risk Service Providers - 2024-08-14

Summary

This proposal suggests deploying an Aave v3 instance focused on providing liquidity for sUSDe and USDe holders to borrow stablecoins.

Motivation

Ethena’s USDe stablecoin has grown to over $3b TVL with success as a supply and borrow asset on the mainnet Aave v3 instance. It has also seen large demand for use as collateral on other lending protocols.

In order to capitalize on this demand, we propose deploying a dedicated Aave v3 instance for Ethena’s stablecoins and related collateral assets.

To start the instance will use sUSDe and USDe as main collateral assets, with USDC, FRAX and GHO as supply assets.

We also propose providing the Ethena instance with an initial GHO facilitator.

The goals for this instance will be to grow Aave market share for borrowers using sUSDe and USDe as well as create an additional steady revenue stream of GHO.

Specification

To further leverage synergies with friendly ecosystem participants, ACI proposes that pyUSD and FRAX to be included as non-collateral assets for the Ethena Aave instance.

Following the recent FRAX Finance AMO proposal which creates an AMO to deposit up to 100,000,000 FRAX in the Aave V3 Frax pool, it is believed this will create significant opportunities which can be capitalised on through this addition to the proposed Ethena Aave instance.

The following Risk Parameters have been updated after feedback from Risk Service Providers:

sUSDe USDe USDC FRAX pyUSD
Isolation Mode No No No No No
Enable Borrow No No Yes Yes Yes
Enable Collateral Yes Yes No No No
E-Mode Category Stablecoin Stablecoin Stablecoin Stablecoin Stablecoin
Loan To Value 72% 72% - - -
Liquidation Threshold 75% 75% - - -
Liquidation Bonus 8.5% 8.5% - - -
Reserve Factor - - 10% 20% 20%
Liquidation Protocol Fee 10% 10% 10% 10% 10%
Supply Cap 10,000,000 90,000,000 80,000,000 15,000,000 60,000,000
Borrow Cap 0 0 72,000,000 12,000,000 54,000,000
Debt Ceiling N/A N/A N/A N/A N/A
uOptimal - - 90% 90% 90%
Base - - 0% 0% 0%
Slope1 - - 5.5% 5.5% 5.5%
Slope2 - - 60% 80% 80%
Stable Borrowing Disabled Disabled Disabled Disabled Disabled
Flashloanable Yes Yes Yes Yes Yes
Siloed Borrowing No No No No No
Borrowable in Isolation No No No No No

For the GHO facilitator, it is recommended a starting cap of 4.75M GHO (5% of total supply) and a borrow rate aligned with the main GHO facilitator.

The proposed Ethena Aave v3 instance will implement the following:

  • The market will include sUSDe, USDe, USDC, FRAX, pyUSD, GHO and a new GHO facilitator. Additional assets to consider post-launch are Ethena’s Pendle PT tokens and rsUSDe.
  • This GHO facilitator will be limited in size to start both to limit risk to Ethena collateral and protect liquidity and the GHO peg.
  • The market will have stablecoin E-Mode to allow for high leverage borrowing between the correlated assets.
  • The market will use the existing sUSDe and USDe capped oracles, so that borrowers are not affected by fluctuations in the secondary market price.
  • Borrow parameters will match the mainnet Aave v3 instance, making it the most competitive location to borrow against sUSDe and USDe.

Useful Links

[TEMP CHECK] Deploy an Ethena Aave v3 Instance

[TEMP CHECK Snapshot] Deploy an Ethena Aave v3 Instance

Disclaimer

The ACI has not been compensated by Ethena for the publication of this proposal.

ACI team members are holders of sUSD, USDe and ENA.

Next Steps

  1. Publish a standard ARFC and collect community & service provider feedback before an ARFC snapshot.
  2. If the ARFC snapshot passes, publish an AIP vote for final confirmation and enforcement.
  3. Work with service providers and Ethena to release and promote the new Aave V3 Instance.

Copyright

Copyright and related rights waived under CC0

1 Like
2

Summary

LlamaRisk supports deploying a new Ethena Aave v3 instance on Ethereum, with sUSDe and USDe as collateral-only assets, USDC and FRAX as borrowable assets, and a GHO facilitator. We’ve conducted due diligence on these assets (PYUSD - Apr 2024, FRAX/sFRAX - Dec 2023) and find that the initial parameters proposed are suitable. Bundling a GHO facilitator with this market represents a good opportunity, allowing USDe and sUSDe to become new collateral types. We propose safe initial mint caps of 4M GHO alongside parameters for DAO consideration, jointly discussed with @ChaosLabs.

We note that further assets, such as DAI and USDT, could be introduced and more exotic collateral types, such as Ethena’s Pendle PT tokens and rsUSDe. These could alter this instance’s risk profile, and consideration for onboarding will warrant a thorough examination.

Motivation

The proposed dedicated instance for Ethena’s sUSDe and USDe aims to enhance leveraged borrowing efficiency within the Aave ecosystem. Currently, both assets are in Isolation Mode on mainnet, limiting their utility while mitigating potential risks to the protocol. This dedicated instance allows for more aggressive parameters by isolating potential risks associated with these assets.

Moreover, introducing a GHO facilitator within this instance and including USDe and sUSDe as new collateral types for GHO will contribute to the expansion and diversification of the GHO stablecoin ecosystem.

Current State of Ethena

After growing to $3.6B of TVL, Ethena protocol has slightly declined this past month. The TVL at the time of writing is now $3.1B, which is rational given current market conditions and associated lower funding rates. In addition to BTC and ETH perpetuals, Ethena introduced SOL as a perpetual futures trading asset. This will help further diversify the risk and increase scaling opportunities regarding open interest.

Ethena has also recently introduced an initial phase of governance, onboarding risk providers to govern different risk parameters, such as reserve fund capitalization, liquid redemption buffer, take rates, collateral allocation between different CEX, etc. LlamaRisk is also part of this risk council and has covered different pain points of the protocol in a dedicated Ethena research. The council will be working on these issues to ensure the stability and lower risk of the protocol.

The ongoing Sats campaign, part of Ethena’s Season 2 initiatives, incentivizes a wide array of DeFi activities. These include ENA/USDe locking, liquidity provisioning, Pendle locks, depositing and borrowing on money markets (among which depositing USDe on Aave V3), and Layer 2 activities. However, it is notable that USDe holders are not encouraged to stake for sUSDe or provide sUSDe liquidity during the current campaign.

USDe liquidity

The main liquidity for USDe resides in Curve and Uniswap V3 Liquidity Pools. The overall size of the liquidity is $245M with a USDe portion of ~$100M (3.3% of TVL). USDe is mostly paired with other stablecoins like FRAX, USDC, USDT, and GHO. The liquidity has remained stable, with minor drawdowns during adverse market conditions (e.g., market stress on August 5th, 2024).

image
Source: USDe liquidity on DEX. Source: Dune Analytics, August 13th, 2024

sUSDe liquidity

1.4B sUSDe is outstanding, representing approximately 45% of the total USDe supply. However, the liquidity situation for sUSDe is significantly worse than that of USDe. The liquidity-to-supply ratio for sUSDe is below 1%, with total sUSDe liquidity remaining under 10M for an extended period - the lowest level since April 2024.

image
Source: sUSDe liquidity on DEX. Source: Dune Analytics, August 13th, 2024

It’s important to note that sUSDe can be unstaked with a current cooldown period of 7 days. This feature reduces the need for immediate liquidity for users who wait through this period, potentially earning additional yield due to sUSDe trading at a discount. Historically, this discount has been modest (maximum of -400 bps), but vigilance is required for potential Ethena scenarios, as capped oracles would be used. If sUSDe begins trading at larger discounts, there’s a risk that discounted sUSDe could be deposited as collateral at a par value while other stablecoins are borrowed as exit liquidity.

image
Source: Historical sUSDe discount. Source: Dune Analytics, August 13th, 2024

Given that the points program incentivizes the USDe liquidity provision while sUSDe liquidity does not, this indicates a problem of non-sustainable liquidity. Therefore, there remains a risk that liquidity could also thin out for USDe in case the points
program is terminated.

Current state of GHO

After an extended conservative growth period, GHO has entered a more aggressive scaling phase. It is a great opportunity to open up new collateral assets for GHO, and the isolated Ethena instance would facilitate that. USDe and sUSDe would be onboarded as new collateral types since they are isolated assets on the main market, and therefore, GHO is not borrowable using them.

image

Source: Chaos Labs GHO Dashboard, August 11th, 2024

It’s important to recognize that 25% of all GHO liquidity currently resides in the USDe/GHO pair on Curve, which remains substantial while down from a previous high of around 50%. Although the existing USDe/GHO pairing has not posed significant risks thus far, introducing additional dependencies between GHO and USDe warrants attention. In particular, the potential for unintended consequences, such as those arising from a USDe de-peg, should be monitored to ensure the continued stability of GHO.

Aave V3 Specific Parameters

We discussed and aligned unified parameter recommendations with @ChaosLabs; they will be provided shortly.

We recommend implementing stablecoins e-mode (90% LTV/93% LT) for this instance. Supply caps for USDe and sUSDe have been chosen conservatively according to liquidity, redeemability, and the fact that Aave will use the internal rate + CAPO.

For the GHO facilitator, we recommend a starting cap of 4M GHO and a borrow rate set 0.5 percentage points lower than the main market GHO facilitator to incentivize borrowing.

3

Summary

Chaos Labs supports a new Ethena Aave v3 instance on Mainnet and provides the recommended parameters below.

Motivation

We support the proposed deployment with the intention of facilitating stablecoin supplying and borrowing. We recommend listing all assets in stablecoin E-Mode, with an LTV of 90%, LT of 93%, and Liquidation Penalty of 1%. Given most users will utilize E-Mode, our supply cap methodology in this instance tests for liquidity under 1% slippage for each asset, then recommends a supply cap at double this value.

sUSDe

  • We recommend aligning sUSDe’s risk parameters with those in the main Aave V3 instance, including making the asset non-borrowable. As discussed in our previous sUSDe recommendation, borrow demand for yield-bearing assets is typically minimal.
  • Additionally, we note that supply has been relatively limited in sUSDe’s existing V3 market, with just $970K supplied as of this writing. The share of USDe staked as sUSDe has has fallen back to its level at listing of 45% following a brief increase to 50%. The Sats campaign has contributed to keeping the staking rate relatively low and we anticipate a shift in supply demand when incentives slow or end.

USDe

  • We recommend also aligning USDe’s risk parameters with those in the main instance. However, in this instance we recommend disallowing borrowing given relatively limited historical demand to borrow the asset, as well as ability to borrow in the main instance. Finally, this does not align with the stated purpose of this instance.

image (79)

  • As discussed above, it is possible that the balance of sUSDe and USDe will shift as Ethena incentives change; should this happen, we will recommend appropriate cap increases and decreases.

USDC, FRAX, and pyUSD

  • We recommend aligning USDC’s parameters with those on the main instance while disabling the asset as collateral, again to better serve the stated purpose of the instance. Allowing the asset as collateral could also lead to USDC-USDC looping if this market is incentivized, taking up valuable borrowable USDC.
  • We note that Slope1 is lower than the main instance’s current Slope1; this is due to MakerDAO’s proposed reduction in the Dai Savings Rate. We will shortly be proposing reductions to stablecoin interest rates on Aave V3.

FRAX and pyUSD

  • We recommend listing these assets with similar parameters as on the main instance. However, we recommend increasing pyUSD’s UOptimal given the setup of this instance, as well as the fact that the asset cannot be used as collateral.
  • Again, we note that Slope1 is lower than what was originally proposed. This is due to the anticipated decrease in the DSR; markets on Aave V3 will be aligned shortly.

Specification

We recommend aligning Stablecoin E-Mode parameters with those on Optimism: 90% LTV and 93% LT.

For the GHO facilitator, we recommend a starting cap of 4.75M GHO (5% of total supply) and a borrow rate aligned with the main GHO facilitator.

sUSDe USDe USDC FRAX pyUSD
Isolation Mode No No No No No
Enable Borrow No No Yes Yes Yes
Enable Collateral Yes Yes No No No
E-Mode Category Stablecoin Stablecoin Stablecoin Stablecoin Stablecoin
Loan To Value 72% 72% - - -
Liquidation Threshold 75% 75% - - -
Liquidation Bonus 8.5% 8.5% - - -
Reserve Factor - - 10% 20% 20%
Liquidation Protocol Fee 10% 10% 10% 10% 10%
Supply Cap 10,000,000 90,000,000 80,000,000 15,000,000 60,000,000
Borrow Cap 0 0 72,000,000 12,000,000 54,000,000
Debt Ceiling N/A N/A N/A N/A N/A
uOptimal - - 90% 90% 90%
Base - - 0% 0% 0%
Slope1 - - 5.5% 5.5% 5.5%
Slope2 - - 60% 80% 80%
Stable Borrowing Disabled Disabled Disabled Disabled Disabled
Flashloanable Yes Yes Yes Yes Yes
Siloed Borrowing No No No No No
Borrowable in Isolation No No No No No
2 Likes