Pendle PTs, the Default Collateral for Aave, Morpho, and Euler

In April, we reported that Pendle had lost 90% of its TVL (total value locked) — dropping from $13.4 billion to $1.39 billion — and we asked how much of that capital was real and how much was "points tourism." Nine months later, there is an answer, and it isn't found in the Pendle token: it's on the balance sheets of the three largest DeFi lending markets. Their principal tokens (PT: the portion of a yield position representing only the principal, redeemable 1:1 at maturity) have become the default collateral for Aave, Morpho, and Euler. According to community dashboard aggregates (Dune and similar; there is no official Pendle cutoff), the combined supply of PT across these three protocols has doubled in six months, exceeding $2 billion; including Silo, the total deposited as collateral surpasses $4 billion. On July 1, Aave made it official by launching its Global Dollar Hub with a PT as the sole entry collateral. This article closes the loop on our own coverage — from the crash to the plumbing — and explores what this shift entails: for the first time, DeFi is stacking billions on top of collateral that has an expiration date.

What turned Pendle's 90% collapse into its best filter?

The $13.4 billion TVL peak that Pendle hit in 2025 was built on three fuels that weren't the product itself: abnormally high yields (stETH above 5%, Ethena above 20%), a points and airdrop narrative that rewarded depositing as early as possible, and a leveraged carry trade (borrowing at a low rate to buy the higher-yielding asset) that multiplied positions as long as the spread held. When yields compressed and airdrops were executed, that capital did what mercenary capital always does: it left. The easy takeaway in April was "Pendle is dying." The correct takeaway was different: the withdrawal was draining the tourists and revealing who was using the protocol for its mechanics rather than the incentives.

What remained after the purge —around 1.390 million in TVL in April, approximately 1.095 by mid-July, with real annual fees and organic volume— turned out to be the substrate upon which infrastructure could be built. A collateral does not need explosive growth; it needs predictability. And an instrument that has survived the exit of its speculative capital is, by definition, more predictable than one inflated by incentives. This apparent death acted as a sieve. What survived ceased to be a trendy product and became plumbing: invisible while it works, decisive when it fails.

What exactly is a principal token and why does it work as collateral?

Pendle splits a yield-bearing asset — such as interest-bearing sUSDe or USDG — into two tradable pieces: the YT (Yield Token), which collects all yield until a fixed date, and the PT (Principal Token), which yields nothing but is redeemable 1:1 for the underlying asset at maturity. Buying a PT below par and waiting for maturity is equivalent to locking in an interest rate in advance, much like a zero-coupon bond.

This mechanic is exactly what a lending market wants in a guarantee: the price of a PT converges to par — toward a value of 1 — as maturity approaches, so an oracle can value it with a discount factor that rises linearly to 1.0 on that day, without depending on the volatile price of the AMM (the automated market where it is traded). For Aave, Morpho, or Euler, this means liquidation thresholds that move smoothly rather than erratically. The problem is the word "future": that value is only certain on a specific date.

How much PT collateral is already in Aave, Morpho, and Euler?

Adoption is already visible on the balance sheets. According to community dashboard aggregates and Dune queries — without an official Pendle cutoff — the supply of PT used as collateral across the three protocols has doubled in six months to exceed $2 billion, and the total deposited across all lending markets (including Silo) surpasses $4 billion. Each protocol has absorbed it with a different architecture, and that difference matters when it comes to maturity.

The Morpho figures — over $1 billion in PT, of which approximately $844 million was linked to USDe positions — are a benchmark from late 2025, not a July 2026 cutoff; the order of magnitude remains relevant, but it is advisable to check the protocol dashboard before citing them as the current state. Euler's jump — from $9 million to $119 million in a year and a half — best illustrates the speed: a market that was an experiment in 2025 is now moving nine figures.

What did Aave launch on July 1 with its Global Dollar Hub?

On July 1, 2026, Aave activated its Global Dollar Hub on Ethereum, the first new Liquidity Hub since V4 went into production and, therefore, the first real test of its hub-and-spoke architecture (a shared liquidity core to which specialized markets connect). The entry collateral was a single asset: PT-USDG-24SEP2026, the principal token of a USDG position maturing on September 24, 2026. With it, users can borrow USDC, USDT, or USDG.

USDG is the stablecoin issued by Paxos — through Paxos Digital Singapore, supervised by the Monetary Authority of Singapore as a Major Payments Institution, and Paxos Issuance Europe, MiCA-compliant and supervised by the Finnish FIN-FSA — with a circulation near $3 billion in mid-2026. It anchors the Global Dollar Network, an ecosystem of over 130 corporate partners including Kraken, OKX, and Mastercard. The combination — a stablecoin regulated in two jurisdictions plus a PT as the access key — explains why the launch coincided with Aave's largest growth in new wallets in five years, a surge that the press estimated at around 900% despite the general drop in crypto prices. The Hub started with approximately $14.8 million supplied in its first days; Aave V4 as a whole exceeded $250 million in deposits by July 6 (the Hub is not reported in isolation), partly due to position migration from V3.

Why does collateral with an expiration date change the rules?

Every collateral that Aave, Morpho, or Euler has accepted until now — ETH, wBTC, stablecoins, liquid staking tokens — shares one property: it exists indefinitely. A PT does not. It is the first DeFi collateral with a maturity date written into the contract, and that detail introduces a risk that no other asset on the books carries: synchronized maturity.

The mechanism works like this. A PT-USDG-24SEP2026 does not expire "when the market decides"; it expires on September 24, 2026, simultaneously for everyone holding it. On that day, three things happen at once. First, the PT can no longer be traded on the AMM: an expired PT is no longer exchanged, only redeemed on Pendle for the underlying asset. Second, its value becomes pegged to par and no longer generates anything, while the loan it guarantees continues to accrue interest — anyone who doesn't close or move their position starts losing money by inertia. Third, and this is the delicate part of Aave's design: since the Global Dollar Hub launched with a single PT series as collateral, all the Hub's collateral matures on the same day. There is no natural staggering.

Traditional banking solves this with bond ladders: maturities are staggered so that everything never matures on the same day. A lending market that accepts a single PT series as collateral does the exact opposite, concentrating maturity on one date, and that is the novelty entering Aave's books. Translated into systemic risk: Aave governance must incorporate a successor series — say, a PT-USDG maturing in December — before September 24, and a mass of positions must migrate within a narrow window around that date. If the migration is well-coordinated, nothing happens: the oracle converges smoothly, the collateral is worth exactly what it says, and positions roll over to the next series. If it is poorly coordinated — if a large volume of collateral reaches maturity without being moved — a wall of positions appears that must be unwound on the same day, with exit liquidity concentrated at a single door. It must be said bluntly: this mechanism has not yet been tested at this scale — no public source documents a mass maturity of PT used as collateral comparable to what the Hub faces — so September will be an unknown, not the repetition of a precedent. It isn't a price failure, which is what oracles usually monitor; it's a calendar failure.

How does each protocol handle maturity day?

The good news is that none of the three were caught by surprise: the answer lies in the architecture, and all three are different.

Aave combines a discount oracle with rollover tools. The discount factor rises linearly to 1.0 at maturity and stays anchored to the redemption value, so there are no post-maturity price swings. For migration, there is a single-transaction transfer that moves the entire position — collateral and debt — to a later-maturing PT without fees, leveraging Aave's E-Mode (the efficiency mode that allows high LTV — loan-to-value — ratios between correlated assets). The weak point isn't the price, but the concentration: by launching with a single series, Aave takes on the challenge of orchestrating that bulk migration and ensuring governance has the successor series ready in time.

Morpho avoids this by design by spreading the risk. Each PT lives in an isolated market, with its own pair, its own oracle, and its own date; an adapter wraps the Pendle oracle to speak the Chainlink interface that Morpho requires. Since maturities are spread across many markets with different dates, there is no single "big day": the blast radius of each maturity is contained within its market. The trade-off is liquidity fragmentation and more markets to monitor.

Euler takes this logic to the extreme with its Vault Kit (EVK), which allows for the deployment of permissionless vaults where the creator chooses the oracle, interest rate curve, hooks (custom functions executed at key vault moments), and operational limits. A PT-sUSDe vault with an LTV of 0.75 allows for loops of up to 4x; the person who deploys it is also the one who determines how it behaves at maturity. Maximum flexibility, maximum responsibility for the creator: the protocol does not impose a single response to maturity because it delegates that decision.

There is also a valuation risk introduced by the linear oracle that is worth mentioning: if the market interest rate of the PT rises above what the oracle assumes in its trajectory, the real price of the PT falls below the price reported by the oracle, and the collateral becomes momentarily overvalued by the protocol. This is known as discount rate risk, which is why risk providers (Chaos Labs, RedStone, Llama Risk) have developed oracles that cap the implied yield instead of relying entirely on a straight line. Convergence to par determines where the price will end up at maturity, but the journey to that date remains exposed to interest rate movements.

What is the lesson from the Pendle cycle?

The full arc fits into a short, dated timeline: less than three months pass between our coverage of the crash (April 27) and the closing of this analysis (July 16). The table summarizes the milestones with their dates and reference figures.

The practical lesson for anyone using these positions: collateral with an expiration date requires a calendar discipline that perpetual collateral did not — knowing which PT series backs your loan, when it matures, and whether your protocol will roll it over for you or leave the migration in your hands. In Morpho and Euler, maturity is a per-market issue; in Aave's Hub, it is currently a collective coordination challenge centered around a single date. Monitoring that date is now as much a part of risk management as monitoring the price of the collateral.

Sources and links: CryptoTimes — Aave V4 Global Dollar Hub · CryptoBriefing — PT-USDG as collateral · Pendle Docs — Principal Tokens · Pendle Docs — Linear Discount Oracle · DeFi Saver — PT Rollover on Aave · Euler Docs — Vault Kit (EVK) · Benzinga — Aave wallet growth