The recent integration of Pendle Principal Tokens (PTs) as collateral on Aave’s core market has been met with remarkable success, attracting over $1.6 billion in supplied assets in just over a month. This surge in demand is largely fueled by favorable broader market conditions, particularly the increased return generated by Ethena’s USDe delta-neutral perpetual strategy. As future yield expectations have risen, PTs have re-emerged as an attractive instrument for users looking to lock in future earnings, and Aave has established itself as a preferred venue for this activity.
Aave enhances the utility of PTs by enabling yield-leveraging strategies. While the option for such strategies existed on other platforms before, Aave’s architecture has provided a competitive advantage through its unified liquidity model, which offers more flexibility and lower rates to borrowers. Furthermore, the Aave DAO has implemented a more efficient pricing mechanism, the dynamic linear discount rate oracle. This approach is designed to provide a more favorable pricing curve for PTs compared to counterpart models, thereby maximizing capital efficiency for users.
However, the rapid growth of PTs on Aave introduce new layers of risk. The primary concern stems from the limited and highly concentrated nature of PT liquidity, which resides almost exclusively on the Pendle AMM. In a scenario of market stress or destabilized liquidity, the price discount on PTs could fall, potentially overwhelming the capacity for orderly liquidations. If a price dislocation moves beyond the operational range of the Pendle AMM, the execution of liquidations cannot be guaranteed, creating a risk of temporary or permanent bad debt for the Aave protocol.
This assessment quantifies the extent of these risks by analyzing the current market state and the feasibility of adverse scenarios. We model the amount of collateral that would become eligible for liquidation under various price decline scenarios, estimate the proportion of liquidations that could be successfully executed against available on-chain liquidity, and calculate the potential bad debt that Aave could accrue as a result.
Summary
Our analysis reveals an imbalance between the high concentration of PT collateral on Aave and the limited on-chain liquidity available to service liquidations. For instance, approximately two-thirds of the entire PT-sUSDe-31JUL2025 supply is deposited on Aave, straining the capacity of the underlying Pendle liquidity pools. At the same time, the risk of liquidations is mostly limited to scenarios where PT token price is revised downwards due to the majority of PT borrows being correlated in base price, as all USDe-related assets on Aave are priced on par with USDT.
In particular, we estimate that a PT price decline of $0.12 could trigger up to $7 million in bad debt, an amount that roughly equates to estimated annual revenue generated from PT-backed loans. A more severe but less likely price drop of $0.21 could escalate potential bad debt to $70 million. The mechanism to address these risks is purely operational: the DAO would need to manually adjust oracle’s dynamic discount rate to trigger and allow liquidations, however, if the price would move outside of the Pendle’s AMM pre-defined range, the exact PT price would be hard to infer, needing to resort to the orderbook ranges.
Given that USDe has become a cornerstone asset in these leveraged strategies, both as the underlying for popular PTs and as a major borrowed asset, its stability is intrinsically linked to Aave’s risk exposure. To address this and cover for Pendle’s pricing risks we will perform further risk quantification calculations and evaluate the integration of USDe into Umbrella module during the next iteration of the system’s expansion. This would provide a dedicated framework for managing the specific risks associated with USDe, ensuring that Aave can continue to capitalize on the PT opportunity while proactively mitigating potential systemic vulnerabilities.
Current Scale of PTs
Aave has integrated Pendle Principal Tokens (PTs) as collateral in its Ethereum core market, allowing users to leverage the yields obtained from the PTs. Several PTs are now supported on the Ethereum Core instance, including PT-sUSDE-31JUL2025, PT-eUSDE-14AUG2025, and PT-USDe-31JUL2025, with PT-eUSDE-29MAY2025 having recently matured. This expansion allows users to borrow against their PTs, which are redeemable 1:1 with their respective underlying assets upon maturity. As of June 8th, PTs accounted for a total value of $1.6B, with $680M USDe and $220M sUSDe supply present in addition to the PTs, highlighting Aave’s exposure to Ethena’s protocol and USDe stablecoin.
Source: LlamaRisk, June 8th, 2025
The scale of PT tokens supply has been achieved in a month, expanding these reserves rapidly and using Risk Stewards to initiate multiple rounds of supply cap increases. As the first PT-eUSDE-29MAY2025 matured, the migration of the supply has also been completed successfully, without requiring prolonged capital re-balancing.
It is notable that the expansion of supply is expected as long as the PT yields are sufficiently larger than the stablecoin borrow rates on Aave’s Core market. The liquidity buffer for the blue chip stablecoins is currently sufficient to accommodate further PT inflows, and there remains capacity until the borrow rates reach the non-profitability threshold.
Pricing of PT Tokens
Lending protocols use different mechanisms to price PT tokens on their markets. Until the deployment on Aave, we have seen two main approaches:
- Using a market rate, tracking Pendle Pools as a unique source, applying a Time-Weighted Average Price (TWAP) to control for short-term volatility
- Using a predefined discount rate that approximates PT’s price path as the asset approaches maturity.
Both of these methods come with their respective trade-offs and are preferred under different risk preferences.
Market Rate (TWAP)
This method uses a Time-Weighted Average Price (TWAP) to reflect the PT’s current market valuation. Therefore, this method directly follows the token’s price path, only adjusting for short-term volatility via a custom TWAP window. Both Morpho and Euler use a 900-second TWAP in some of their PT markets.
While this method follows the real asset price trajectory with the highest accuracy, the drawback is the unique source of liquidity (Pendle AMM), which makes the liquidity highly concentrated and susceptible to short-term instability. Due to that, PT market price fluctuations could happen, triggering unexpected liquidations.
Linear Discount Rate
This method addresses the single point of dependency problem, disregarding the market price of an asset. Linear discount rate method prices PTs using a fixed, pre-defined baseDiscountPerYear. This makes the discount reduce linearly as the PT approaches maturity. The formula is:
PT_Price = 1 - (baseDiscountPerYear * TimeToMaturity)
This results in a linearly increasing price path for the PT asset with respect to maturity and asymptotically reflects the PT asset’s price evolution. This approach is mainly used by Morpho Spark vaults.
One crucial consideration for this pricing method is the setup of the baseDiscountPerYear parameter, which needs to reflect the real discount rate within safe bounds, without aggressively over-pricing the asset. This value is particularly hard to tune as the real discount rate fluctuates based on underlying market conditions and the market’s balance.
Dynamic Linear Discount Rate (Aave’s Approach)
Aave addresses the problem of the constant discount rate and prices PTs using a linear formula:
PT_Price_Aave_Oracle = 1 - (configured_discountRatePerYear * TimeToMaturity).
The configured_discountRatePerYear is not static; it can be updated if the PT’s market price deviates by a certain threshold from the PT_Price_Aave_Oracle, or via a heartbeat mechanism. This makes the mechanism more adaptable to the PT price fluctuations in PT liquidity pools, but introduces the need for careful supervision.
An important aspect of Aave’s approach is the maxDiscountRatePerYear. This parameter caps the discount rate, thereby setting a floor for the PT oracle price. This is intended to reflect the maximum implied APY that the Pendle AMM can support, as SY token composition in those pools cannot exceed 96%. It is notable that there was a historical instance where PT token pool, namely PT-USDO+±30Jan2025, hit its maximum implied APY on the AMM but traded on the Pendle order book at an even higher implied APY (i.e., a lower PT price) than what the AMM reflected.
Source: Chaos Labs, May 29th, 2025
Important Note on Oracle Operation: Currently, the dynamic rate adjustment mechanism for Aave’s PT oracles is not active. Price updates for the
configured_discountRatePerYearmust be implemented manually.
While the dynamic linear discount rate oracle was selected, we have discussed other possible pricing methods more extensively, also covering the trade-offs in a previous governance forum discussion. The final DAO’s consensus is that Aave’s critical pricing mechanisms should be separating the pricing infrastructure providers from risk service functions, which is what the dynamic linear discount rate oracle partially adheres to.
Comparison
Even a light difference in the reported prices of each pricing method can impact the demand and utility significantly. Aave, as mentioned before, uses the Dynamic Linear Discount; Euler, the Market Rate; and for Morpho, different curators use different oracles. It can be observed that for the PT-sUSDE-31JUL2025, PT prices reported by Aave’s oracle reflect the market price closely, albeit with a slight over-pricing of not more than 20 bps. This, combined with the deterministic future price path expectation, makes Aave an attractive venue among users willing to perform PT yield leveraging.
Source: LlamaRisk, Jun 6th, 2025
While Euler and Morpho markets using TWAP of PT’s market price are closely related, it’s important to note that a key difference is that Morpho has a fixed USDe pricing where USDe equals to a constant 1 USD value, while Euler uses the actual market rate for USDe/USD ratio. Additionally, the price from Morpho’s linear discount oracle, which is used in one of the PT markets is lower than the others due to employing a higher Discount Rate.
PT Exposure & Risks
Currently, there is a notably high exposure to PTs on Aave’s Core market. For example, when PT-sUSDe-31JUL2025 was listed on Aave it had an initial supply of around $120M and approximately $50M in liquidity on Pendle. As of Jun 8, its supply has grown to around $1600M ($1.6B) with $125M in AMM liquidity on Pendle, changing the supply-to-liquidity ratio from 2.4 to 12.8. In the same period, the Aave supply cap for this PT has increased from $85M to $1.300M ($1.3B), with roughly 3/4 of the PT’s total supply now deposited on Aave.
PT-eUSDE-14AUG2025 and PT-USDe-31JUL2025 contain a similarly high percentage of their total supply deposited on Aave. However, these PT assets exhibit less aggressive supply-to-liquidity ratios of 2.7 and 30, respectively, considering solely the AMM liquidity.
The elevated supply concentration in conjunction with high supply-to-liquidity ratios necessitates a further look into Aave’s ability to liquidate positions if PT prices fall to their oracle-defined minimums (i.e., when maxDiscountRatePerYear is hit). While these minimums do not represent the lowest possible value at which the PT tokens would trade, these thresholds represent moderate scenarios that could materialize under stressful market conditions, possibly amplified by temporary USDe secondary market price de-pegs.
To clearly quantify the risks PTs introduce, it is important to estimate:
- The amount of collateral that would enter liquidation as PT prices drop.
- The cumulative liquidity (AMM and orderbook) available and how it changes as prices move down, to assess if these positions could be liquidated effectively.
- The amount of bad debt that would occur, also taking successful liquidations into account.
Collateral at Risk
We first inspect the current loan distribution in order to estimate collateral at risk and the total leverage of the system. As expected, the majority of borrowers perform yield looping at the most aggressive levels, making the loans susceptible to liquidation risk if the PT token’s reported price on Aave needs to be adjusted downwards as the PT market price decreases. Nonetheless, a small part of the borrowers remain cautious, without leveraging the PTs within the maximal LTV threshold.
At the same time, the predefined maxDiscountRatePerYear provides an anchor that suggests the minimal expected AMM price for the PT assets and therefore provides a clearer view on the feasibility of the different projected downturns. Moreover, PTs with longer maturities have the minimum price set to lower levels, which also means that the liquidity is distributed over a wider price range. This reflects the maturity vs. stability trade-off, where long-term maturities present a longer-term fixed-rate opportunity while exhibiting more volatility early in the PT pool’s deployment.
PT-sUSDe-31 JUL 2025:
For PT-sUSDe, due to a relatively low maxDiscountRatePerYear, implying a higher floor price, and users adopting more conservative leverage, fewer positions would be eligible for liquidation within the oracle’s minimum theoretical price. However, successfully liquidating the positions that would become eligible might be challenging given the liquidity distribution, which gets even more complicated as the market price (suggested by the orderbook) would surpass the theoretical minimum limits.
Source: LlamaRisk, June 3rd, 2025
PT-eUSDe-14 AUG 2025:
As for PT-eUSDe, this maturity pool has only been bootstrapped very recently, with the majority of users migrating from the older May maturity pool. This pool being of longer maturity, the projected minimal theoretical price stands lower, meaning that the majority of supply could become susceptible to liquidations as these theoretical limits would be reached. In particular, this demonstrates a larger risk of successfully performing liquidations if they become needed.
Source: LlamaRisk, June 3rd, 2025
PT-USDe-31 JUL 2025:
We observe a similar result with PT-USDe loan distribution, except that the strong uptick in liquidatable value would happen earlier into the price decline, making the majority of the supply liquidatable at a mere $0.011 PT market price drop.
Source: LlamaRisk, June 3rd, 2025
Overall, an approximate estimate, assuming a best-case scenario where no external liquidity is withdrawn and Aave is the sole venue processing these liquidations, suggests the following amounts of collateral would be liquidated if PT prices drop to their respective oracle-defined minimums:
- PT-sUSDe-31JUL2025: $40M.
- PT-USDe-31JUL2025: $20M.
- PT-eUSDE-14AUG2025: $60M.
While the Value-at-Risk figures are important, it is crucial to understand the context behind the events that could cause a PT price decline. Normal price fluctuations are driven by changing expectations of future yield. A scenario could arise where a sharp increase in expected future returns causes the current PT price to fall as it adjusts to reflect a higher implied yield. This type of event is often self-correcting, as the attractive yield can spur new demand and lead to a price rebound. Moreover, risk naturally diminishes as any PT approaches its maturity date and its price converges toward its redemption value.
More direct risks stem from instability in the underlying assets themselves, such as distress in the exchange rate between sUSDe and USDe, which would negatively impact the market value of the corresponding PT. The most complex scenario involves a depeg of USDe. Due to Aave’s current oracle configuration, which values USDe on par with USDT, a depeg would not directly trigger liquidations for the large part of these leveraged positions. However, it would likely cause a loss of confidence, prompting a mass exit from USDe-based PT positions on the open market. It is this resulting crash in the PT’s market price, not the USDe depeg itself, that would trigger liquidations and expose the protocol to the bad debt scenarios.
Potential Bad Debt by PT Token
The potential bad debt that Aave could accrue if no liquidations occur as the price of the PT token falls below users’ liquidation thresholds. It is observed that the bad debt would not directly start to accrue as the oracles would reach theoretical minimal prices. Instead, the accumulation of bad debt would start with more aggressive PT price downturns, leaving a risk buffer for these asset reserves in the short term.
Source: LlamaRisk, June 6th, 2025
In reality, the overall bad debt that would be incurred depends on the ratio of the liquidations that would still be executed, which in turn depends on the available liquidity and liquidator’s true risk aversion.
No Liquidations Scenario
Under a conservative scenario where no liquidations would effectively happen, the potential total bad debt from PT-sUSDe-31JUL2025, PT-USDe-31JUL2025, and PT-eUSDE-14AUG2025, assuming co-movement in all PT prices, would start to accrue at price downturns of at least 0.17$. This would require significant market stress or even solvency issues to materialize before such a price downturn could happen.
Source: LlamaRisk, June 3rd, 2025
Factoring in Potential Liquidations
Considering the liquidations that would adjust for the available PT liquidity, the overall bad debt could be reduced by up to $200M. This estimation is also assuming co-movement in all PT prices. In reality, while all three assets (USDe, eUSDe, sUSDe) are almost perfectly correlated, their respective PT pools could experience different downturns.
Source: LlamaRisk, June 3rd, 2025
Risk/Reward of PTs for Aave
From a portfolio management perspective, which is also important in light of Umbrella’s module deployment, it is useful to evaluate whether having the current amount of PTs as collateral is financially beneficial for Aave. For that, the revenue generated from PT assets needs to be compared against the potential risks.
Revenue Generation:
To quantify the financial benefit to the Aave DAO, we can estimate the annual revenue generated from loans collateralized by PTs. This revenue is derived from the interest paid by borrowers, a portion of which is directed to the treasury via the Reserve Factor (RF).
Currently, approximately $1.2 billion in stablecoins is borrowed against PT collateral. This is composed of $145 million in USDe, which has a 25% RF, and $1.04 billion in other stablecoins (USDT, USDC, etc.), which have a 10% RF. Assuming a conservative average borrow APY of 5% across all assets, an estimated total annual revenue for Aave’s Treasury is approximately $7 million from this activity.
Comparing Revenue Against Potential Risk
This steady stream of revenue must be weighed against the potential for sudden, significant losses from bad debt. Our bad debt analysis provides a dual comparison:
- A moderate price decline of just ($0.12) across the PT assets could result in approximately ($7) million of bad debt. This potential loss is striking as it is roughly equivalent to the entire estimated annual revenue generated from these positions. In this scenario, a single adverse market event would effectively erase a full year’s worth of earnings.
- In a more severe but still plausible scenario, a price drop of ($0.21) could lead to ($70) million in bad debt. To put this figure in perspective, it would take the Aave DAO approximately a decade to cover such a loss at the current revenue rate.
This highlights an asymmetry in the risk profile, where the potential for a one-time loss could outweigh the steady, incremental gains.
Risk Outlook
While this analysis reveals a significant risk concentration and a challenging risk/reward profile, it is important to contextualize these findings. The identified risks, while considerable, do not pose an immediate systemic threat to the overall safety and solvency of the Aave Core market. The potential bad debt is concentrated within this specific asset class. Given the scale and diversification of the entire Aave protocol, along with the protection afforded by the Safety Module, the protocol is well-capitalized to withstand such a shock. Rather, these findings highlight a specific edge scenario vulnerability that requires more conservative future risk management.
Growing USDe Importance
Our further analysis has revealed that USDe has emerged as a prominent stablecoin for borrowing against PTs, particularly for PT-USDe-31JULY2025 and PT-eUSDE-14AUG2025. This preference may be due to a slightly better LTV and LT for USDe compared to other stablecoins when used with these PTs.
Source: LlamaRisk, June 3rd, 2025
At the same time, USDT remains the most popular choice as a borrowed asset, specifically due to the unique property of the USDe being priced at 1:1 ratio with USDT market price. While this allows higher efficiency for users, it results in Aave bearing the risk of USDe de-pegs.
As PT tokens gain a larger market share, this dependence on USDe’s peg would continue to increase, therefore requiring more careful risk management.
Recommmendations
Given USDe’s importance within Aave, particularly its role in leveraged PT strategies and as a significant borrowed asset, it would be rational to consider USDe’s inclusion to Umbrella module as the next iteration is being considered. This, in conjunction with existing USDT and USDC coverage, would allow to cover the PT asset risks, while allowing the PT markets to continue operating at aggressive supply-to-liquidity levels.
We will continue monitoring the Aave market changes introduced by Risk Stewards and remain cautious to changes in PT asset liquidity on Pendle, as market stress would require promptly adjusting the discount rates of PT price oracles on Aave if such a situation arises.
Disclaimer
This review was independently prepared by LlamaRisk, a community-led decentralized organization funded in part by the Aave DAO. LlamaRisk is an Ethena’s Risk Committee member and an independent attestor of Ethena’s PoR. LlamaRisk did not receive 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.