Credit · Fixed Income · PENDLE
Pendle tokenizes yield-bearing assets into Principal Tokens (PT) and Yield Tokens (YT), creating an on-chain market for fixed-rate yield and yield speculation/hedging. Pendle V2's specialized AMM and vote-escrowed vePENDLE coordinate liquidity across dozens of maturity-dated yield markets, and Boros extends the model to funding-rate trading.
Tokenized yield: split principal and yield (PT/YT).
PENDLE price
$1.34
-0.4% 24h
Latest data · 15 min delay
Typed risks scored by severity, likelihood and impact, with the incident record behind them.
42
severity-weighted points across 14 typed risks · 4 critical
assessed Jul 2026Market
12· 4 risks · 1 critical
Technological
15· 5 risks · 1 critical
Counterparty
7· 2 risks · 1 critical
Governance
6· 2 risks · 1 critical
Regulatory
2· 1 risk
Every score is recomputed at render time from the typed risks below; nothing is hand-assigned. Each risk contributes its severity weight (critical 4 · high 3 · medium 2 · low 1) to its category; the headline number is the sum across all categories. Bars scale to this entity's highest category score. Scores are not comparable across tags because a permissionless-market protocol concentrates technological risk while an institutional-credit protocol concentrates counterparty and regulatory risk, and a single league table would misrepresent both.
Market 4 × → 12
Technological 5 × → 15
Counterparty 2 × → 7
Governance 2 × → 6
Regulatory 1 × → 2
Pendle's own core contracts have never been drained, and PendleMarketV3 is immutable apart from pausing, which is a real strength. The risk lives everywhere else. Governance is concentrated in a 2 of 4 multisig and a single guardian key with no timelock, which was useful in September 2024 when Pendle paused everything within a day of the Penpie exploit and protected around 105 million USD, but the same power could be turned the other way. The bigger systemic issue is that PTs became collateral across Spark, Morpho, Aave and others, priced by a discount oracle that Chaos Labs showed can drift far enough for borrowers to walk away with debt, on markets running liquidation LTVs above 91%. Add underlying concentration in sUSDe and USD0++, issuers who can change redemption terms overnight, liquidity that exists only on Pendle's own AMM, and TVL that has fallen to under 1 billion USD from a 6.5 billion USD peak, and the profile is a technically sound protocol carrying other people's risks. Treat Pendle exposure as exposure to the specific underlying and to the lending market accepting the , not to Pendle in the abstract.
Market4
Technological5
Counterparty2
Governance2
Regulatory1
4 documented incidents · 3 Sep 2024 to 2024 · how the protocol behaved under stress, with sources.
3 Sep 2024
Penpie, a yield booster built on Pendle, was drained for about 27 million USD through a reentrancy bug in its batchHarvestMarketRewards function. The attacker permissionlessly registered a fake Pendle market at 0x0ab305033592E16dB7D8e77d613F8d172a76ddc9 to inflate a staking balance. Penpie had been audited by WatchPug and Zokyo.
DetailsPenpie, a yield booster built on Pendle, was drained for about 27 million USD through a reentrancy bug in its batchHarvestMarketRewards function. The attacker permissionlessly registered a fake Pendle market at 0x0ab305033592E16dB7D8e77d613F8d172a76ddc9 to inflate a staking balance. Penpie had been audited by WatchPug and Zokyo.
Outcome: PNP fell about 40% and PENDLE about 9%. Pendle paused all its contracts and stated its own funds were secure. LlamaRisk records that the pause safeguarded roughly 105 million USD. Exploiter address 0x7a2f4d625fb21f5e51562ce8dc2e722e12a61d1b.
SourceFeb 2025
PT-ENA experienced a five minute price wick down in its oracle feed, an example of the AMM derived TWAP producing a price unrepresentative of fair value.
DetailsPT-ENA experienced a five minute price wick down in its oracle feed, an example of the AMM derived TWAP producing a price unrepresentative of fair value.
Outcome: Cited by Chaos Labs as evidence that PT TWAP feeds can be transiently unreliable, supporting the killswitch and safety margin design.
SourceJan 2025 to Feb 2025
The guaranteed redemption price of USD0++ was changed from 1.00 USD to 0.87 USD, repricing PT-USD0++-27MAR2025 held as collateral in external lending markets. USD0++ implied APY had been capped at 30% in the oracle configuration while the asset traded above 50%.
DetailsThe guaranteed redemption price of USD0++ was changed from 1.00 USD to 0.87 USD, repricing PT-USD0++-27MAR2025 held as collateral in external lending markets. USD0++ implied APY had been capped at 30% in the oracle configuration while the asset traded above 50%.
Outcome: Documented by Chaos Labs as a case demonstrating that PT oracle configurations can be invalidated by unilateral issuer action, contributing to its recommendation against enabling PT borrowing.
Source2024
For PT-weETH-27JUN2024 the Pendle team incorrectly estimated the maximum expected yield used in the market's rate configuration.
DetailsFor PT-weETH-27JUN2024 the Pendle team incorrectly estimated the maximum expected yield used in the market's rate configuration.
Outcome: Documented by Chaos Labs as a parameterisation error affecting PT valuation in downstream lending markets.
SourceEvent-type chips are categorised by event type against the risk taxonomy (a documented presentation mapping, not a dataset assessment). Amounts are the dataset's published figures; missing values were never published.
