[ARFC] Aave Umbrella - activation

[LlamaRisk]

LlamaRisk’s Umbrella Methodology

As the activation of Umbrella draws near, we aim to provide the Aave DAO with insights into how we will recommend parameters and manage risk in collaboration with other members of the Aave Finance Committee (AFC). At the core of this effort is LlamaRisk’s Umbrella methodology, a quantitative framework specifically designed to determine appropriate Umbrella Fund Caps for assets covered by the mechanism.

This document highlights the key components of LlamaRisk’s Umbrella methodology, including data inputs, analytical processes, outputs, and governance considerations. By leveraging data-driven analysis and scenario-based simulations, our methodology ensures that the protocol is adequately capitalized to withstand predefined market stress scenarios. These recommendations aim to enhance Aave’s risk management, optimize capital efficiency, and support the protocol’s sustainable growth.

Background

Large-scale liquidations pose systemic risks to lending protocols. Insufficient market liquidity can leave undercollateralized debt, leading to bad debt and potential insolvency. To mitigate this, protocols use Umbrella Funds (or Safety Modules) as backstops for shortfalls from failed liquidations. These funds must balance being large enough to handle crises and avoiding inefficient capital allocation.

Aave enhances this approach with Umbrella, a per-asset Umbrella Fund that builds on stkAAVE and stkGHO Safety Modules. The Umbrella methodology ensures Aave can withstand predefined market stress (e.g., bad debt equal to 10% of an asset’s borrow cap) by analyzing liquidity, user positions, and price shocks. It recommends fund caps that ensure safety while minimizing unnecessary capital lock-up.

Methodology Inputs

The accuracy and robustness of the Umbrella methodology depend on the quality and comprehensiveness of input data captured at a specific timestamp. The methodology utilizes two main categories of data:

1. Protocol State Data:

• Detailed parameters for every supported asset (e.g., current market prices, supply and borrow caps, Loan-to-Value (LTV) ratios, Liquidation Thresholds, Liquidation Bonuses, and E-Mode category inclusion).

• A comprehensive snapshot of every user’s account, capturing their collateral and debt balances and E-Mode status.

2. External Market Data:

• Historical price time series for all relevant assets, critical for modeling potential future price shocks using techniques like Value-at-Risk (VaR) and correlation analysis.

• Data on market liquidity and depth for relevant major decentralized exchanges (DEXs) trading pairs.

Methodology Flow and Outputs

The Umbrella simulation follows a multi-stage process:

1. Data Ingestion & Preparation: Collect and consolidate all the necessary Protocol State and Market Data for the chosen timestamp. Filter user positions based on criteria like minimum debt size and potentially filter by E-Mode category if analyzing a specific E-Mode.

Figure: Example state of an asset and the market parameters on Aave

2. Synthetic Price Shock Generation: Based on historical price data analysis, calculating daily returns, identifying stressed regimes, computing VaR99 and inter-asset correlations during drawdowns, generating many distinct, synthetic price shock scenarios. Each scenario represents a potential future market state, assigning a specific negative price change (e.g., -5%, -10%) to each asset within Aave’s Core market, reflecting correlated downturns.

Figure: Example of generated price shock samples

3. Liquidation Simulation (Iterated per Shock Scenario):

• Select one shock scenario and apply the specified price changes to the current market prices of all assets.

• Re-evaluate the Health Factor of all user positions using the shocked asset prices and the protocol’s Liquidation Thresholds (or E-Mode specific thresholds). Identify all positions falling below the liquidation threshold (HF < 1).

• For the specific asset being analyzed, sum up the total debt owed across all identified liquidatable positions. Also, determine the composition and total amount of all collateral assets backing this specific debt.

• Model the process of liquidating the aggregated collateral. This involves simulating swaps through the relevant liquidity pools, accounting for the price impact caused by the liquidation volume. The simulation finds the Optimal Scale factor: the maximum fraction of the total aggregated liquidatable debt (for this asset, in this scenario) that rational liquidators can liquidate profitably, considering the Liquidation Bonus and the simulated price impact. This is achieved using iterative search algorithms to pinpoint the profitability threshold.

• Store the calculated Total Liquidatable Debt and the corresponding Optimal Scale factor for the target asset resulting from each simulated shock scenario.

4. Outlier Filtration: Analyze the distribution of Total Liquidatable Debt and Optimal Scale values across all scenarios. Apply a statistical outlier detection using predefined multipliers. Scenarios falling outside these calculated bounds are flagged as outliers and removed from the subsequent calculation to prevent extreme, potentially unrealistic scenarios from unduly influencing the outcome.

Figure: Example of detected outlier samples

5. Market Liquidation Capacity Calculation: Using only the filtered inlier data points, calculate the mean of the Total Liquidatable Debt values and the mean of the Optimal Scale values. The Market Liquidation Capacity is then computed as:

This value represents the methodology’s estimate of how much debt (in the specific asset) the market can reliably absorb through liquidations under the stress conditions defined by the inlier shock scenarios without causing losses to liquidators.

6. Umbrella Fund Requirement Calculation: Compare the calculated Market Liquidation Capacity against the predefined Safety Target parameter, defining a Safety Threshold over a total Borrow Cap of an asset.

The final Umbrella fund size is then determined by the gap between the Safety Target and Market Liquidation Capacity:

Governance Parameter Management

The Umbrella methodology provides a structured computational framework that needs to align with the protocol’s overall strategy and risk appetite. The Safety Target is the primary parameter for governance interaction, defined as the percentage of an asset’s Debt Cap that the protocol aims to protect through market liquidity and the Umbrella Fund. The percentage of Aave revenue allocated to Umbrella is also a key parameter that will need to be decided by governance, representing another important lever for calibrating the Umbrella mechanism to the protocol’s risk management objectives and financial sustainability goals.

The flexibility of adjusting the Safety Target allows the Aave DAO and the AFC to align Umbrella Fund levels with various considerations. For instance, the inherent risk profile of an asset, its specific borrow concentration patterns, prevailing market conditions, and broader strategic growth initiatives involving risk parameter changes (like LTV adjustments) can all inform the appropriate level for the Safety Target. This adaptability ensures that Umbrella levels reflect both current risks and future ambitions.

LlamaRisk initially recommends starting with moderate Safety Thresholds across major assets. As the DAO proposes and implements strategic market risk adjustments, the Safety Target for affected assets can be reviewed and modified accordingly by the AFC. This creates a direct, dynamic link between risk management and protocol growth objectives. Adjusting the Safety Target becomes a collaborative process, ensuring that the Umbrella methodology’s Umbrella Fund Cap recommendations remain aligned with the evolving strategic direction and risk tolerance of the Aave protocol. Other methodological parameters are generally considered more stable but remain transparent and subject to periodic review.

Initial Scope: Assets and E-Modes

The Umbrella methodology, particularly in its initial implementation, primarily aims to analyze and recommend Umbrella Fund Caps for the most systemically important assets within the protocol. This includes major stablecoins like USDC, USDT, and DAI, constituting a large portion of borrowing activity. It also covers large-cap volatile assets such as WETH (and its staked derivatives like wstETH) and WBTC, representing significant collateral value and borrowing demand.

The methodology is designed with flexibility regarding E-Modes. The structure allows for running the analysis specifically for assets within a designated E-Mode, using the appropriate E-Mode-specific Liquidation Threshold. The goal remains consistent: ensuring sufficient backing (market liquidity + Umbrella) to handle liquidations up to the Safety Target within the specific risk parameters of that asset or E-Mode category.

[ChaosLabs]

Summary

Chaos Labs fully supports the integration of Umbrella within Aave’s core protocol architecture. Embedding Umbrella as a native mechanism will facilitate the prompt resolution of insolvent positions through a dynamically risk-weighted junior tranche structure, enabling efficient absorption and socialization of bad debt under adverse market conditions.

We intend to collaborate closely with the Aave Finance Committee to deliver ongoing, data-driven recommendations for fine-tuning Umbrella’s parameters. This includes adaptive calibration of tranche thresholds, emission schedules, and risk mitigation strategies to ensure resilient performance across a range of market states and stress scenarios. Below, we offer a preliminary breakdown and key insights ahead of the release of our full quantitative methodology in the coming days.

Umbrella Cost Dynamics

Currently, Aave distributes substantial incentives through stkAAVE, which represents a significant ongoing expense. Under Umbrella, incentive and cost dynamics have various implications that can be considered when optimizing markets and quantifying parameters and expenses.

One key dynamic arises from Umbrella’s design, which prevents staked aTokens from being used as collateral. As a result, staking into Umbrella is inherently more attractive for assets with low collateral demand—such as stablecoins—where the opportunity cost of forfeiting rehypothecation is minimal. This typically aligns with assets that have high borrow demand, reinforcing the need for robust Umbrella coverage.

However, this relationship does not always hold. Assets like WETH, for example, exhibit both strong borrow demand and significant collateral usage. In such cases, Umbrella introduces a higher opportunity cost for participants, as staking requires foregoing valuable collateral utility. To attract sufficient liquidity under these conditions, the yield offered through Umbrella staking must be correspondingly higher to offset this trade-off.

WETH Borrow and Collateral Distribution

The financial sustainability of Umbrella is also partially correlated with the borrowing intensity of supported assets and, as such, with their necessity for Umbrella protection. This dynamic is caused by the direct correlation between borrowing demand and revenue derived by SVR and the Reserve Factor.

However, some assets, combining low interest rates and revenue with high market utilization, such as wstETH being the primary borrow option against LRTs, amplify costs disproportionately, creating a significant opportunity cost for the DAO to incentivize such markets while accruing limited revenue from them. This highlights the importance of onboarding assets and strategies with a balanced or positive risk-reward for the DAO in order to contain Umbrella expenses.

While initially, Umbrella is planned to be funded purely through AAVE incentives, the emission curve is likely to shift to a structure whereby underlying revenues from SVR or Reserve Factor of the interested pool are reinvested into the market for incentives. This reallocation incentivizes sustainable growth and improved system resilience and allows for reduced spending from the current security budget, amounting to over $55M annually.

If necessary, a long-term strategy to balance incentives can stem from increasing the Reserve Factor to distribute to Umbrella Stakers. However, this strategy will reduce revenue directed at vanilla suppliers and thus is highly reserve-dependent, taking into account the aforementioned components below.

Parameterization of maxEmissionPerYear and Target Liquidity

The derivation of maxEmissionPerYear—understood as the optimal incremental yield offered to users staking into Umbrella—is informed by a multifaceted opportunity cost analysis. This analysis captures both user-level trade-offs and systemic risk dimensions inherent to the composable architecture of the protocol. The following considerations form the foundation of its calibration:

Collateral Utility Opportunity Cost

Staking tokens in Umbrella precludes their use as collateral, in contrast to traditional aToken holdings that can be rehypothecated. The opportunity cost is therefore linked to the share of aTokens typically used as collateral within a reserve. This relative collateral utilization reflects the user trade-off between earning additional Umbrella rewards and preserving the capital efficiency of leveraging their assets elsewhere.

Native Reserve Contribution to External Reserve Risk

This component accounts for systemic risk arising from the reserve’s exposure to overcollateralized debt positions, quantified through the nominal health factors of users across the reserve. When aggregate health factors are low and a large portion of the reserve is actively used as collateral, the protocol faces elevated liquidation risk.

Given the composable and rehypothecative structure of the system—where assets often serve dual roles as both collateral and supplied liquidity—a risky distribution of such assets can trigger cascading liquidations. If collateralized positions are forcefully unwound, corresponding aTokens are removed from the reserve, driving up utilization rates. In extreme scenarios, this dynamic can impair the protocol’s ability to effectively liquidate risk and maintain solvency.

By increasing incentives for users to stake into Umbrella under these conditions, the protocol discourages excessive relative collateral usage, reinforces reserve stability, and reduces the likelihood of systemic stress propagation.

Market Demand Reflected in Interest Rate Regimes

Since maxEmissionPerYear is expressed in raw token terms, it must be responsive to the prevailing interest rate environment, which serves as a proxy for market-wide demand for leverage. As rates rise, signaling intensified competition for capital and heightened returns on leveraged positions, the opportunity cost of supplying to Umbrella significantly increases.

In such high-rate regimes, capital agility becomes more valuable, and the trade-off of forfeiting collateral utility grows steeper. To remain attractive under these conditions, Umbrella’s reward emissions must scale accordingly, compensating users for the elevated marginal cost of locking capital in a less flexible position. This ensures that staking into Umbrella remains a rational choice even when market incentives favor rehypothecation and aggressive leverage.

Duration and Cooldown Risk

Although aTokens continue to accrue yield and remain slashable during the cooldown period, the temporal delay before assets become accessible again introduces duration risk. This immobility hampers a user’s ability to respond promptly to market shifts, reinforcing the need for a compensatory yield uplift. The longer the effective lockup, the higher the premium required to offset the strategic disadvantage of delayed capital deployment.

Target Liquidity

Target Liquidity aims to be defined by leveraging a VaR framework, applied to the distribution of underlying collateral that supports debt within the reserve. This measure seeks to estimate the required coverage to safeguard the protocol in the face of adverse market scenarios. Given the inherent composability and interdependencies within Aave, determining this value necessitates a simulation-driven, quantitative approach that accounts for the system’s complex dynamics.

This process ultimately frames the derivation of the Target Liquidity parameter as a minimization problem—where the goal is to identify the optimal nominal stkToken supply required to mitigate tail risks while respecting capital efficiency. In an ideal state, the Aave Finance Committee would have full flexibility to allocate as many resources as needed toward Umbrella coverage. However, as the reserve grows, the optimal share of Umbrella coverage is not expected to scale linearly. Instead, it is likely to follow a logarithmic growth pattern, reflecting diminishing marginal risk exposure relative to reserve size.

In our forthcoming post outlining the complete umbrella reserve methodology, we will detail the full mathematical derivation and supporting simulations.

Cooldown Period Parameterization

A 20-day cooldown and 2-day withdrawal window align well with existing Safety Module practices. However, this introduces some risks tied to oracle responsiveness—delays or inaccuracies in fundamental oracles can allow stakers early exits before slashing events are adequately reflected in the oracles and bad debt is accrued. An example of such a situation is with Lido bunker mode activation and correlation penalty, which only impacts the oracle price 18 days following the initial slashing. A further example lies in assets that require manual oracle updates through the governance process to reflect their underlying value following a backing loss.

Careful selection of cooldown and withdrawal window for these specific assets may be required in order to ensure Umbrella Liquidity at the time of oracle update.

Application of Automation

Automation can play a pivotal role in Umbrella by enabling dynamic parameter adjustments across multiple components of the system. In addition to the previously discussed elements—such as Target Liquidity and maxEmissionPerYear, which are influenced by Value-at-Risk (VaR) assessments and reserve opportunity cost—automation can also govern the initial denomination of rewards in AAVE tokens, given by the mismatch between reward token denomination and the underlying reserve denomination.

Given that the underlying objective function is expressed in raw token terms, and that market-specific nominal risk scales with reserve growth, it can be beneficial to implement a system that can continuously recalibrate these values in real time. This ensures reward emissions remain aligned with both market dynamics and protocol-level risk exposure.

Disclaimer

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

Copyright

Copyright and related rights waived via CC0

[Kene_StableLab]

@ChaosLabs

While initially, Umbrella is planned to be funded purely through AAVE incentives, the emission curve is likely to shift to a structure whereby underlying revenues from SVR or Reserve Factor of the interested pool are reinvested into the market for incentives.

We hope that as SVR revenue continues to grow, we will be able to repurpose those funds for incentives, if this works out, it will create an excellent flywheel mechanism.

[bgdlabs]

Umbrella activation parameters

After receiving feedback from and discussing this with other service providers, we think proceeding with a progressive activation of Aave Umbrella is reasonable.

Umbrella initial parameters

To start with, we recommend activating only a subset of assets in Aave v3 Ethereum Core in Umbrella to simplify this initial Phase, and promptly expand with other AIP to more assets and pools/networks.

The initial configuration we propose regarding Aave v3 Ethereum Core is the following.

Important. Aside from proposed configurations (assets, target liquidity, cooldown/unstake window, offset), both on the proposed Umbrella parameters and legacy, some variables need to be assumed for modelling, like the AAVE or ETH prices, average supply on Aave, or expected staked assets on legacy Umbrella, amongst others.
Our objective is to showcase a reasonable model that later on will materialise with more/less deviation on the production system.

Staked asset	Covered asset	Target Liquidity *	Max emission (rewards at target liquidity)	Umbrella APY range (up until excess liquidity) **	Total APY (Aave + Umbrella)***	Cooldown/unstake window	Deficit offset
aUSDC (wrapped)	USDC	85’000’000 USDC	3’000’000 aUSDC/year	1.76%-7.06%	6.8%-12.5%	20/2 days	100’000 aUSDC
aUSDT (wrapped)	USDT	85’000’000 USDT	3’000’000 aUSDT/year	1.76%-7.06%	6.8%-12.5%	20/2 days	100’000 aUSDt
aWETH (wrapped)	WETH	25’000 ETH	550 aWETH/year	1.1%-4.4%	3%-6.3%	20/2 days	50 aWETH
GHO	GHO	12’000’000 GHO	1’200’000 aGHO/year	5%-20%	5%-20%	20/2 days	100’000 GHO

*Target Liquidity is denominated in the contracts in wrapped aTokens, increasing over time in exchange rate. That means the Target Liquidity itself will grow slightly over time and rewards will need to be periodically adjusted.
For the sake of simplicity, the number on the table is in equivalent terms of underlying (USDC, USDT, WETH), not in wrapped aTokens.
** Umbrella has an upper limit of APY as max emission is capped, but technically no lower limit. However, going over the table’s lower point of APY would mean there are way more deposits than expected, hence the market pricing “cheaper” the risk of staking.
*** 1y average supply rates for each asset on Aave are taken as reference.

In addition, on the Umbrella side is important to highlight the following items on the activation proposal:

• Set as Rewards Admin the Aave Finance Committee, with a timelock of 1 day. This will be done by introducing a new Permissioned Payloads Controller, mirroring the architecture of the execution layer of the Aave Governance, but oriented to less critical flows like rewards updates.
• Set the Aave Ethereum Collector as the reward payer: the address from which rewards will be distributed. The configured allowance will be 50% of the defined yearly budget, which will allow enough but controlled flexibility for the Finance Committee to increase rewards during the initial month.
  At any point, the Aave governance can modify this as required.
• All outstanding deficits on assets to be covered by the initial Stake tokens will be “cleaned” with Collector funds on the AIP, up to a limit of ~$2’000 on each asset. The current aggregated deficit is up to ~$600 in USDC, USDT, WETH, and GHO, so the $2’000 margin seems more than enough.
• A DeficitOffsetClinicSteward smart contract will be given an allowance from the Aave Collector up to the Deficit Offset configured for each asset. This will allow the Aave Finance Committee to trigger deficit offset coverage without going through a governance proposal.
• The Deficit Offsets proposed are substantially above the bad debt levels historically accrued on Aave v3 Core on the covered assets. That means said offsets act as a very substantial “tranche” that should protect stakers from slashing under normal conditions.
• Stake Umbrella tokens, as by design, will cover exclusively the deficit created on the associated borrowed assets on Aave v3 Core Ethereum: Staked aUSDC will cover exclusively the deficit on USDC, staked GHO exclusively the deficit on GHO, and so on.

Legacy Safety Module progressive deprecation

Together with the activation of the initial Umbrella instances, the AIP will also start a progressive deprecation of the legacy Safety Module to transition to Umbrella. The principles of this progressive deprecation are:

• Global coverage of Aave pool and instances should remain at similar levels, including Target Liquidity of Umbrella + broad slashable funds on the Legacy Safety Module.
• Coverage infrastructure should be fully migrated in those cases where Umbrella supports. E.g., legacy stkGHO on what concerns coverage will be fully deprecated in favour of Umbrella’s stkGHO.

Legacy Safety Module configuration (as of 15/05/2025)

Staked asset	Rewards/day (AAVE)	Rewards/year (AAVE)	Rewards/year ($)	Average total staked ($)	Slashing eligibility	Total Slashable ($)	Swap effectivity (inverse slippage)	Effective coverage ($)
AAVE	360 AAVE	131’400 AAVE	21’000’000	500’000’000	30%	150’000’000	75%	112’500’000
AAVE/wstETH Balancer v2	240 AAVE	87’600 AAVE	14’000’000	180’000’000	30%	54’000’000	80%	43’200’000
GHO	100 AAVE	36’500 AAVE	5’800’000	150’000’000	100%	150’000’000	100%	150’000’000

Proposed Legacy Safety Module post-Umbrella Phase 1 activation

Staked asset	Rewards/day (AAVE)	Rewards/year (AAVE)	Rewards/year ($)	Average total staked ($)	Slashing eligibility	Total Slashable ($)	Swap effectivity (inverse slippage)	Effective coverage ($)
AAVE	216 AAVE	78’840 AAVE	15’768’000	450’000’000	20%	90’000’000	85%	76’500’000
AAVE/wstETH Balancer v2	216 AAVE	78’840 AAVE	15’768’000	162’000’000	20%	32’400’000	90%	29’160’000
GHO	0	0	0	0	0	0	0	0

*Merit incentives will not be touched on the legacy stkGHO, and coverage incentives will be moved to the new stkGHO.

High-level implications of the activation

• Currently, on the AAVE/ABPT components, the legacy Safety Module has a total coverage of ~$204m (30% of ~$780m), but with very low capital efficiency from the stkAAVE/stkABPT side, making this number realistically closer to ~$155m. This coverage is heterogeneous in the case of AAVE and stkAAVE (it applies to multiple pools, to all assets on those pools).
  With the reductions of slashing eligibility and rewards on stkAAVE and stkABPT, we expect this number to still be in the order of ~$110m, which will remain as the heterogeneous coverage mechanism for assets not covered by the Umbrella day 0.

• On the legacy stkGHO size, the current coverage is very outsized compared with the realistic requirements. So the slashing removal from the legacy stkGHO will change the coverage of GHO to more sane levels compared with outstanding borrowings: from ~$150m average on legacy stkGHO to a target of $12m in Umbrella Staked GHO.

• On the legacy Safety Module, the reduction of rewards in percentage is higher than the reduction in slashing eligibility, because on low-level percentages of slashing, decreases have proportionally more weight.
  The end goal is to remove completely slashing from them, but given their current size, we think it is more responsible to do it progressively until Umbrella’s liquidity builds up in this initial Phase and follow-ups.

• Even considering the absolute value reduction of GHO coverage and reducing slashing percentage on legacy stkAAVE/stkABPT, the global coverage will be higher post-Umbrella:

  • Pre-Umbrella Phase 1: ~$305m.
  • Post-Umbrella Phase 1: ~$345m

• The comparison of rewards’ expenses pre- and post-Phase 1 Umbrella is as follows:

  • Pre-Umbrella Phase 1: ~$51m/year (in AAVE token).
  • Post-Umbrella Phase 1: ~$40m/year ($31.5m in AAVE token + $8.5m in aTokens).

  This shows major cost savings for more absolute coverage, and the number will substantially decrease as in the follow-up phases, as the legacy Safety Module continues its deprecation, realistically settling to levels of ~$15-30m, if assuming zero expenses of AAVE as rewards.

• Realistically, on follow-up proposals, the coverage of other networks will be approximately proportional to the pool’s size in comparison with Ethereum, and any ad-hoc consideration.

  This implies that expenses and coverage would be 10-20% more once the other pools and networks are added.

• By keeping Merit rewards, but removing slashability, legacy stkGHO will factually become a non-slashable staking contract, with a cooldown.

Next steps

After some days in this forum, we will proceed with creating an ARFC Snapshot for these final parameters, which will precede the AIP for the Aave governance to apply them in production.

[LlamaRisk]

Summary

LlamaRisk supports the proposed initial parameters and asset selection for the launch of Umbrella. Based on our Umbrella methodology simulations, these asset choices and the allocated budget are well-suited to provide optimized risk coverage for Aave’s Core market. The gradual deprecation of the legacy Safety Module, alongside Umbrella’s activation, will ensure that protocol coverage remains sufficient until Umbrella’s scope is expanded.

Additionally, we recommend a minor optimization to the Target Liquidity parameters for USDC and USDT. Our methodology indicates that loans backed by these two stablecoins exhibit different levels of market shock resilience. As such, there is an opportunity to maximize the marginal utility of each dollar within the Umbrella module by allocating more resources to USDT, which has a lower intrinsic liquidity capacity relative to its market size. The proposed adjustments are as follows:

• Decrease USDC Target Liquidity from $85M to $66M.
• Increase USDT Target Liquidity from $85M to $104M.
• Adjust reward emissions proportionally, keeping the same total maximal emission rate.

These changes would improve risk-adjusted coverage without increasing overall costs for the DAO. As market conditions and loan compositions on Aave evolve and Umbrella coverage expands, we will continue to recommend further adjustments as needed.

Evaluation of Risk Coverage

To evaluate the proposed parameters for the initial Umbrella activation, we applied our Umbrella-specific methodology based on the price shock simulations and market liquidity capacity estimations. The output measures bad debt coverage needs and helps us guide the risk-adjusted parameter recommendations on a per-asset basis.

The specific simulation run was executed on May 15th, 2025. A snapshot of asset prices and loan compositions on Aave was taken and historical price observations were taken into account for further synthetical price shock generation. Overall asset metrics were as follows:

Asset	Price	Supply	Borrowed	Borrow Cap
WETH	$2,590	2,473,578 WETH	2,139,972 WETH	2,700,000 WETH
USDC	$1	2,463,512,109 USDC	2,032,575,656 USDC	4,320,000,000 USDC
USDT	$1	3,464,547,040 USDT	3,178,590,933 USDT	4,720,000,000 USDT
GHO	$1	-	180,000,000 GHO	180,000,000 GHO

Liquidity Capacity and Debt Coverage

A core measure in this analysis is Liquidity Capacity, defined as the maximum amount of debt that can be liquidated profitably before price impact losses render further liquidations unfeasible. This capacity is crucial because the total liquidations a borrow asset can sustain with an active Umbrella coverage is conceptualized as the sum of the Target Liquidity and Liquidity Capacity.

The simulations employed Value at Risk (VaR) based shock scenarios to estimate Liquidity Capacity. These simulations estimated liquidatable debt and collateral under various stress conditions, scaling them to the point where liquidations cease to be profitable. The scaled liquidatable debt at this point is the Liquidity Capacity metric. The simulation run for the initial Umbrella asset on Core market provided the following measures:

Asset	Borrow Cap	Liquidation Capacity	Liquidation Capacity ($)	Mean Liquidatable Debt ($)
USDC	4,320,000,000 USDC	50,400,000 USDC	$50,400,000	$1,115,000
USDT	4,720,000,000 USDT	12,200,000 USDT	$12,200,000	$192,000
WETH	2,700,000 WETH	6,900 WETH	$17,870,000	$1,508,663
GHO	180,000,000 GHO	5,000,000 GHO	$5,000,000	$250,000

Market Shock Resilience

To evaluate the robustness of the Target Liquidity in Umbrella’s module adjusted for Liquidity Capacity, market price shock simulations were applied, and the Shock Intensity multiplier metric was acquired. The Shock Intensity metric represents a multiple of a VaR99 based price shock paths such that the Liquidatable Debt would equate to the sum of the Target Liquidity and Liquidity Capacity.

Source: LlamaRisk, May 15th, 2025

The resulting Shock Intensity multipliers differ based on the asset types, depending on asset price volatility and the collateral type distribution on Aave’s market. This leads to a large shock resilience for assets primarily used in E-Mode setups. In particular, WETH is largely resilient to market price shocks because most WETH borrows (90.3%) are used in ETH LST E-Mode. Consequently, in this simulation run, we omit the Shock Intensity evaluations for WETH, including USDT, USDC, and GHO stablecoins.

Source: LlamaRisk, May 15th, 2025

A key observation is an imbalance in the Shock Intensity resilience between USDC and USDT. With $85M projected Target Liquidity and ~$50.4M Liquidity Capacity, USDC demonstrated resilience up to a ~3.9x shock. Conversely, USDT, with an identical $85M projected Target Liquidity but only ~$12.2M Liquidity Capacity, showed resilience to a ~3.6x shock, at which point a large uptick in liquidatable debt is projected. This disparity suggests a potential inefficiency in the allocation of Umbrella funds relative to the underlying market’s inherent liquidation absorption capabilities.

Proposed Optimizations

Adjusting the Umbrella fund allocations between USDC and USDT would lead to a more balanced coverage of their respective borrow caps and enhance the protocol’s resilience to price shocks. By increasing the Target Liquidity for USDT and slightly reducing it for USDC, both assets would have similar total debt coverage in dollar terms, and their coverage as a percentage of borrow cap is more closely aligned. This optimization would ensure that the protocol can withstand market volatility better, particularly for the largest stablecoins, without increasing the Umbrella budget.

Asset	Target Liquidity	Total Debt Coverage	Coverage of Borrow Cap, %	Shock Intensity
USDC	$85,000,000 → $65,900,000	$135,400,000 → $116,300,000	3.13% → 2.69%	4.30 → 3.80
USDT	$85,000,000 → $104,100,000	$97,200,000 → $116,300,000	2.06% → 2.46%	3.59 → 3.61
WETH	$57,500,000 (unchanged)	$75,370,000 (unchanged)	1.18% (unchanged)	>10x
GHO	$12,000,000 (unchanged)	$17,000,000 (unchanged)	9.44% (unchanged)	4.40

Future Considerations

As the market landscape evolves and loan risk profiles shift, it will be essential to review Target Liquidity levels on an asset-by-asset basis regularly. Anticipated growth may also require more assertive risk parameters, so any such adjustments must be accompanied by a corresponding rebalancing of Umbrella’s Target Liquidity allocations.

The yield premium offered by Umbrella is expected to attract additional supply to Aave’s markets, thereby increasing liquidity available to borrowers. However, ensuring that the utilization rate never exceeds the supply of assets within Umbrella’s module remains critical so that assets remain available for slashing if necessary.

The Umbrella’s activation phase is expected to directly improve and optimize the risk profile of Aave’s Core market. In the subsequent stages, it will be crucial to further refine the parametrization and position Aave for sustainable growth. That will be a common effort and a priority of AFC’s members.

Disclaimer

This analysis was independently prepared by LlamaRisk, a community-led 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.

[TokenLogic]

Summary

TokenLogic supports the proposed parameters presented by @bgdlabs with one minor amendment that enables the pre-Umbrella stkGHO to transition into sGHO.

Here are the key changes this proposal aims to put forward:

• Umbrella Launch

  • Following @bgd recommendation, allocate 3M USDC, 3M USDT, 550 WETH, and 1.2M GHO to fund Umbrella.
  • Lower GHO emissions; emissions covered by existing DAO reserves.
  • Yield tied to Aave aToken performance for market-aligned returns.

• sGHO Transition

  • Remove Cooldown and Slashing from pre-Umbrella stkGHO.
  • Rebrand as sGHO with risk-free yield and improved UX.
  • Merit rewards to be distributed weekly; focus on boosting base yield, expected yield ~7.50%.

• stkABPT Adjustments

  • ~318k BPT withdraw-able in June; potential 50% TVL drop.
  • Recommend further reduction in AAVE emissions to target a 10-12% yield if exits occur.

• stkAAVE Strategy

  • Buybacks to slow in June after acquiring 34k+ AAVE.
  • stkAAVE emissions now funded by protocol revenue.
  • Anticipated yield reduction; potential rise of derivative strategies.
  • Recommended AAVE emissions of 315 AAVE/day, with a 4% target yield.

Evaluation of Spending

Umbrella’s forecasted annual emissions represents less than Q1 2025 Borrow Fee (Reserve Factor) revenue from each respective reserve. However, Q3 and Q4 2024 revenue, highlights the significant growth Aave Protocol has enjoyed and if market conditions change, we expect the Target Liquidity parameters and Emissions rate to be adjusted accordingly.

A testiment to the DAO’s strong financial position, the annual budget can be drawn from existing reserves without impacting any Service Provider, or existing approved DAO expenses. Future funding requests shall be configured to ensure Umbrella is adequately funded at all times.

By utilising aTokens from the Aave Protocol as the base yield and providing an additional incentive yield, the Umbrella yield remains correlated to the broader market cycle. This is an attractive quality that ensure users depositing into Umbrella receive a yield that expands and contracts with broader market.

stkGHO to Temporary sGHO

By removing both the Cooldown and Slashing feature from stkGHO, GHO deposited into the pre-umbrella staking contract receives a risk-free interest rate. By increasing the cadence of Merit rewards, the pre-umbrella stkGHO becomes increasingly more like sGHO user experience. By removing the Cooldown feature users who want to migrate between pre-umbrella stkGHO and the new Umbrella stkGHO can do so seamlessly.

Given the relative performance of GHO emissions via Merit relative to the SSR, we expect capital to flow into the pre-umbrella stkGHO contract which can be rebranded as sGHO to better reflect the new risk profile. The current Merit APR offers a Base yield 7.43% before applying any Boost factor, this exceeds the SSR and generate Aave Protocol deposit rate. We recommend refining the Boosts to focus on the Base yield to amplify the yield and also for @AaveLabs to introduce sGHO to the frontend to better reflect the new risk profile.

stkABPT

Since the recent stkABPT Emission update, minimal liquidity withdrawals from the liquidity pool has occurred and it was not until soon after the initial Umbrella configuration were shared did two wallets trigger Cooldown. 317,952 of the 630,471 BPT tokens staked in stkABPT can be withdrawn on the 5th June 2025. The net affect is the pools TVL will reduce by 50.4%, or $108.7M in TVL. Whilst the liquidity pool remains at over $100M, the largest on Balancer Protocol, the pool is more than adequate to facilitate the overwhelming majority of swaps.

If the two address do withdraw BPT for AAVE and wstETH, leaving the liquidity pool, we recommend reducing the AAVE emissions bring the yield back to more recent levels. The AAVE emission adjustment from 240 AAVE/day to 216 AAVE/day, could further to 10-12% yield.

stkAAVE

With the DAO to spend 10M in stablecoins on AAVE buybacks, the $5,857,304 spent to date has acquired 34,334 AAVE comparing favourably to the 78,840 proposed for maintaining stkAAVE emissions. The DAO has progressively shifted emissions to being funded from fees generated by the Aave Protocol.

Now the Umbrella’s development has progressed and the overwhelming success acheived from the buyback program, future capital allocation towards the buyback program is to be revised. The DAO is adequately capitalised with AAVE and is in a great position to fund stkAAVE emissions. We support the direction as outlined by @bgdlabs and expect the yield to drift lower as the Slashing risk is reduced. However, given the financial health of the DAO, we favour a lesser reduction in AAVE emissions, instead of 216 AAVE/day, we recommend 315 AAVE/day resulting in an expected yield of around 4.00% at the current deposit level.

We anticipate as rewards reduce, some actors will favour more sophisticated derivative strategies that offer a higher yield. We support the continued use of AAVE emissions until such time as stkAAVE evolves into a staking role that acts as a liquidity sink receiving rewards generated from the protocol revenue and other protocol benefits, as mentioned in the AAVEnomics update. Until this aspect of the AAVEnomics is ready we prefer only minor amendments to stkAAVE emmissions whilst reducing slashing risk.

[bgdlabs]

Thanks @LlamaRisk and @TokenLogic for the feedback.

Given that the requests for changes are pretty reasonable, we will be applying them on the final ARFC Snapshot to be created imminently:

• [@llamarisk] Change Target Liquidity of aUSDT and aUSDC to $104m and $66m respectively, adjusting emission to still target the same APY ranges as on the initial proposal.

• [@TokenLogic] Remove cooldown from legacy stkGHO, with the asset becoming a de facto temporary sGHO.

• [@TokenLogic] Reduce less than originally proposed stkAAVE rewards, for the change to not be aggressive, to 315 AAVE/day.

As a summary, the final proposed parameters will be the following.

Umbrella

Staked asset	Covered asset	Target Liquidity *	Max emission (rewards at target liquidity)	Umbrella APY range (up until excess liquidity) **	Total APY (Aave + Umbrella)***	Cooldown/unstake window	Deficit offset
aUSDC (wrapped)	USDC	66’000’000 USDC	2’330’000 aUSDC/year	1.76%-7.06%	6.8%-12.5%	20/2 days	100’000 aUSDC
aUSDT (wrapped)	USDT	104’000’000 USDT	3’670’000 aUSDT/year	1.76%-7.06%	6.8%-12.5%	20/2 days	100’000 aUSDt
aWETH (wrapped)	WETH	25’000 ETH	550 aWETH/year	1.1%-4.4%	3%-6.3%	20/2 days	50 aWETH
GHO	GHO	12’000’000 GHO	1’200’000 aGHO/year	5%-20%	5%-20%	20/2 days	100’000 GHO

*Target Liquidity is denominated in the contracts in wrapped aTokens, increasing over time in exchange rate. That means the Target Liquidity itself will grow slightly over time.
For the sake of simplicity, the number on the table is in equivalent terms of underlying (USDC, USDT, WETH), not in wrapped aTokens
** Umbrella has an upper limit of APY as max emission is capped, but technically no lower limit. However, going over the table’s lower point of APY would mean there are way more deposits than expected, hence the market pricing “cheaper” the risk of staking.
*** 1y average supply rates for each asset on Aave are taken as reference

Legacy Safety Module new configuration

Staked asset	Rewards/day (AAVE)	Rewards/year (AAVE)	Rewards/year ($)	Average total staked ($)	Slashing eligibility	Total Slashable ($)	Swap effectivity (inverse slippage)	Effective coverage ($)
AAVE	315 AAVE	114’975 AAVE	22’995’000	500’000’000	20%	100’000’000	85%	85’000’000
AAVE/wstETH Balancer v2	216 AAVE	78’840 AAVE	15’768’000	162’000’000	20%	32’400’000	90%	29’160’000
GHO	0	0	0	0	0	0	0	0

*Merit incentives will not be touched on the legacy stkGHO, and coverage incentives will be moved to the new stkGHO.
*AAVE price $200 avg for the modelling

To refresh aggregated numbers with the last modifications:

• The global coverage will be higher post-Umbrella:
  • Pre-Umbrella Phase 1: ~$305m.
  • Post-Umbrella Phase 1: ~$353m
• The comparison of rewards’ expenses pre- and post-Phase 1 Umbrella is as follows:
  • Pre-Umbrella Phase 1: ~$51m/year (in AAVE token).
  • Post-Umbrella Phase 1: ~$47m/year (~$38.5m in AAVE token + ~$8.5m in aTokens).

[bgdlabs]

Following the procedure, we have published the ARFC Snapshot with the final proposed configuration of this initial activation of Umbrella.

Voting will start in approximately 24 hours, participate
https://snapshot.box/#/s:aavedao.eth/proposal/0xbe792a1db33cd7803e23810553e5a6a728c3ac15827ad2652aa6de1858fa5596

[bgdlabs]

Following the pre-approval of the final activation parameters of Umbrella in ARFC Snapshot, we have created the proposal (AIP) for the Aave Governance to activate Aave Umbrella on Aave v3 Ethereum Core.

Voting will start in 24 hours, participate
https://vote.onaave.com/proposal/?proposalId=320

[bgdlabs]

With the final activation proposal already in the voting stage, we would like to share with the community the following FAQ, on specific aspects surrounding how Umbrella will look on release, and also implications for stakers on the legacy Safety Module.

All the following points will apply once Aave governance proposal 320 gets executed, currently estimated for Thursday, 5th June ~13.20 GMT.

I’m staking GHO on the legacy Aave Safety Module, what should I do?
It depends.
You can stay on the legacy stkGHO without any more risk of slashing and a cooldown of 20 days, but with lower rewards.
Or you can migrate to Umbrella stkGHO, with higher rewards but risk of slashing and 20-day cooldown. If you opt for this, you need to do the following once the governance proposal is approved and executed:

1. Cooldown (required step, but you don’t need to use any cooldown period) on the legacy stkGHO.
2. Unstake your GHO.
3. Stake your GHO on Umbrella, for example, on stake.onaave.com, an instance of the official Aave Umbrella UI (owned by the DAO), run by BGD Labs.

I’m staking AAVE and ABPT on the legacy Safety Module (stkAAVE, stkABPT), anything I should be aware of, or that I need to do?
On this initial activation phase of Umbrella, no action needs to be taken if you are staking on stkAAVE or stkABPT. The only aspects you need to be aware of are that the slashing percentage (how much the system can slash from your staked balance) is reduced from 30% to 20%, and that rewards are slightly reduced too. But the system remains fully functional.

Which risks do I need to be aware of when staking on Umbrella?
The primary Umbrella risk is the risk of slashing, which occurs if deficit accrues in the associated asset of the connected Aave pool (USDC on v3 Ethereum Core, for instance, if you are staking aUSDC). Historically, the Aave DAO has never slashed stakers of the previous Safety Module, and Umbrella has a mechanism of offset by which the DAO will always cover first-loss up to X amount (e.g., 100’000 in the case of USDT or USDC). But you should understand that it can happen. Additionally, like any smart contract system, there can potentially be bugs that create any type of problem. This has been minimized by all the security procedures disclosed in this forum post and the proposal, including all our internal processes of development at BGD, and four external security audits.

Can you elaborate on what the Umbrella deficit offset means in simple terms?
For example, on this initial proposal, on staked USDT, the DAO configures an offset of 100’000 USDT. What that means is that more than 100’000 USDT bad debt on the Aave v3 Core pool should be accrued before a single USDC staked asset gets slashed. Basically, the DAO steps up to that amount first to cover any loss, before stakers.

So by staking on Umbrella, I will always earn the APY defined in the previous forum posts?
Not exactly. Umbrella yield has two components:

1. When staking any non-GHO token (USDC, USDT, WETH initially), you will be earning aToken yield, e.g., ~4% on aUSDT as of the 1st of June. This will accrue automatically in your staked assets, as they will grow in value if no slashing happens and there is yield on Aave.
2. Additionally, for assuming the slashing risk Umbrella gives you rewards continuously. These can be modified by the DAO over time, but it should always be an extra yield percentage on top of the basic Aave yield. To understand the dynamics of how the Umbrella Emission Curve works, we highly recommend reading the explanation, but as a high-level rule of thumb: if Target liquidity would be 100m USDC and the configured rewards at that point would cause a 5% extra yield, that means that whenever nobody is staking, yield would be 10% (double), and if total staked is well above the 100m USDC, yield percentage will be lower (but always more than what you earn just supplying on Aave).

How do I accrue the Umbrella rewards? And how do I need to claim them?
Once staking, you accrue Umbrella rewards continuously, and you can claim them at any point you want, but doing a blockchain transaction, for example, via stake.onaave.com. The system is fully on-chain.

I don’t have WETH, but only ETH in my wallet. Can I still stake?
Yes, if you use stake.onaave.com, the UI will build the transaction to wrap your ETH into WETH before staking.

What do I need to do exactly to unstake from Umbrella?
The flow is very similar to the legacy Safety Module: first, you activate the cooldown for the amount you are currently staking. After waiting for the cooldown, currently configured to 20 days, you can unstake your assets. It is not part of the system, but as Umbrella is tokenized, it is also possible that a secondary market for staked Umbrella tokens will appear, which would allow you to get the underlying without cooldown (but probably for a fee).

I have different assets than the ones enabled on this proposal, or I am using Aave on another network, non-Ethereum, will I be able to use Umbrella?
The plan is to expand Umbrella to at least part of the other Aave pools in the very short future, to some of the assets, as not all of them require staking (if they don’t accrue deficit). So yes, even if you don’t have USDC, USDT, ETH, or GHO on Ethereum to stake now, with some other assets in the future, you should be able to, always depending on the Aave governance.

I see that when staking on Umbrella, my aTokens don’t increase in balance, but in value, how so?
Technically, the assets you stake in Umbrella are wrapped aToken (previously known as stata/static a tokens). These are just a special version of aTokens growing in value (exchange rate) and not in balance. Still, in terms of yield, both are exactly equivalent.
If you use stake.onaave.com, the interface will recognize any type of underlying asset and transparently wrap it appropriately for you. This means that no matter if you have e.g. USDT, aUSDT (rebasing, growing in balance), or waUSDT (wrapped already), the interface will allow you to stake.
On unstaking, the interface will also allow you to choose which type of underlying to want to receive. In the previous example USDT, aUSDT or waUSDT.

I am staking aUSDT in Umbrella, the UI shows me $1000, but that being only ~888 aUSDT staked, how so?
What you are staking is wrapped aTokens, based on exchange rate. It means that 1 wrapped aUSDT has a higher value than USDT or rebasing (balance growing, non-wrapped aUSDT), and so your ~888 staked aUSDT value is $1000.

What is the role of BGD Labs on the project?
As service providers to the Aave DAO, we designed and developed Umbrella. But as with another smart contract of the Aave DAO, we don’t have any type of control over the instances that will be activated with the governance proposal, so our role going forward is to monitor that everything works properly and securely, together with suggesting improvements and supporting other service providers working around Umbrella. Additionally, we developed and ran one instance of the official Aave Umbrella UI owned by the DAO on stake.onaave.com, but this is totally permissioned software that everybody else can run, and only helps users of the Umbrella smart contracts build transactions and visualize data.

[Benji533]

Is there an estimated date for adding other stablecoins (USDS and USDe) to umbrella?

[bgdlabs]

No especific date at the moment @Benji533 , it will depend also on the opinion of risk providers and the need for instances on all other assets.

[Kitten]

probably an oversight but stkAave rewards can still be restaked in the legacy safety module

[EzR3aL]

This is correct. You will still be able to stake in the legacy module as it will be depreciated over time only. This means the rewards and slashing will be reduced step by step.

[0xmonk]

Any idea when ANTI-GHO is coming? Thought it was part of umbrella. @ACI

[EzR3aL]

Anti GHO was part of the other proposal, the Aavenomics. It wasn’t said that with Umbrella everything will be there day one. Umbrella was just the first step towards all of this. I would assume we will likely see it within the next 3 months being activated. But again, step by step as many things had been outlined in that proposal.