Chaos Labs - Monthly Community Update

ChaosLabs · October 7, 2024, 9:43pm

September 2024

This update highlights Chaos Labs’ activities and proposals in September.

Highlights

WBTC Parameter Adjustment

We first published our recommendations to change WBTC’s parameters to reflect its new custody arrangement. Following community feedback, we aligned with other service providers to recommend parameter adjustments that will balance the user experience with the asset’s altered risk profile.

LTV and LT Alignment

To continue optimizing the protocol and improving the user experience, we published an exhaustive recommendation to align LTV and LT for the same assets across networks. This improved the borrowing power for many assets, including AAVE, LINK, and rETH by increasing their LTVs on some networks. This added, for example, over $4M more borrowing power to LINK on Ethereum at current supply levels.

cbBTC Listing

We provided recommendations to list cbBTC in a timely manner, conducting research before the asset had been listed by conversing with the relevant teams. This allowed Aave to be one of the first lending protocols to list the asset, whose supply has grown quickly and for which we have facilitated cap increases.

What’s Next

In the coming months, the Chaos team will continue its focus on the following areas:

GHO: ongoing recommendations, including updating parameters for GHO on Arbitrum.
Continuation of the V2 risk parameter updates to gradually reduce capital efficiency across V2 collateral assets.
Continuous optimization of risk parameters on all V3 deployments.
Implement Chaos Labs Risk Oracles.
Analysis and parameter recommendations for new assets and markets.

ChaosLabs · November 5, 2024, 10:52pm

October 2024

This update highlights Chaos Labs’ activities and proposals in October.

Highlights

Migration of WETH-WETH Loopers and E-Mode Optimization

We published a set of recommendations to discourage WETH-WETH looping in the Lido instance. This included providing time to deleverage, reducing the LB of WETH outside of E-Mode, removing WETH from E-Mode collateral, restoring the liquidation bonus, and finally increasing the LTV of WETH outside of E-Mode. This was critical, as we found that forced transitions out of E-Mode could trigger $30M in liquidations.

Gnosis DAO Credit Line Instance

Following a post by ACI, we provided recommendations for a Gnosis DAO credit line instance, a novel instance for Aave. This required the development of ad hoc methodologies to ensure that the credit line remains healthy while also allowing for greater borrowing than would otherwise be facilitated by GNO’s on-chain liquidity.

ezETH and wstETH Liquid E-Modes

We followed up on our previous listing recommendation to provide new E-Mode categories for ezETH and wstETH, including ezETH/USDS and ezETH/wstETH, alongside optimized wstETH/WETH configurations on L2s. Additionally, we updated our non-E-Mode parameter recommendations to account for these new E-Modes. We also updated our CAPO recommendation to account for the recent distribution of EIGEN rewards in ezETH’s exchange rate.

What’s Next

In the coming months, the Chaos team will continue its focus on the following areas:

GHO: ongoing recommendations, including updating parameters for cross-chain GHO.
Continuation of the V2 risk parameter updates to gradually reduce capital efficiency across V2 collateral assets.
Continuous optimization of risk parameters on all V3 deployments.
Further updates on the development of Chaos Labs Risk Oracles.
Analysis and parameter recommendations for new assets and markets.

ChaosLabs · December 3, 2024, 2:45pm

November 2024

This update highlights Chaos Labs’ activities and proposals in November.

Highlights

Onboard and Enable sUSDe Liquid E-Mode

We published a detailed analysis of sUSDe risks with its current oracle setup, specifically measuring the price deviations of USDe and sUSDe. Additionally, we analyzed the liquidity of the sUSDe/sDAI Curve pool. We used the maximum discount observed below (2.47%) as a guide for setting the LB to 3%. The initial recommendation included USDS and USDC as the borrowable assets; a later recommendation added USDT as another borrowable asset.

Adjust weETH Interest Rate Parameters on Aave V3 Scroll.

Following Scroll’s TGE, we observed changes in the V3 Scroll deployment, particularly in the weETH and WETH markets. This was largely related to users removing their WETH supply after the Scroll airdrop. We recommended reducing weETH’s Slope 1 to a level where borrowing weETH against WETH becomes profitable, allowing users to arbitrage rates and reduce WETH’s without depositing new supply.

Deployment of Aave on Linea

We completed an analysis of Linea, including its technical architecture, its ecosystem and market, and its DEXes, tokens, and oracles. Ultimately, we found that it is appropriate for Aave to create a new deployment on Linea. Using historical liquidity, we provided recommendations for listing parameters for six assets. We also provided parameters for an ETH-correlated E-Mode and a wstETH/WETH E-Mode with more aggressive parameters.

Add PAXG to Aave V3 Main Instance on Ethereum

Chaos Labs conducted a detailed technical review of PAXG, which represents direct ownership of physical gold. The listing was complicated by the fact that PAXG is tradable 24/7, differing from the trading hours of the underlying gold. This causes deviations from the gold oracle price, which is not updated on the weekend. We observed a 22.2% divergence between the PAXG and XAU oracle prices in April 2024. Ultimately, we recommended using the XAU market price oracle, finding that it acts as a reduced LTV during upside deviations and it is statistically improbable for bad debt to accrue during downside deviations based on the historical performance of gold.

What’s Next

In the coming months, the Chaos team will continue its focus on the following areas:

GHO: ongoing recommendations, including updating parameters for cross-chain GHO.
Parameterization for new Liquid E-Modes.
Continuation of the V2 risk parameter updates to gradually reduce capital efficiency across V2 collateral assets.
Continuous optimization of risk parameters on all V3 deployments.
Further updates on the development of Chaos Labs Risk Oracles.
Analysis and parameter recommendations for new assets and markets.

ChaosLabs · February 5, 2025, 7:03am

January 2025

This update highlights Chaos Labs’ activities and proposals in January.

Highlights

sUSDe and USDe Price Feed Update

We published two highly detailed analyses of sUSDe/USDe price feeds, recommending that they be switched to USDT price feeds. Our analysis found that USDe’s fundamental value is closely linked to USDT, as the majority of its perpetual hedges are denominated in USDT. We presented numerous scenarios, including exchange failure, finding that pricing USDe according to USDT would lead to preferable outcomes for the protocol and users.

sUSD Risk Parameter Adjustments

Following a depeg of sUSD, we moved rapidly to recommend a decrease in its supply and borrow caps to prevent growth in the market. As an additional step, we later analyzed recent activity in the market and sUSD’s peg dynamics and mechanism design. We found that its current economics had contributed to a destabilized peg, even as it remained over-collateralized. As a result, we recommended setting sUSD’s LTV to 0%, as well as reducing the LTV and LT in the Optimism Stablecoins E-Mode.

ggAVAX Listing

We recommended listing ggAVAX, an AVAX LST that utilizes MiniPools, a relatively unique mechanism that required further analysis. We had previously analyzed the asset, finding that its withdrawal system could pose a problem, as some tokens were locked for extended durations. Upon analyzing again, we found that this dynamic had changed, and average lock duration had dropped significantly. This fact, coupled with an analysis of the asset’s market pricing relative to its exchange rate, led us to recommend listing the asset with an E-Mode to facilitate leveraged yield strategies.

Deploy stataUSDC and statUSDT GSMs on Ethereum

We responded in favor of a proposal to migrate existing USDT and USDC GSMs on Ethereum to their stataToken counterparts. We noted that, in a previous analysis, we had supported the proposal to use tokens representing deposits on Aave in GSMs. Upon another analysis, we reiterated this support, finding that there was always substantial borrowable liquidity in the markets in question.

What’s Next

In the coming months, the Chaos team will continue its focus on the following areas:

Risk Oracle integrations leveraging Edge infrastructure.
Continuous optimization of risk parameters on all V3 deployments.
Analysis and parameter recommendations for new assets and markets.
Continuation of the V2 risk parameter updates to gradually reduce capital efficiency across V2 collateral assets.
Parameterization for new Liquid E-Modes.
GHO: ongoing recommendations, including updating parameters for cross-chain GHO.

ChaosLabs · March 4, 2025, 10:23pm

February 2025

This update highlights Chaos Labs’ activities and proposals in February.

Highlights

Aave V2 Deprecation Update

As the protocol continues its migration away from V2, we conducted an analysis of V2 markets, including the bad debt distributed by asset. We found that virtually all of the bad debt is on Ethereum, and AMPL and BUSD account for a high share of the bad debt. Based on these findings, we recommended reducing the interest rate for markets in which bad debt represented the majority of borrows to 1%. We also recommended disabling new borrows for all V2 assets.

Insights from Market Events

We published multiple timely updates on market events, the first being the high volatility at the beginning of the month, and the latter being the Bybit exploit. The Bybit exploit was of special interest given Ethena’s use of the exchange. However, our detailed analysis found that Ethena continued to process redemptions effectively, minimizing the observed depeg, especially on chain. Ultimately, there were six liquidations during the event, with $22M of sUSDe collateral seized.

Onboard tBTC to Aave

We conducted an analysis of tBTC, finding that it uses threshold cryptography to distribute key generation among 100-node signer groups, each requiring a stake in T tokens. We observed that its security relies on a (51,100) threshold scheme and strict slashing rules, ensuring no single entity controls private keys, despite the potential for manipulation if excessive stake was accumulated. We concluded that market analysis showed stable liquidity and volatility parameters, supported by a limited pool of 35 whitelisted operators.

Risk Oracles

Supply and Borrow Caps

Building on the integration of Aave Generalized Risk Stewards, we went live with the Supply and Borrow Cap Risk Oracle, which automates the generation and application of cap recommendations for Aave, leveraging a proven simulation engine that has successfully supported manual cap adjustments through the Risk Steward for years. Depending on whether the cap utilization surpasses or falls below specific thresholds, either the cap increase or cap decrease simulation can be triggered, in addition to the frequent time-based triggering of simulations for all respective markets. The simulation then determines the optimal new cap value based on a range of risk metrics. The introduction of Risk Oracles will significantly reduce operational overhead and allow for faster response to changing cap utilization.

The chart below shows the amount of supply and borrow cap changes since October 1, 2024.

Pendle Oracle

We proposed a Risk Oracle system that integrates a volatility-structured pricing mechanism for Pendle’s Principal Tokens on Aave, ensuring manipulation-resistant rate updates and a killswitch for minimum price scenarios. This includes a dynamic liquidation threshold and bonus that adjusts as the PT nears maturity, safeguarding against tail risks accounting for the associated duration risk within the asset.

What’s Next

In the coming months, the Chaos team will continue its focus on the following areas:

Risk Oracle integrations leveraging Edge infrastructure.
Continuation of the V2 risk parameter updates to gradually reduce capital efficiency across V2 collateral assets.
Parameterization for new Liquid E-Modes.
Continuous optimization of risk parameters on all V3 deployments.
Analysis and parameter recommendations for new assets and markets.
GHO: ongoing recommendations, including updating parameters for cross-chain GHO.

ChaosLabs · April 3, 2025, 8:54am

March 2025

This update highlights Chaos Labs’ activities and proposals in March.

Highlights

Aave <> Chainlink SVR v1. Phase 1 Activation

We recommend activating the first phase of Chainlink’s Smart Value Recapture (SVR) on Aave to enable the protocol to reclaim non-toxic MEV generated during liquidations triggered by Chainlink Price Feeds. SVR uses a Dual Aggregator design, where a single Chainlink oracle network publishes the same price update to two destinations: one to the SVR Price Feed via Flashbots MEV-Share, and the other to the Standard Price Feed via the public mempool. This structure allows Aave to capture MEV value from searcher bidding while maintaining compatibility with existing Chainlink interfaces. If the SVR update is delayed or fails to land onchain, a fallback mechanism kicks in after a configurable delay, pulling the price from the Standard Feed to ensure price availability. This design preserves liveness while ensuring liquidators can’t bypass MEV-Share to avoid contributing value back to the protocol.

We support the deployment of SVR, which we believe will introduce a new revenue stream for Aave while incurring minimal expected losses. Specifically, our simulation using position data from the Aave Ethereum Core market within a selected time window shows that the total expected bad debt remains negligible, even with oracle delays introduced by SVR. Simultaneously, the OEV capture rate is projected to remain around 40–50%, translating into significant protocol revenue. However, to ensure robust risk management, we also recommend several mitigation strategies, including asset-specific volatility assessments, deviation threshold adjustments, and protocol fee tuning for SVR-integrated assets.

Listing USR

We recommended listing USR on Aave V3’s Core instance based on our technical assessment of its liquidity profile, market adoption, and underlying smart contract infrastructure. We evaluated USR’s volatility characteristics, collateralization parameters, and historical performance to propose an appropriate set of risk parameters—encompassing supply and borrow caps, liquidation thresholds, and interest rate curves—that would protect Aave’s existing markets while promoting USR’s healthy utilization. By integrating USR with these tailored safeguards, we believe Aave can offer users a new stable and robust asset option without compromising the protocol’s overarching risk standards.

New Deployments

We recommended deploying Aave V3 on INK, Celo, Soneium, and MegaETH based on comprehensive risk assessments that evaluated each network’s security architecture, bridging mechanics, and on-chain liquidity prospects. By proposing tailored risk parameters—covering collateral factors, supply/borrow caps, and interest rate models—we aimed to safeguard Aave’s core markets while facilitating sustainable liquidity growth. For Plasma, we instead advised a cautious approach: our analysis highlighted the need for additional technical details (e.g., whitepaper clarifications, stable testnet performance) before finalizing any deployment recommendation. This distinction ensures that Aave’s expansion to new chains remains aligned with rigorous safety criteria and the protocol’s broader growth strategy.

Risk Oracles

Supply and Borrow Caps

Following the introduction of automated supply cap updates using Risk Oracles, we are able to report on the first month that they have been utilized. As plotted below, they have primarily been active in gradually reducing the supply caps for underutilized assets like FRAX and and DAI. This significantly reduces the manual work required to make consistent cap updates, including lowering the caps of underutilized assets.

What’s Next

In the coming months, the Chaos team will continue its focus on the following areas:

Supply and Borrow Cap Risk Oracle integration on additional Chains leveraging Edge infrastructure.
Parameterization for new Liquid E-Modes.
Continuous optimization of risk parameters on all V3 deployments.
Analysis and parameter recommendations for new assets and markets.
GHO: ongoing recommendations, including updating parameters for cross-chain GHO.
Umbrella parameterization and methodology
Continuous monitoring of SVR and associated parameterization
Pendle Dynamic Risk Oracle for each PT asset deployment
Circuit Breaker for LSTs and LRTs
Interest Rate Curve Automation

ChaosLabs · May 6, 2025, 6:33pm

April 2025

This update highlights Chaos Labs’ activities and proposals in April.

Highlights

Staking Penalties on Ethereum’s Consensus Layer: Implications for wstETH and Other LSTs and LRTs

We conducted in-depth research on the impact of Ethereum consensus-layer slashing and penalties, with a focus on Aave’s exposure through LSTs and LRTs. We considered the changes in risk to the protocol before and after Pectra, which alters how validator balances and slashing penalties are handled. This includes changes to the initial slashing penalty and an improvement to the correlation penalty. We found that Lido’s current approach to the Pectra upgrade is focused on ensuring full compatibility, rather than immediately adopting new features. We recommended a variety of mitigation measures that can reduce staking risks, including client diversity and a preference for minority clients. Finally, we modeled a variety of stress scenarios, from a low-scale isolated slashing event to a CL minority bug with slashing. We found that most events would result in manageable losses for Aave, while the severe events are highly unlikely. This analysis ultimately acted as the basis of a more capital-efficient parameterization of wstETH within the Aave Protocol.

SVR Dashboard

We released a new dashboard detailing the use of SVR on Aave. This includes a breakdown of the number of liquidations and the profit for Aave.

It also includes a number of charts showing a variety of metrics, including recapture rate by asset, showing the top five and bottom five.

Finally, it includes a granular listing of all SVR liquidation events on Aave, providing greater insight into these events.

The dashboard can be found here.

PT Token Listings

We recommended the listing of Principal Tokens (PTs) on Aave and launched our risk oracle to enable continuous risk parameterization for these assets. This initiative aligns with our Principal Token Risk Oracle Framework and leverages the dynamic linear discount rate oracle implementation, as detailed in our methodology here and here. Since their listing, PTs have seen strong demand.

New Deployments

We provided an update to our recommendation to deploy Aave V3 on Mantle, maintaining a cautious stance on asset listings due to the network’s relatively volatile liquidity. However, we were able to provide preliminary listing recommendations for WETH, USDC, USDT, and USDe. We also recommended a new deployment on Aptos, providing an overview of network activity, tooling, DEXes, and available tokens, and were able to provide listing recommendations because the network is already established.

Risk Oracles

Supply and Borrow Cap Oracles

The supply and borrow cap risk oracle on Arbitrum continued to function flawlessly, this month making incremental changes to DAI’s supply and borrow caps, while making continued reductions to wstETH and WBTC’s borrow caps.

We deployed cap oracles on Avalanche and Base this month, leading to reductions in DAI.e’s caps on Avalanche and an increase in AUSD’s supply cap. There were fewer changes on Base, where the oracle recommended changes to cbETH’s borrow cap and increases in EURC’s caps.

PT Risk Oracle

The newly listed assets utilizing our PT risk oracle, described above, have reached $485M in combined TVL shortly after listing.

Edge UI

Forum Activity

What’s Next

In the coming months, the Chaos team will continue its focus on the following areas:

Supply and Borrow Cap Risk Oracle integration on additional Chains leveraging Edge infrastructure.
Continuous monitoring of SVR and associated parameterization
Pendle Dynamic Risk Oracle for each PT asset deployment
Risk Oracle integration for automated interest rate curve adjustments in response to demand
Circuit Breaker for LSTs and LRTs
Umbrella parameterization and methodology
GHO: ongoing recommendations, including a comprehensive quantitative framework for the GHO savings rate.
Continuous optimization of risk parameters on all V3 deployments.
Parameterization for new Liquid E-Modes.
Analysis and parameter recommendations for new assets and markets.

Interest Rate Curve Risk Oracle

While continuing the rollout of the Supply and Borrow Cap Risk Oracles, this month, we presented our methodology for an Interest Rate Curve Risk Oracle. This methodology utilizes demand models to drive protocol optimization. The current method of stablecoin IR updates requires considerable operational effort, with manually created transactions for more than 60 stablecoin markets on 16 deployments. The risk oracle will be able to perform more granular and frequent updates, maintaining a higher utilization for the stablecoin markets on which it operates. As shown below, this would lead to higher protocol revenue.

ChaosLabs · June 4, 2025, 9:06pm

May 2025

This update highlights Chaos Labs’ activities and proposals in May.

Highlights

Staking Penalties on Babylon’s Bitcoin Staking Protocol

Following last month’s in-depth research on the impact of Ethereum consensus-layer slashing and penalties, this month we turned our focus to Babylon’s staking protocol. Specifically, we reviewed the protocol workflow, key protocol components, and examined Bitcoin Supercharged Networks and the Babylon Genesis Chain. Our analysis continued on the penalties that can be accrued via the remote slashing architecture, including for equivocation by finality providers and inactivity or downtime. We also examined the slashing protection measures, including client-level protections and external anti-slashing tools. Finally, we conducted a slashing scenario analysis, finding that the net effect of a slashing scenario would be a negligible impairment in collateral value.

PT Token-Related Adjustments and Listings

We responded to two proposals — to onboard USDe July and August PT token on Ethereum Core. Additionally, we recommended the adoption of a PT token and USDe E-Mode, in which USDe-related PT tokens would be used to borrow USDe, reducing protocol risk by using the underlying asset for leverage. This necessitated a change in USDe’s parameters, specifically altering its interest rate curve to capture additional protocol revenue. The PT token to USDe E-Modes without any changes would be expected to generate greater user profit because of its more permissive LT parameters; to account for this we proposed a series of hikes in USDe’s target interest rate that would make its yield comparable with that of the normal stablecoin E-Mode.

Ultimately, these changes, coupled with a higher Reserve Factor for USDe, has made USDe a significant revenue generating market for Aave, especially considering its size, currently accruing nearly $7M annualized; these figures can be viewed on our newly updated Community Analytics page.

New Deployments

We provided responses to proposals for new Aave deployments on the following networks:

Rootstock
- We supported the deployment on the network contingent upon the availability of adequate price feed infrastructure. We recommended listing USDT, WRBTC, and ETHs as initial assets.
Soneium
- We supported the proposed deployment and recommended listing WETH, USDC.e, and USDT0.
Tron
- We also supported the proposed deployment on Tron and recommended listing USDT and WTRX as initial assets.

Risk Oracles

Supply and Borrow Cap Oracles

The supply and borrow cap risk oracles on Arbitrum, Avalanche, and Base continued to function efficiently, executing changes across a wide variety of assets. The most significant changes came on Base - EURC, where the oracle allowed for significant increases in the asset’s supply and borrow caps.

In the other direction, it has also frictionlessly reduced caps for assets with lower demand, including USDC.e on Arbitrum and LBTC on Base.

PT Risk Oracle

The newly listed assets utilizing our PT risk oracle have reached $1.456B in combined TVL. The chart below shows two of the three parameters dictated by risk oracles: Liquidation Bonus and Liquidation Threshold, as well as total supply and supply cap, represented by the dashed lines. In the case of PT-eUSDe-May, it maxed out its LT by May, experiencing only reductions in its LB.

PT-sUSDe-July, however, had a consistently increasing LT, and thus required numerous supply cap increases; supply increased nearly immediately following the implementation of cap increases. The new listings of PT-USDe-July and PT-eUSDe-August have already achieved significant adoption.

Forum Activity

What’s Next

In the coming months, the Chaos team will continue its focus on the following areas:

Supply and Borrow Cap Risk Oracle integration on additional Chains leveraging Edge infrastructure.
Continuous monitoring of SVR and associated parameterization
Pendle Dynamic Risk Oracle for each PT asset deployment
Risk Oracle integration for automated interest rate curve adjustments in response to demand
Circuit Breaker for LSTs and LRTs
Umbrella parameterization and methodology
GHO: ongoing recommendations, including a comprehensive quantitative framework for the GHO savings rate.
Continuous optimization of risk parameters on all V3 deployments.
Parameterization for new Liquid E-Modes.
Analysis and parameter recommendations for new assets and markets.

ChaosLabs · July 8, 2025, 8:01pm

June 2025

This update highlights Chaos Labs’ activities and proposals in June.

Highlights

Addition of USDe to Ethena PT E-Modes

The previous creation of a USDe-specific E-Mode allowed users to leverage USDe-linked PT tokens against USDe itself was a success, as USDe has generated a significant revenue relative to its size.

However, to move between the Stablecoin and USDe E-Modes was cumbersome for users, requiring the full unwinding of a position. To address these frictions, we recommended adding USDe to the Stablecoin E-Mode, allowing users to easily migrate between the two without unwinding their position.

USDT IR Adjustments

We conducted active monitoring of markets leading to rapid-response recommendations for USDT, helping to mitigate rate spikes as the result of user behavior. HTX (formerly Huobi) was engaged in a pattern of depositing and withdrawing large amounts of USDT on Core; this became increasingly problematic as utilization increased because of USDT borrow demand from PT token deposits.

To address this, we first lowered Slope 2, then lowered it further and increased UOptimal, then finally reverted Slope 2 and UOptimal to their previous values. This led to limited disruptions to users and continued growth on both the supply and borrow sides.

USDT on Core

Onboard wstLINK to Aave V3 Core Instance

Chaos Labs supported the listing of wstLINK on Core, finding that the asset was suitable from a risk perspective and may provide a significant increase in LINK borrow utilization. Chainlink Staking currently operates in v0.2, which is composed of the Community and Node Staking Pools, the former being individual or institutional holders who do not provide oracle services directly and are not subject to slashing. Node Operators are responsible for running oracle services and can be slashed.

Based on the staked LINK market dynamics, stLINK has traded at a slight premium to LINK on secondary markets, reflecting that airdrops from Chainlink BUILD projects have begun, and users were expecting it to be listed on Aave, which would effectively increase its yield. Along these lines, we also recommended an adjustment to LINK’s IR and borrow cap parameters to better optimize for this use case.

Early Renewal Proposal

Chaos Labs proposed a 12-month early renewal of our risk management partnership with the Aave DAO as a new and separate engagement stream in parallel with the current engagement. This proposal reflected Chaos Labs’ transition from periodic risk reviews to delivering deeply integrated, protocol-native risk automation infrastructure. Additionally, we noted that we have conducted:

64 asset listing analyses in the past 7 months (vs. 46 in the entire previous year), reflecting a sharp increase in volume and diligence scope.
13 new deployment listings, with 10 scheduled to go live shortly, significantly expanding protocol reach and impact.
226 supply cap, 203 borrow cap, and 572 interest rate curve updates executed through the Risk Stewards, scaling operational coverage.
260 total forum posts spanning proposal authorship, parameter reviews, research, and real-time risk insights, reinforcing Chaos Labs’ governance leadership.

The section below covers two of our most prominent Risk Oracle deployments, managing caps and PT token collateral parameters, respectively. We have also significantly enhanced our monitoring platform, especially with a new SVR dashboard that displays value recaptured over time, amongst other metrics.

Risk Oracles

Supply and Borrow Cap Oracles

The supply and borrow cap risk oracles on Arbitrum, Avalanche, and Base continued to function efficiently, executing changes on numerous markets. The most significant changes came on Arbitrum’s wstETH market, where the oracle recommended increasing the borrow cap from 1,400 to 3,075 over the course of multiple adjustments.

This allowed for frictionless increases in wstETH borrows on Arbitrum as a result of increased leverage against rsETH.

wstETH on Arbitrum

PT Risk Oracle

The value of assets utilizing the PT risk oracle has eclipsed $2B in combined TVL. The chart below shows two of the three parameters dictated by risk oracles: Liquidation Bonus and Liquidation Threshold, as well as total supply in USD terms.

We can observe that PT-sUSDE’s LT remained steady while its LB decreased, reflecting its approaching maturity. Meanwhile, the LTs of both PT-USDe and PT-eUSDE increased throughout the month in conjunction with their LBs falling. The oracle continues to function optimally and has allowed for significant, responsible growth in these assets.

Forum Activity

What’s Next

In the coming months, the Chaos team will continue its focus on the following areas:

Supply and Borrow Cap Risk Oracle integration on additional Chains leveraging Edge infrastructure.
Continuous monitoring of SVR and associated parameterization
Pendle Dynamic Risk Oracle for each PT asset deployment
Risk Oracle integration for automated interest rate curve adjustments in response to demand
Circuit Breaker for LSTs and LRTs
Umbrella parameterization and methodology
GHO: ongoing recommendations, including a comprehensive quantitative framework for the GHO savings rate.
Continuous optimization of risk parameters on all V3 deployments.
Parameterization for new Liquid E-Modes.
Analysis and parameter recommendations for new assets and markets.

ChaosLabs · August 4, 2025, 9:24pm

July 2025

This update highlights Chaos Labs’ activities and proposals in July.

Highlights

Stress Testing Ethena: A Quantitative Look at Protocol Stability

We published a detailed report outlining Aave’s increasing exposure to Ethena-linked assets, especially PT tokens, finding that nearly 50% of USDe is currently on Aave. Next, we examined the risks associated with the exposure, including PT token liquidity (mitigated using risk oracles), and USDe liquidity risk. This liquidity risk was further broken down into supply concentration risk, finding that two entities account for over 60% of the USDe on Aave, as well as USDe rehypothecation risk.

We found that to appropriately mitigate these liquidity risks, it is important to maintain a more conservative UOptimal for USDe relative to USDT and USDT, as well as to integrate USDe into PT stablecoin E-Modes to reduce frictions when utilizing Debt Swaps. Finally, we examined the risks posed by USDe’s backing, modeling numerous scenarios to determine risks (the chart below shows the quantitative assessment of Ethena hedging before an exchange failure is confirmed by a custodian), ultimately finding that Ethena’s hedging and collateral management systems are robust, and the probability of bad debt accruing to Aave is extremely low.

Dynamic Calibration of CAPO Parameters via Risk Oracles

We proposed a formal framework for calibrating CAPO parameters using Risk Oracles, which would enforce a time-weighted upper bound on the permissible growth of a yield-bearing asset’s exchange rate relative to its base asset. Specifically, it is critical to refresh the snapshotRatio at a consistent cadence to ensure that the maxRatio is in line with real-world observations. When snapshotRatio is not refreshed, it leads a model with reduced responsiveness, as shown in the charts below. Taking this into account, we provided details on our proposed CAPO risk oracle, designed to improve the efficiency and risk mitigation properties of CAPO.

Our framework accommodates assets with either regular or irregular yield distributions, stressed via backtesting. In the case of ezETH, its March distribution of EIGEN rewards into its exchange rate provided an important test. Under backtesting, the rate-setting algorithm executed a parameter recalibration in response to the updated reward trajectory. Importantly, the required minimum 3-day MA of APY, used as a guardrail to prevent excessive rate increases, converged below the observed 3-day APY. This meant the new max rate was deemed within the bounds, increasing the associated maxYearlyGrowthPercent.

Adjust USDT and WETH IR Curves

We continued to monitor and frequently update the IR curves of both WETH and USDT in response to large withdrawals by the HTX account from Aave V3. The WETH withdrawals in particular created ripple effects, as users who were leveraging LSTs and LRTs were faced with unprofitable positions, and unwinding caused minor peg deviations in these assets. To address this, we temporarily lowered Slope2 and increased UOptimal; similar actions were taken in the USDT market, where large withdrawals caused a spike in borrow rates.

Risk Oracles

Supply and Borrow Cap Oracles

The supply and borrow cap risk oracles on Arbitrum, Avalanche, and Base continued to function efficiently, executing changes on numerous markets. The most significant changes came on Base’s EURC market, where the oracle recommended multiple supply and borrow cap increases.

This allowed for a frictionless rise in EURC utilization on Base, in part because of increased borrowing against cbBTC.

PT Risk Oracle

The value of assets utilizing the PT risk oracle has eclipsed $3B. The chart below shows two of the three parameters dictated by risk oracles: Liquidation Bonus and Liquidation Threshold, as well as total supply in USD terms.

PT-sUSDE-31JUL2025’s LT remained steady while its LB decreased, reflecting its approaching maturity at the end of July. Meanwhile, the LT of PT-sUSDE-25SEP2025 has been consistently increasing since listing, while its LB has decreased slightly. The oracle continues to function optimally and has allowed for significant, responsible growth in these assets.

Forum Activity

What’s Next

In the coming months, the Chaos team will continue its focus on the following areas:

Supply and Borrow Cap Risk Oracle integration on additional Chains leveraging Edge infrastructure.
Continuous monitoring of SVR and associated parameterization
Pendle Dynamic Risk Oracle for each PT asset deployment
Risk Oracle integration for automated interest rate curve adjustments in response to demand
Circuit Breaker for LSTs and LRTs
Umbrella parameterization and methodology
GHO: ongoing recommendations, including a comprehensive quantitative framework for the GHO savings rate.
Continuous optimization of risk parameters on all V3 deployments.
Parameterization for new Liquid E-Modes.
Analysis and parameter recommendations for new assets and markets.

Sberex · August 5, 2025, 11:20am

Hi @ChaosLabs, can we consider introducing an E-Mode category for EURC and increasing Max LTV and Liquidation threshold for EURC market on AAVE V3 from the current 75%/78% to 90%/92%? Possibly adding EURC to existing E-Mode options like popular sUSDe stablecoins. Thank you!

Topic		Replies	Views
Keyrock Delegate Platform Delegate Platforms	34	4855	September 19, 2024
StableLab Delegate Platform Delegate Platforms	99	7302	August 12, 2025
Chaos Labs - Aave V3 Ethereum (Initial Parameters) Risk	1	2713	January 20, 2023
[ARFC] Chaos Labs <> Aave Risk Management - Renewal Governance	7	2247	December 7, 2023
LBS Blockchain Society Delegate Platform Delegate Platforms	684	17944	February 28, 2024