Pendle lost 90% of TVL — why can't anyone compete?

Pendle lost 90% of its TVL (Total Value Locked) — from $13.4 billion to $1.39 billion — and is still 28 times larger than its closest competitor. It is the only DeFi protocol that allows separating a yield-bearing asset into two pieces — principal and future yield — and trading each separately. Lock in an 8% rate on stETH for six months. Bet that Ethena's yield will rise. Hedge against falling USDC rates. All on-chain, without an intermediary. In January 2026, it completely reformed its tokenomics: new governance token, new fee distribution, new emission rules. The token fell 82% from its highs.

The 90% drop in its TVL is not just the story of one protocol: it's the story of how much of the yield in DeFi was real capital and how much was points tourists. To understand why it matters — and whether the drop is an opportunity or a sign that this market is smaller than it seemed — you first need to understand what Pendle does.

What is Pendle and how does yield trading work?

Pendle is a permissionless protocol (open, no approval or intermediary needed) that allows separating a yield-bearing asset into two parts: the principal and the future yield. Each part becomes an independent token that can be bought, sold, or held until maturity. It is, in essence, DeFi's fixed-income market.

In traditional finance, this has existed for decades: zero-coupon bonds are principal without a coupon, and Treasury strips separate principal from interest. Pendle does the same, but with DeFi assets — stETH (Ethereum staked with Lido), sUSDe (Ethena's synthetic dollar), aUSDC (USDC deposited in Aave) — and executes it on-chain without a custodian.

The mechanism has three parts:

• SY (Standardized Yield) — a standardized wrapper that converts any yield-bearing asset into a uniform format. stETH, aUSDC, sUSDe: all become SY before entering Pendle. It's the compatibility layer.
• PT (Principal Token) — represents the right to receive the face value of the asset at maturity. If you deposit 1 stETH and separate it, the PT is worth 1 ETH at maturity. Before that date, it trades at a discount — and that discount is the implied rate. If an stETH PT maturing in 6 months trades at 0.96 ETH, you are locking in a yield of ~8% annualized. Buying PT = locking in a rate.
• YT (Yield Token) — represents the right to receive all the yield generated by the underlying asset until maturity. If stETH yields 4% and you hold the YT, you collect that 4%. If the yield rises to 8%, you collect 8%. If it falls to 0%, you collect nothing. The YT converges to zero at maturity — it's a leveraged position on future yield. Buying YT = betting that the yield will rise.

How does Pendle's AMM work and why is it different from Uniswap?

Pendle's AMM (Automated Market Maker) does not concentrate liquidity by price range — it concentrates liquidity by implied yield range. The difference matters: in Uniswap, the LP (liquidity provider) chooses the price range they want to operate in (e.g., ETH between $2,000 and $3,000). In Pendle, the curve automatically adjusts as the PT approaches maturity, because the PT's price converges to face value by design.

This reduces impermanent loss (IL — the loss an LP suffers when the price of the pair moves outside its range) associated with the passage of time. Additionally, the pool pair is PT/SY (both derivatives of the same asset), which maintains high correlation and reduces IL compared to pools of uncorrelated assets like ETH/USDC.

There's a data point that reveals the market's true sophistication: in 2025, limit orders (off-chain signed orders with on-chain settlement) facilitated $23 billion in volume — 45% of Pendle's total. In high-volume pools like USDe and sUSDe, they exceeded 90%. Pendle's liquidity is not just the pool's TVL: it's the combination of AMM + specialized order flow.

Why did Pendle lose 90% of its TVL since 2025?

Pendle's peak was a perfect accident of favorable conditions: high yields (stETH above 5%, Ethena above 20%), points narrative (protocols distributed points redeemable for airdrops to those who deposited in Pendle), and leveraged carry trade — the strategy of borrowing at a low rate and depositing in high-yield pools to keep the difference. When all three conditions compressed, capital exited.

What left: mercenary capital seeking airdrops and points, carry trades that ceased to be profitable when yields fell, and YT speculation that lost meaning with compressed yields. What remained:

The honest question: does capital return? It depends on yields. If staking and lending yields rebound, demand for locking in rates (PT) and betting on yield increases (YT) reactivates. If yields remain compressed, Pendle functions as a niche infrastructure — useful, but with a TVL that reflects a low-rate market, not an aggressive carry economy.

What does sPENDLE change compared to the old vePENDLE?

In January 2026, Pendle replaced vePENDLE (vote-escrowed, the model where you lock tokens to vote and collect) with sPENDLE (staked PENDLE). The change was profound:

The implication for the investor: sPENDLE turns PENDLE into a token with a more direct value flow. 80% of the protocol's fees are used to buy back PENDLE on the open market — it's not a dividend, but it's buying pressure proportional to the protocol's use. The annualized revenue for sPENDLE holders is $7.5 million, with 97.3 million PENDLE staked. Token liquidity improved: you no longer need to lock up for 2 years to participate.

AIM changed the incentive dynamic: the pool that generates the most TVL and fees now wins, not the one that pays the most bribes. New pools receive an initial boost to get started (bootstrapping); afterward, they must demonstrate economic traction. Pendle also co-invests: it contributes $0.40 for every $1 of external incentives in PENDLE, or $0.30 in other tokens.

Which chains does Pendle operate on and where does it have real liquidity?

Ethereum accounts for 64% of TVL and 95% of trading volume. Plasma (a specialized EVM chain) is second with 21% of TVL, but its trading volume is modest. Sonic (Fantom's successor L1) and HyperEVM (Hyperliquid's EVM environment) show nascent TVL. The concentration on Ethereum makes sense: the deepest yield markets (stETH, sUSDe, aUSDC) exist on mainnet — where underlying liquidity is greater.

Tokenized assets on Pendle go beyond ETH staking. As of April 2026, four main categories are traded:

Additionally, Pendle uses LayerZero to enable cross-chain PTs: PTs created on Ethereum can be used as collateral on networks where Pendle does not have full deployment, concentrating deep liquidity on mainnet.

Who competes with Pendle in yield tokenization?

No one, at the relevant scale. And that is both its greatest strength and its greatest cautionary sign.

Pendle's advantage is not just "more TVL": it's the combination of a secondary rate market + principal/coupon tokenization + specialized order flow + integration as collateral in money markets (Aave, Morpho, Spark) + a distribution layer for issuers. Replicating one part is feasible; replicating the complete combination has not been done by anyone.

The caution: a market without real competition can mean two things. Either the product is so good that no one can compete — or the total market is too small to justify more than one player. With an MC/TVL of 0.16, the market seems to lean towards the latter interpretation.

What are the real risks of using Pendle in 2026?

What happens if the underlying asset loses value?

PT and YT inherit the risk of the underlying protocol. If Lido suffers an exploit and stETH loses its peg, stETH PT loses value. If Ethena cannot maintain the USDe peg, sUSDe PT is not worth $1 at maturity. Pendle does not add counterparty risk — but it also does not eliminate it. Official documentation explicitly states: loss in third parties directly affects PT and LP.

What is the risk of negative yield and watermark?

If the underlying asset enters negative yield (generates losses instead of gains), the PT may redeem below expectations. Additionally, Pendle uses a watermark mechanism: the YT does not generate yield again until the asset recovers its previous yield level. This is a Pendle-specific risk that does not exist in a simple staking deposit.

Do Pendle pools have liquidity risk?

Pendle markets have maturity. When a pool expires, LPs must migrate to a new market or withdraw liquidity. Unrolled positions (not migrated to a new maturity) redirect their yield to the treasury — not to the user. This creates an operational risk: if you do not actively manage your positions, you can lose yield.

Has Pendle's core been hacked?

There are no confirmed exploits of Pendle's core as of April 2026. The protocol has multiple audits (six auditors for V2, including ChainSecurity, Spearbit, and Dedaub), a $250,000 bug bounty, and adopted Security Alliance's Safe Harbor to facilitate whitehat rescues. Since V4/V5 factories, they removed the permit() function in new PT/YT/LP tokens as it was a phishing vector.

The most relevant incident was in February 2026: a third-party staking protocol that sent unvalidated data to the Pendle Router. The loss was from the external vault, not the core. But it illustrates the real risk in 2026: the danger is no longer just the Pendle contract, but everything built on top — vaults, wrappers, money markets that use PT as collateral.

Is Pendle permanent infrastructure or a cyclical trade on carry narratives?

Both. And the proportion between the two determines the investment thesis.

The infrastructure argument: Pendle solved a problem that no one else has solved at scale — on-chain yield trading with principal and yield tokenization. sPENDLE improved tokenomics. AIM rationalized emissions. Boros — Pendle's adjacent product for trading funding rates (the rates paid to maintain leveraged positions in perpetual contracts) — has already processed $11.5 billion in notional with a peak of $270 million in open interest. Restaking, yield-bearing stablecoins, and RWAs need rate markets — and Pendle is the only one that offers them.

The cyclical argument: the 90% drop in TVL was not an accident. It was the revelation of how much of Pendle's capital was mercenary — attracted by points, airdrops, and leveraged carry. When conditions normalized, it left. If yields rise again, capital will return — but it will also leave when they fall. The MC/TVL of 0.16 says the market has already learned that lesson.

The most honest thing to say: Pendle is real infrastructure that works best when yields are high and rate volatility exists. In compressed yield environments, it works but with a fraction of the capital. The risk-adjusted return of providing liquidity in Pendle depends on the cycle — and DeFi yield cycles are shorter and more violent than in TradFi. Locking in a rate with PT remains one of the smartest strategies in bear markets. Betting with YT remains one of the riskiest in any market.