Pendle. Mention it on DeFi Twitter twelve months ago and you got "points farming" as the reply. Mention it now and the answer has shifted. PTs, the protocol's fixed-yield tokens, are quietly listed as collateral on Aave, Morpho and Spark. The carry trade looks clean on a spreadsheet. The panda has been reading the parameters, and the spreadsheet leaves a few things out.
What is a Pendle PT, exactly?
Pendle splits any yield-bearing asset into two pieces. Take a position in sUSDe, Ethena's staked synthetic dollar. Deposit it in a Pendle market. You receive two tokens: a Principal Token (PT) and a Yield Token (YT). The PT is what you redeem at maturity. The YT collects the variable yield until then.
A PT-sUSDe maturing on a fixed date trades at a discount to sUSDe. Hold it to maturity and you redeem one sUSDe per PT. That discount, annualised, is the fixed yield. According to Pendle's official documentation, this turns variable APYs into something closer to a zero-coupon bond.
That makes a PT structurally legible to lending markets in a way that a rebasing LST or a points-laden LRT never was. There is a known maturity date, a known redemption asset, a known payoff. Lending protocols can model a discount curve. Risk teams can quote a haircut. The product fits a category that already exists in TradFi, just rebuilt on chain.
The mechanism: a clean carry, on paper
The trade is straightforward. A user buys PT-sUSDe at a discount. They deposit it as collateral in Morpho or Aave. They borrow USDC or USDe against it. They swap the borrow back into PT-sUSDe and loop. By maturity, the fixed yield on the PT is higher than the borrow rate, and the spread becomes the profit.
According to DefiLlama's Pendle protocol page, Pendle sits among the top non-stablecoin DeFi protocols by TVL, with the largest pools denominated in sUSDe, weETH variants and USDS. A non-trivial share of that TVL is collateralised somewhere else, which is the point.
The mechanic relies on three numbers being aligned. The PT discount must be wider than the lending borrow rate. The loop must survive the collateral haircut. And the PT's maturity must not collide with a rate move that flips the spread negative.
Morpho's curator-driven vaults, covered in our note on the Morpho curator economy, are well suited to this kind of product. Each vault sets its own collateral parameters. Each curator chooses which PT maturities and which haircuts to accept. Risk is priced per vault, not globally. That separation matters when the underlying yield asset changes shape, as it tends to do.
Why are lending markets accepting them?
Because the alternative collateral options are getting thinner. LRTs sold a points story in 2024 and 2025. That story has cooled, as covered in our reading of EtherFi's post-points yield stack. Naked ETH staking yield is in the low single digits. Memecoin collateral is too volatile for serious credit lines.
PTs offer something different: known duration, known counterparty asset, and a yield that mostly comes from a single, auditable source. Ethena's USDe, which underpins most large Pendle markets today, sources its yield from delta-neutral basis trades. The funding-rate exposure is real, as we noted in our view on USDe's funding-rate test, but at least the source is named.
According to DefiLlama's Ethereum chain dashboard, Ethereum's DeFi TVL stood at $37.61 billion on June 9, 2026. The slice that flows through PT-collateralised loops is a meaningful piece of the credit market on Ethereum mainnet. Quiet, but structural. The DeFi pillar round-up gives the wider context for how these credit primitives are stitching together.
Risks
Several things can break the carry. The list matters because the headline yield does not include any of them.
Smart-contract risk lives in three places at once: Pendle, the lending protocol, and the underlying yield asset. A bug in any of them propagates. Pendle's PT market itself is well audited, but a leveraged loop multiplies any failure mode in the layers below.
Depeg risk is the headline. If sUSDe loses its peg, the PT-sUSDe collateral reprices on chain at the worst possible moment. Liquidations cascade. Borrow positions get unwound at unfavourable curves. The carry trade flips into a loss before the maturity date ever arrives.
Oracle risk is subtler. The PT price feed used by lending markets is not the spot market price. It is usually a discount-rate model. If the model lags actual market repricing during stress, undercollateralised positions sit on the books for hours before the liquidation engine catches them.
Governance risk runs through every layer. Pendle's vePENDLE controls fee distribution and market emissions. Morpho's curators control vault parameters. Aave's DAO controls listing and risk parameters. A vote in any one of those changes the cost basis of an outstanding loop overnight.
Funding-rate risk applies specifically to USDe-backed PTs. When perpetual funding rates flip negative and stay there, the underlying yield compresses. A PT held to maturity still pays, but the next maturity prices in the new reality.
What to watch next
Three numbers are worth watching. The Pendle TVL share routed through lending collateral. The number of Morpho vaults whose curators have whitelisted PT maturities. And the USDe funding-rate trend. If any one of them inverts, the loop starts unwinding from the leveraged tail first.
Pendle did the unflashy thing. It built a primitive that fits inside the rest of DeFi instead of demanding a new category. The lending markets noticed. The yield-trade desks noticed. The risk reports are still catching up.
The panda continues to read the parameters. No moonshots, no rugs to call. Just a structurally interesting piece of plumbing earning its keep, with risks that are not yet priced into the headline APY.
