[ARFC] Aave v3 Gnosis Instance Updates

Proposal edited to include Risk Providers feedback

2025/01/02 proposal edited to reflect a 2 part implementation waiting for osGNO review

[ARFC] Aave v3 Gnosis Instance Updates

Author: @ACI

Date: 2024-12-20


Summary

This ARFC proposes several updates to the Aave v3 Gnosis instance to improve capital efficiency and add new use cases on the network. The key changes include removing GNO from isolation mode, adjusting the reserve factor for EURe, onboarding osGNO, making sDAI borrowable, and creating new relevant E-modes.

Motivation

GNO has demonstrated strong stability and market presence on Gnosis Chain, making isolation mode unnecessarily restrictive and hindering network growth. Removing GNO from isolation mode will facilitate further expansion of the network.

The reduction in the EURe reserve factor aligns with the asset’s performance and incentivizes increased lending activity.

The introduction of osGNO adds utility for GNO holders within the protocol and enables looping strategies using osGNO as collateral to borrow GNO. Leveraged staking use cases have been a key growth driver for Aave DAO, and enabling this on Gnosis Chain is expected to gain similar traction.

Making sDAI borrowable and introducing new E-modes will enhance capital efficiency and foster synergies between stable assets. The unique combination of EUR and USD borrowing opportunities is a distinct advantage for Gnosis Chain.

sDAI, as an asset with intrinsic yield, primarily serves as collateral. It is unsuitable to apply an interest rate curve for sDAI similar to other stablecoins on Aave.

Therefore, we propose a “flat” interest rate curve with a base rate of 1%, a slope1 of 1.5%, and a slope2 of 30%. The optimal utilization ratio should be set at 90%, with a reserve factor (RF) of 80%.

These parameters will not significantly contribute to sDAI’s yield. However, as sDAI is inherently a yield-generating collateral, its primary use case is to serve as collateral for borrowing other assets.

Specification

The proposal includes the following changes:

Specification

  • Remove GNO from Isolation mode

Lower Reserve Factor for EURe

Asset Chain Current Reserve Factor Rec. Reserve Factor
EURe Gnosis 20% 10%

Create sDAI/EURe E-mode

Parameter Value Value
Asset sDAI EURe
Collateral Yes No
Borrowable No Yes
Max LTV 85% 85%
Liquidation Threshold 87.5% 87.5%
Liquidation Bonus 5% 5%

this will be implemented in a second AIP

Parameter Value
Asset osGNO
Isolation Mode N/A
Borrowable No
Collateral Enabled Yes
Supply Cap 4,750
Borrow Cap -
Debt Ceiling -
LTV 0.05%
LT 0.10%
Liquidation Bonus 7.5%
Liquidation Protocol Fee 10%
Variable Base -
Variable Slope1 -
Variable Slope2 -
Uoptimal -
Reserve Factor -
Stable Borrowing Disabled
Flashloanable Yes
Siloed Borrowing No
Borrowable in Isolation No
Parameter Value Value
Asset osGNO GNO
Collateral Yes No
Borrowable No Yes
Max LTV 90% -
Liquidation Threshold 92.5% -
Liquidation Bonus 2.50% -

Onboard osGNO and create osGNO/GNO E-mode

  • List osGNO as a new non-borrowable asset
  • Create dedicated E-mode category for GNO-correlated assets
  • Proposed E-mode LTV: 90%
  • Proposed E-mode Liquidation Threshold: 92.5%
  • Proposed E-mode Liquidation Penalty: 2.5%

Implementation

These changes will be implemented through a payload executing the following actions:

AIP part One:

  • Configuration updates for GNO isolation mode
  • Reserve factor adjustment for EURe

AIP PArt Two:

  • Asset listing for osGNO
  • E-mode configuration for both new categories

Next Steps

  1. Publication of a standard ARFC, collect community & service providers feedback before escalating proposal to ARFC snapshot stage.
  2. If the ARFC snapshot outcome is YAE, publish the necessary AIPs for final confirmation and enforcement of the proposal.

Copyright

Copyright and related rights waived under CC0

6 Likes

Overview

A review of proposed changes for the Aave V3 Gnosis instance.

Onboard osGNO and create osGNO/GNO E-Mode

osGNO Analysis

osGNO is a GNO LST created by Stakewise. The token represents GNO staked in “Vaults” and earns yield from validators. osGNO can be permissionlessly minted against any Gnosis Chain node. Unlike traditional LSTs, whereby the associated receipt token represents the pooled underlying at a 1:1 rate, minting osGNO requires a user to be overcollateralized in staked GNO with respect to his osGNO, at a ltvPercent ”minting threshold” of 90%. The rationale behind this is to leverage the excess backing of staked GNO to protect holders against slashing and poor performance risks from permissionless nodes. Effectively, this implies that the associated stakers in Vaults are exposed only to the performance of their respective Vault’s validators, while external holders of osGNO benefit from this GNO-denominated slashing loss buffer. To this effect, while stakers can only mint osGNO worth up to 90% of their staked GNO, they still retain rewards on 100% of their stake.

Protocol Incentives

In addition to the autonomy presented at the node operator/vault level, the protocol effectively creates and maintains a competitive marketplace between respective vault operators through economic (dis)incentive alignment at the protocol level. In addition to siloed slashing risks within respective vaults, the protocol enshrines a “fair exchange rate,” as referenced in the PriceFeed contract, which represents the weighted average APY across all vaults with feePercent ≤ 15%, acting as the effective reward baseline for all osGNO in circulation in vault-standard fashion. In the event that a user delegates to underperforming vaults, stemming from operator MEV extraction, substantial operator fees or suboptimal infra, the underlying LTV of the associated osETH vault positions can scale accordingly, as the effective number of shares collateralizing the underlying staked GNO is no longer in line with the protocol.

External Redemptions

If the value of a user’s osGNO position reaches 91.5% of the staked GNO value, any osGNO holder can redeem for GNO held in the vault, totally ordered with respect to the LTV of the position. The amount redeemed is calculated using the following equation.

image - 2024-12-23T173932.392

These third-party redemptions can only be conducted to bring a user’s health level back to 90%, minimizing the amount that stakers can lose, and are only profitable to perform under the assumption that the market price of osGNO is trading lower than the associated exchange rate, thereby implicitly aligning incentives to maintain protocol safety while explicitly contributing to peg reversion.

Liquidations

Furthermore, if the value of an osGNO position deviates further, past the enshrined liqThresholdPercent of 92% relative to the staked GNO, maintaining protocol solvency is no longer conditionally performed as a function of the negative market price deviation from the PriceFeed value, rather the position can be fully liquidated with a liqBonusPercent of 1% to the external actor.

Withdrawals

Burning osGNO and withdrawing staked GNO follows the underlying dynamics of the Gnosis exit queue, which, with a CHURN_LIMIT_QUOTIENT value of 2^12 as opposed to Ethereum’s 2^16 and 3.5x fewer validators, allows users to receive the underlying GNO exponentially faster than staked ETH, typically occurring in 24 hours.

Market Price Deviations

As a result of the mechanisms at play, we have observed periods where the asset was trading at a large premium relative to its exchange rate. In this sense, there exists an associated deterrent in capital efficiency such that arbitraging the upward deviation requires a user to adhere to the ltvPercent, capping the effective maximum upward deviation at ~1.11 osGNO/GNO, nearly akin to a CDP stablecoin such as LUSD. This effect, coupled with the relatively quick exit queue and redemption/liquidation mechanisms, makes it such that negative deviations will be short-lived, while positive deviations solely benefit the protocol.

The associated market dynamics overlayed with mint/burn events further exemplify this. Significant upward deviations lead to an implicit incentive to mint osGNO, as the effective APY realized will thus be higher. While an upward market price deviation can deter Stakewise redemptions and liquidations from occurring, the net accrual rate of the LTV is significantly reduced given by the rate difference between the lower-performing vault positions and the weighted average staking APY, thus we expect the market price to revert into a profitably liquidatable range within such a long timeframe. Moreover, as the LTV converges upward, the associated collateralizing stake relative to the total osGNO in circulation would thus be deteriorating, thereby resulting in the expected market price reacting accordingly.

Market Cap and Liquidity

The asset currently has a total supply of 18,804 osGNO ($5.2M) and one significant DEX liquidity pool — the osGNO-GNO pool on Balancer. While the total supply has fallen since mid-November, it is likely that an Aave listing would induce demand and cause growth in its supply.

LT, LTV, and LB

Given the asset’s somewhat limited liquidity on Gnosis, as well as its anticipated use case in leveraged yield generation strategies, we propose setting the asset’s non-E-Mode parameters such that the asset will not be used to borrow uncorrelated assets (similar to rsETH’s listing on Ethereum). In E-Mode, we concur with the proposed parameters of 90%, 92.5%, and 2.5% for LTV, LT, and Liquidation Penalty, respectively.

Given the overcollateralized nature of osGNO, which enables upward deviations under a 1.11 GNO market price, liquidating osGNO/GNO positions on Aave becomes increasingly profitable during events of upward deviations. Hence, we consider the proposed 2.5% Liquidation Bonus to be sufficient to maintain a competitive and profitable liquidation market.

Supply and Borrow Caps

We recommend setting the supply cap at 1x the liquidity available beneath the Liquidity Penalty of 2.5%. This leads to a recommendation of 4,750 osGNO. Given the novelty of the asset and the high upside deviation demonstrated, we do not recommend setting the supply cap at 2x as the Chaos Labs methodology suggests. However, given liquidations are unlikely given the setup, we are prepared to facilitate supply cap increases.

Oracle

We recommend using the GNO/USD oracle augmented with the exchange rate contained in Stakewise’s PriceFeed contract.

Specification

Parameter Value
Asset osGNO
Isolation Mode N/A
Borrowable No
Collateral Enabled Yes
Supply Cap 4,750
Borrow Cap -
Debt Ceiling -
LTV 0.05%
LT 0.10%
Liquidation Bonus 7.5%
Liquidation Protocol Fee 10%
Variable Base -
Variable Slope1 -
Variable Slope2 -
Uoptimal -
Reserve Factor -
Stable Borrowing Disabled
Flashloanable Yes
Siloed Borrowing No
Borrowable in Isolation No
Parameter Value Value
Asset osGNO GNO
Collateral Yes No
Borrowable No Yes
Max LTV 90% -
Liquidation Threshold 92.5% -
Liquidation Bonus 2.50% -

CAPO Configuration

Following the methodology in this post, we recommend using a 14-day MINIMUM_SNAPSHOT_DELAY and a maxYearlyRatioGrowthPercent of 13.01%.

maxYearlyRatioGrowthPercent ratioReferenceTime MINIMUM_SNAPSHOT_DELAY
13.01% monthly 14 days

Remove GNO from Isolation Mode

GNO is currently listed in isolation mode with a debt ceiling of $2M, 1.3M of which is utilized. It has an LTV of 48% and LT of 53%. At current price and supply levels, this would increase the borrowing power of the market from $2M to just over $10M. The primary considerations when implementing this change are volatility and liquidity. Regarding the former, the asset has been moderately volatile, with a 30-day daily annualized volatility of 80.46%, up slightly from 77.94% over the last 180 days. This volatility level does not present a barrier to removing the asset from isolation mode.

The asset’s liquidity on Gnosis has grown since September and only dropped after the recent price drop.

Given the significant growth of the asset’s liquidity and the major GNO suppliers maintaining positions with a health score of 1.3 or higher, the market’s liquidation risk is greatly reduced. Together, these four top positions represent nearly 92% of supply while posing little threat of liquidation in their current composition. Given these factors, we approve of removing GNO from isolation mode.

Specification

Remove GNO from isolation mode.

Lower Reserve Factor for EURe

EURe has demonstrated stable liquidity on Gnosis in recent months.

Additionally, EURe’s volatility relative to EUR has remained stable at 10.95%.

These factors indicate that the asset’s risk profile is sufficient to reduce its RF in line with other stablecoins. While doing so is expected to reduce Aave DAO revenue from the asset by roughly $35K/year given current usage, this adjustment can be effective in driving additional EURe supply thanks to the increase in supply rate.

Specification

Asset Chain Current Reserve Factor Rec. Reserve Factor
EURe Gnosis 20% 10%

Make sDAI Borrowable

As yield-bearing assets such as sDAI generally generate limited borrow demand and minimal revenue, while there isn’t a significant risk in making the asset borrowable, we do not recommend making sDAI borrowable.

Additionally, given the high concentration of supply positions within the sDAI market, enabling the asset as collateral would only be possible by maintaining a conservative UOptimal parameter.

We will continue to review the market and will adjust our recommendation when a valid usecase for borrowing sDAI will be present.

sDAI/EURe E-Mode

As discussed in a previous section, EURe has maintained a somewhat stable peg and liquidity, making it suitable for inclusion in E-Mode, albeit with more conservative parameters than those on other deployments, given its lower correlation to USD-denominated stablecoins.

While both sDAI and EURe have volatile liquidity at the 1% slippage level, their liquidity below the 5% slippage level has remained relatively stable over the past three months.

With all these factors in mind, we concur with the proposed parameters for the E-Mode, finding that the 5% Liquidation Penalty is sufficient to properly incentivize liquidators, and the 87.5% LT point provides a sufficient buffer from the bad debt accrual point at 95.2%.

Specification

Parameter Value Value
Asset sDAI EURe
Collateral Yes No
Borrowable No Yes
Max LTV 85% 85%
Liquidation Threshold 87.5% 87.5%
Liquidation Bonus 5% 5%

Disclaimer

Chaos Labs has not been compensated by any third party for publishing this ARFC.

Copyright

Copyright and related rights waived via CC0

2 Likes

Summary

LlamaRisk supports this proposal and endorses the detailed analysis conducted by @ChaosLabs. We align with the proposed parameters based on their rationale and our independent research. We want to highlight a few additional considerations for the DAO:

osGNO

The majority of liquidity in the Balancer pool was contributed by a single address approximately 90 days ago, creating significant liquidity concentration risk. This presents a critical point of failure - if this major liquidity provider withdraws, it could severely destabilize the pool. While osGNO’s underlying collateral redemption mechanism provides some protection, Aave liquidators may still struggle to execute profitable liquidations efficiently.

The Balancer pool’s asymmetric composition (66% osGNO / 33% GNO) further compounds these risks, as the path to exit could become particularly challenging during periods of market volatility.

Given the asset’s thin liquidity, the highly limited LTV and LT parameters for non-E-mode osGNO are prudent. This conservative approach reduces bad debt risk by restricting leveraged positions to GNO borrowing only, capitalizing on the assets’ inherent price correlation.

GNO


Source: GNO supplier market positions, Chaos Labs Community Analytics, 23 December 2024

LlamaRisk supports removing GNO from isolation mode with the recommendation that supply caps are not increased under current market conditions. The asset on Aave is primarily used to collateralize USDC.e loans that are unlikely to be liquidated. To allow a greater blend of assets will not only increase GNO’s attractiveness as collateral but may also facilitate the Gnosis Credit Line proposal put forward recently.


Source: GNO supplier market wallet health, Chaos Labs Community Analytics, 23 December 2024

The primary concern for this proposal centers on GNO’s limited liquidity. A relatively modest swap of $1.1M (~4500 GNO) causes a 7% price impact. With the removal of isolation mode, the protocol may attract additional GNO deposits given the ~20K supply cap room. At full capacity, total GNO TVL would reach ~$23M, resulting in LTV-adjusted debt potentially exceeding 10x available liquidity. This creates significant risk during large-scale liquidations, particularly since many of the largest GNO positions on the network are non-liquidatable. Given these market conditions and potential bad debt, we strongly recommend maintaining the current supply cap without any increases.

sDAI

Making sDAI borrowable introduces significant risks. Primarily, sDAI’s inherent APY can be modified at short notice by Sky contributors. Users may face unexpected interest rate changes without adequate preparation, creating an operational burden for the DAO to maintain aligned IRM parameters.

This risk is amplified on the Gnosis Chain, where staked DAI yields are higher than native due to an additional layer of yield redirection. Almost all xDAI represents DAI staked on Ethereum and bridged to Gnosis. Users may experience rapid changes in position health and face potential liquidation if they fail to manage their collateral and debt positions actively.

Current market conditions suggest a limited initial borrow cap of $2M, indicating insufficient financial infrastructure for profitable shorting opportunities. Given these constraints and the lack of clear benefits, we cannot support onboarding sDAI as a borrowable asset without a more compelling strategic rationale.

Disclaimer

This review was independently prepared by LlamaRisk, a community-led non-profit decentralized organization funded in part by the Aave DAO. LlamaRisk is not directly affiliated with the protocol(s) reviewed in this assessment and did not receive any compensation from the protocol(s) or their affiliated entities for this work.

The information provided should not be construed as legal, financial, tax, or professional advice.

The current proposal has been escalated to ARFC Snapshot, including latest updates from Risk Service Providers.

Vote will start tomorrow, we encourage everyone to participate.

After Snapshot monitoring, the current ARFC Snapshot ended recently, reaching both Quorum and YAE as winning option with 859,2K votes.

Therefore [ARFC] Aave v3 Gnosis Instance Updates has PASSED.

Next step will be the publication of an AIP for final confirmation and enforcement of the proposal.

After discussing the implementation with @bgdlabs, the OsGNO onboarding will be handled in a separate AIP, pending an asset technical review.

Since all other elements of the current AIP have already been approved at the Snapshot stage, we will proceed to the AIP phase without the OsGNO portion. A second AIP will be published once we receive the green light for OsGNO.

2 Likes

Quick comment on why I believe it is correct to make sDAI borrowable.

On Gnosis there should be no reason to deposit xDAI into Aave. sDAI does not introduce any additional risk - so any rational actor would rather deposit sDAI than xDAI. Currently the only meaningful deposits for xDAI come from GnosisDAO - essentially subsidizing the ability to borrow stables cheaper.


Therefore I believe the correct setup would be to off-board xDAI and fully replace it with sDAI - ideally xDAI could remain as a UI option but under the hood all xDAI deposits would be turned into sDAI. In this setup however I think it is important that sDAI can be borrowed.

1 Like

This topic was automatically closed 30 days after the last reply. New replies are no longer allowed.