Comparison · yield aggregator
Pendle vs Convex: Yield Trading vs Yield Aggregation
Side-by-side breakdown of Pendle and Convex on what they do, who they're built for, fee structures, and risk profiles. Two yield protocols, very different jobs.
2026-05-19
Verdict at a glance
| Use case | Pick |
|---|---|
| Locking in a fixed yield on a stablecoin or LST | Pendle Pendle's PT (principal token) is the only on-chain mechanism for fixed-rate exposure. |
| Maximum boosted CRV yield on Curve LP positions | Convex Finance Convex's permanent veCRV lock gives boosted rewards without depositors needing to lock CRV themselves. |
| Speculation on future yield going up | Pendle Pendle YT (yield token) is the only liquid leveraged bet on yield direction. |
| Set-and-forget Curve pool deposit | Convex Finance Convex auto-stakes, auto-claims, auto-compounds; depositor never touches gauges. |
| Capturing points / airdrop campaigns | Pendle Pendle YT positions compress months of points farming into a single trade. |
| Lowest layer of additional protocol risk | Either Both add one smart-contract layer; differences are operational not security. |
Live data on each protocol's TVL, supported chains, and pool list: Pendle · Convex Finance
What you’re comparing
Despite both being major DeFi protocols built around yield, Pendle and Convex answer completely different questions.
Pendle is a marketplace for trading yield itself. It splits any yield-bearing asset into two tokens: a Principal Token (PT) that redeems for the original asset at maturity, and a Yield Token (YT) that captures all yield accrued from now until maturity. Buying PT locks in a fixed return; buying YT is a leveraged speculation on yield going up. Markets exist for major LSTs (stETH, weETH, ezETH), restaking tokens, real-world-asset stables (sUSDS, sUSDe), and “points” campaigns from emerging projects.
Convex is a yield aggregator built on top of Curve. Depositors give Convex their Curve LP tokens; Convex stakes them into Curve’s gauge system on the depositor’s behalf, claims the CRV emissions, votes via its own permanent veCRV lock for the highest-emitting pools, and redirects the boosted rewards back to depositors. The depositor earns trading fees plus boosted CRV plus CVX plus any partner emissions without needing to lock CRV themselves.
They’re not alternatives. They’re complementary tools - many sophisticated DeFi positions use both (e.g. Pendle wrapping a Convex deposit as the underlying yield-bearing asset to trade the future yield separately from the principal).
Pendle: yield as a tradable asset
Pendle’s core insight is that yield-bearing assets contain two distinct economic claims: a claim to the principal (which exists regardless of yield) and a claim to the yield stream (which depends on the underlying’s performance). On most DeFi protocols these two claims are bundled in a single token. Pendle separates them.
If you hold 1 ETH worth of stETH:
- PT-stETH entitles the holder to redeem for 1 ETH worth of stETH at maturity
- YT-stETH entitles the holder to all staking rewards generated until maturity
Markets exist for each token. Buying PT at a discount locks in a fixed rate (you know exactly what you’ll receive at maturity). Buying YT is a leveraged bet that future staking yield will exceed the YT’s purchase price. Selling YT and keeping PT lets you continue holding the principal while monetizing the yield stream upfront.
This makes Pendle the closest analog DeFi has to a bond market. The “implied yield” at any given moment is what the market thinks the yield asset will produce until maturity; it’s published on Pendle’s UI and updates in real time as PT and YT prices move.
The dominant use cases as of mid-2026:
- Fixed-rate stablecoin yield: buy PT-sUSDS or PT-sUSDe at a discount, hold to maturity, redeem at par. Functions as a fixed-income instrument.
- Points speculation: many new protocols (Ether.fi, Ethena, Resolv, EtherFi staking variants) distribute points before their token launches. Pendle YT positions compress months of points farming into a single trade. High-volatility bet on whether the eventual airdrop justifies YT entry price.
- Long-yield-direction trading: if you believe staking yield is heading up, buy YT-stETH and effectively go long Ethereum staking returns with leverage.
Convex: yield maximization on Curve
Convex’s core insight is that Curve’s veCRV system creates an aggregation arbitrage. Locking CRV for four years to maximize your Curve LP rewards is something most LPs don’t want to do (it locks their capital up). Convex does the locking once, permanently, for everyone - and shares the boosted rewards across all Convex-routed deposits.
The mechanism:
- Depositor sends Curve LP tokens to Convex.
- Convex stakes those LP tokens into Curve’s gauge system, claiming the full veCRV-boosted CRV emissions on the depositor’s behalf.
- Convex’s permanent veCRV lock votes weekly for the highest-emitting pools (or, more precisely, the pools its vlCVX holders direct it to vote for).
- Boosted CRV + partner emissions + CVX (Convex’s own token) are streamed back to the depositor.
Net to the depositor, yield is roughly 1.5-2x what they’d earn depositing into Curve directly with no CRV lock. The trade-off is one more contract layer (Convex on top of Curve) and exposure to CRV emission schedule changes.
Risk surface
Pendle’s risk profile is dominated by the underlying yield assets it wraps. A depeg in stETH or sUSDe propagates immediately through every Pendle market that uses them. The protocol’s own contracts have been audited extensively (Spearbit, ChainSecurity, Ackee, WatchPug) and have no exploits since launch. The most material historical incident was a frontend hijack in April 2024 (malicious approval prompt) that some users approved before the team patched it.
Convex’s risk passes through from Curve almost entirely. If Curve has a problem (Vyper compiler bug, gauge manipulation, governance attack), Convex depositors are affected first. Convex’s own contracts have been live since 2021 with no exploits on its booster or vlCVX systems.
In both cases the additional protocol layer is well-audited and battle-tested, but it’s still an additional layer. For position sizes above $100k, the marginal yield benefit of either should be weighed against the extra smart-contract risk.
Fees
Pendle takes a small fee on YT-to-yield conversion (typically 3% of yield in current markets). PT redemptions are fee-free at maturity. The protocol’s revenue model is the YT spread, not LP fees.
Convex takes 17% of CRV emissions and partner rewards: 16% goes to vlCVX holders, 1% to Convex DAO operations. The boost more than compensates for the fee in most market conditions - if your LP would earn 4% CRV without lock and 8% CRV with full veCRV boost, the Convex-routed position earns ~6.5% net after fees, comfortably above the no-lock rate.
When to pick which
Pick Pendle if you want fixed-rate exposure (PT positions function as on-chain bonds), if you want to speculate on yield direction with leverage (YT), or if you want to participate in points campaigns through a single trade rather than long-term holding.
Pick Convex if you have Curve LP exposure (or want it) and want maximum yield without managing veCRV locking yourself. Convex is essentially the default routing for any Curve LP that doesn’t have a strong opinion about gauge voting.
Combine both for sophisticated yield strategies. A common construction: deposit USDC into a stablecoin Curve pool, route through Convex for boosted yield, then use the Convex receipt as collateral in a Pendle market to lock in the fixed-rate portion of the yield. Each protocol adds value at a different layer.
Live data on each protocol’s TVL, current yields, and supported markets: Pendle · Convex Finance.
Reader Q&A
› Do Pendle and Convex actually compete with each other?
Not really - they solve different problems. Pendle is a yield-trading marketplace for converting variable yields into fixed-rate or leveraged-yield instruments. Convex is a yield aggregator that maximizes Curve LP rewards through pooled veCRV voting. The closest overlap is Pendle's PT-on-Convex markets, where Pendle wraps Convex positions as the underlying yield asset - which is composition, not competition.
› If I want a fixed 5% yield, how does Pendle deliver that?
You buy PT-sUSDS (or PT on any other supported yield asset) at a discount. If the underlying redeems at 1.000 at maturity and you bought PT for 0.978, your locked-in yield is the gap divided by time. The yield is fixed because the redemption value is fixed; what moves is the underlying's actual yield, which is the YT holder's variable. Pendle is structurally the on-chain equivalent of a zero-coupon bond market.
› Why would I use Convex instead of depositing into Curve directly?
Three reasons. First, the boost - Convex permanently holds one of the largest veCRV positions, which gives every Convex-routed Curve LP the maximum gauge boost without requiring the depositor to lock their own CRV. Second, the aggregation of partner emissions - Convex farms most of Curve's whitelisted reward tokens automatically. Third, governance via vlCVX - if you want voting weight in Curve's gauge wars, vlCVX is the dominant route.
› How do Pendle 'points' markets work?
Many new protocols distribute 'points' off-chain before launching their token. Holding the underlying asset (e.g. eETH from Ether.fi, or sUSDe from Ethena) accumulates points. Pendle lets you split that asset into PT (no points exposure, fixed-rate yield) and YT (all the points, plus the variable yield). Buying YT compresses months of points-farming into a leveraged bet on the eventual airdrop being more valuable than the YT's market price. It's high-volatility - whether the airdrop pays off vs YT entry price is the entire question.
› Which one is older and more battle-tested?
Convex (2021) is older and has been through more market stress. The protocol's largest historical risk passed through from Curve - the July 2023 Vyper compiler exploit drained Convex-routed pools, with affected depositors compensated through Curve's governance process. Pendle (2021 on Ethereum but redesigned for V2 in 2022) has had no contract-level exploits and one notable frontend hijack in April 2024 (a malicious approval prompt, mitigated by alert users and quickly patched). Both are blue-chip-audited; neither has had on-chain exploit losses to date.