I tracked Pendle through a 90% TVL crash ($13.4B → $1.36B in six months). What happened next was remarkable.

Everyone was watching TVL move. Meanwhile, fees were telling a completely different story.
Q4 2025: TVL dropped 41.7%, but fees only fell 15.7%.
That's a massive gap. And the reason why explains how Pendle actually makes money.
So here's the thing: pool maturations force people to make a choice. When your position expires, you either exit or roll into a new maturity.
Every exit? That's a swap fee. @pendle_fi literally makes money on both sides of volatility. Movement in either direction = revenue.
Then Q1 2026 brought a different challenge.
DeFi yields compressed significantly. Monthly fees went from $4.44M in August 2025 down to $552K by March 2026. The counter-cyclical model showed its boundary: extended calm means less repositioning, which means lower fees.
But here's where it gets interesting.
During this period, the team executed a complete structural upgrade:
sPENDLE (launched Jan 2026):
🔸 80% of all fees → PENDLE buybacks
🔸 Buybacks now exceed emissions 4.4x
🔸 Emissions cut ~30% via algorithmic model
🔸 Net deflationary even at minimum revenue
🔸 Staking went from 20% to 57.7% of supply
Boros (funding rate trading):
🔸 $12B+ notional volume since launch
🔸 New revenue stream independent of DeFi yields
🔸 Taps the $63B perps market (currently 0.1% penetration)
RWA Integration:
🔸 $151M TVL built from near zero in 8 months
🔸 Paxos USDG, Apollo - institutional heavyweights as direct participants
🔸 PT tokens work as Aave V3 collateral (creates recursive demand)
The previous vePENDLE had 20% holder participation and many underperforming pools.
New model sends 80% of fees straight to buybacks. Net deflationary. Token holders benefit directly from protocol revenue.
Current snapshot (DefiLlama, May 2026): $1.59B TVL, $10.5M annualized fees, ~$320M market cap.
That's a 0.20 MC/TVL ratio. Pendle is 28x larger than its closest competitor.
They command 50-60% of the entire yield tokenization market. The moats are substantial: custom AMM built for time-decaying assets, deep Aave/Morpho integration, $23B in limit order volume (45% of total activity).
What emerged from this period is a fundamentally different protocol.
Evolution from single-revenue DeFi protocol to multi-vector infrastructure: yield tokenization + derivatives + institutional RWAs. Stayed profitable throughout.
The protocol at $1.36B has completely different economics now: buyback-aligned tokenomics, multiple uncorrelated revenue streams, institutional infrastructure, deflationary even at minimum revenue levels.
Heading into the next cycle:
🔸 Market position strengthened through the entire transition
🔸 Institutional adoption accelerating through RWAs
🔸 Revenue diversified across multiple independent streams
🔸 Demonstrated resilience through major TVL movement while maintaining profitability
The transformation is complete and the opportunity is still early.
Data: @DefiLlama