The AUSD rollover trap and Pendle liquidity risks

Liquidity walls and Aave collateral

Traders face a liquidity wall when they attempt to roll over Pendle positions. On October 2, 2026, the December 17 AUSD principal-token market held only $1.61 million in liquidity. This small amount contrasts with the $67.4 million in PT-AUSD tokens that users supplied to Aave V3 Monad markets as collateral for the October 8 expiry. If you attempt to migrate your collateral from an expiring market to a new one, you will encounter significant price impact. Many users rely on the fixed yield to stay profitable, but the slippage from moving tens of millions into a pool with only $1.61 million in liquidity will likely consume the entire spread. Active AUSD loans on Monad jumped 113% to $8.7 million in just 15 days before the expiry. On August 31, 2026, LlamaRisk found that the 18 largest suppliers on Aave carried debt with a median health factor of 1.02. These borrowers hold debt in USDC, GHO, USDe, and USDT0. Aave initially launched the collateral market with a 20 million-token supply cap, but users filled it quickly, forcing recommendations to increase the cap to 40 million and then 80 million tokens. The total capital in the October market grew far beyond the initial 20 million token supply cap. Will the December market scale fast enough to absorb this $67.4 million of collateral?

Comparing yield venues and fee shifts

Spectra competes for liquidity by using Curve’s Stableswap pools. It grew its TVL from $20 million to $100 million over the last 12 months. While Pendle uses a custom AMM to price time-decaying assets, Spectra builds its markets on specialized infrastructure. Term Finance also addresses fixed-yield needs through auction-based lending instead of splitting assets into PT and YT. Pendle’s market dynamics changed recently when the platform updated its fee structure on September 18, 2026. The new model targets 10% of projected yield and has a maximum cap of 10 basis points. This replaced the previous flat fee of 5 basis points. Pendle’s V2 system decouples the SY, YT-PT, Market, and sPENDLE components to reduce bugs. The PENDLE governance token uses a vote-escrow model where holders lock tokens for up to two years. This model is winding down as sPENDLE replaces it. Under the new system, 80% of Pendle V2 Yield and Swap fees fund PENDLE buybacks. These repurchased tokens go to active sPENDLE holders every two weeks rather than to weekly lockers.

The December PT implied yield is 5.64%. This rate stays below the borrowing rates for several stablecoins. Borrowing USDC costs 6.09%, while mUSD costs 4.28% and GHO costs 4.64%. You must watch these spreads because the incentive that pushed the December rate to 6.64% is temporary.

Maturity mismatch and interest rate risks

The maturity mismatch in fixed-rate positions creates a trap for unsuspecting traders. When interest rates rise, the market value of existing bonds falls. I see traders who ignore the sensitivity of zero-coupon bonds to interest rate swings. A PT buyer locks a rate, but the market price of the PT fluctuates based on prevailing yields. If you buy PT and interest rates rise, the value of your position drops. You must hold the PT until maturity to receive the 1:1 redemption. Pendle’s TVL reflects this volatility, as it dropped from $2.08 billion at the end of Q1 2026 to $1.24 billion as of September 15, 2026. PT holders also take on the smart contract and custody risks of the underlying assets like sUSDe. If sUSDe depegs at expiry, the PT redeems to a depegged asset rather than par value. Traders who buy $10,000 of PT-sUSDe at a price of 0.94 receive 10,638 units, which yields a 6.4% locked APY if they hold the position until the March 2026 expiry. In contrast, a $10,000 outlay for YT at 0.06 controls yield on $166,000 of underlying, providing roughly 16x exposure to the yield rate. YT holders benefit from points farming, where they receive points multiplied by their YT notional, a 10x to 30x amplification on their capital. However, if Ethena pays 3% yield while the YT implies 25%, the buyer loses money because the YT only pays out the actual yield the underlying produces. If realized yield underperforms implied yield, YT decays to zero faster than the yield stream replenishes it. This asymmetric loss means YT behaves like a yield call option with a 100% downside cap.