A $320,000 Pendle Trade Wave Sparked $36.4 Million in Morpho Liquidations

A roughly $320,000 burst of trading on Pendle early Tuesday ended up cascading into about $36.4 million in liquidations on Morpho—without any rule being broken. On August 25, a wallet tagged 0x854e…690d executed 11 back-to-back trades on Pendle between 04:28 and 04:37 UTC. The trader swapped about $320,000 of SYreUSD into more than 9.5 million YTreUSD. The rapid buying spree lifted the implied annual yield in the PTreUSD/YTreUSD market above 20%, briefly knocked PTreUSD down by roughly 3%, and pushed leveraged borrowers into liquidation territory on Morpho. Pendle's mechanics sit at the center of the move. Depositing a yield-bearing asset on Pendle splits it into a Principal Token (PT) and a Yield Token (YT). PT tracks the principal redeemable at maturity, while YT represents the yield stream until that date. Their prices move inversely: when YT rises, PT tends to fall, and vice versa. By aggressively bidding up YTreUSD, the wallet increased implied yields and simultaneously pressured PTreUSD lower. That price drop hit Morpho users who had posted PTreUSD as collateral. With loan-to-value ratios around 91.5% before the trades, a ~3% decline was enough to push positions underwater and activate Morpho's automated liquidation engine. The liquidation tally included 33 events across roughly 19 to 20 borrower positions. Liquidators seized more than 38 million PTreUSD as collateral and repaid about $35.19 million in USDC debt plus around $960,000 in USDT debt. In total, approximately $36.4 million in positions were liquidated. One key detail: the episode did not generate bad debt. Collateral levels were sufficient to cover outstanding loans, leaving the protocol solvent. The setup was unusually fresh. Pendle launched a USDC vault on Morpho around August 4—about three weeks before the incident. The vault quickly drew more than $15 million in deposits, much of it routed into PTreUSD markets. Ahead of the trades, implied yields for PTreUSD were near 11%, with maturity set for December 10, 2026. At those levels, borrowing against PTreUSD was appealing, but an LTV near 91.5% leaves less than a 10% cushion before liquidation risk escalates. Oracle design amplified the fragility. Pendle's PTreUSD oracle takes the lower of two inputs: a 15-minute market average or a fixed discount curve that implies roughly a 6% annual discount. In a market with limited liquidity, 11 trades in under nine minutes can still move the 15-minute average meaningfully. Pendle said it is reviewing oracle configurations following the incident. The figures highlight the leverage of thin markets: for every dollar spent by the trader, about $114 in positions were liquidated. The broader takeaway is a familiar DeFi tradeoff. Oracles must reflect market prices to function, yet market prices can be moved by sufficient capital—especially when liquidity is shallow. The risk grows when collateral is a derivative instrument like a PT token rather than a deep, spot asset such as ETH or USDC. In this case, buying yield tokens raised implied yields, pushed down principal token prices, and triggered liquidations on a separate lending protocol—a cross-protocol contagion path designers will need to model more explicitly. Longer oracle windows can reduce manipulation risk but increase stale pricing. Shorter windows track reality more closely, but become more vulnerable to fast, concentrated trading.