DeFi strategy case study · prepared for Veris · 19 August 2026
Three positions cleared both screens, on three unrelated drivers: insurance underwriting, trader losses and a levered fixed-rate spread. Each is sized to the depth its venue shows today, and stated at the net yield it pays after every cost. Strategies 01 and 03 each carry a live model on their pages.
Portfolio construction
Strategies 01 and 03 are both principal-token loops. The resemblance is structural only.
| Position | Underlying | Who pays the return | What impairs it | Venue · chain |
|---|---|---|---|---|
| 01 · PT-ONyc | ONyc — reinsurance NAV | Insurance underwriting | Catastrophe claims, reserve losses | Loopscale · Solana |
| 02 · JLP | Jupiter liquidity pool | Trader losses and fees | Sustained trader profitability, borrow capacity | Kamino + Hyperliquid |
| 03 · PT-sUSDe | PT-sUSDe — fixed to 22 Oct | PT yield less USDe borrow | A rising USDe borrow rate, Ethena impairment | Pendle + Aave V3 · Plasma |
Every size, capacity ceiling and available-liquidity figure in this book was read from venue state on 18–19 August 2026 and treated as immediately available at that moment.
01 · Strategy
PT-ONyc · Loopscale · Solana · PT issued by Exponent. The underlying is regulated reinsurance premia originated by OnRe, so underwriting performance pays the return. Bought at 0.9901, redeeming at par into ONyc on 10 September, financed by a fixed-rate fixed-term loan matched on an order book. Both legs are fixed at entry: the PT redeems at par on 10 September, and the loan is written at a fixed rate for a fixed term. The 610bp entry spread is therefore set before the position opens. A 50%-of-book discipline supports $1.65M of equity at 1.50x.
02 · Strategy
Kamino Multiply · Solana · hedged on Hyperliquid. JLP is the house. It is the counterparty to every position on Jupiter Perps, and its NAV carries three things: the price move on its basket, a 75% share of platform fees, and the pool’s P&L against traders. Lever it on Kamino, short the basket on Hyperliquid, and the price move is gone. The two components left both rise in a downturn: traders are typically net long, so a sell-off transfers their losses to the pool, and volatility lifts trading and borrow volume, so fee revenue climbs at the same moment. Kamino’s published Multiply APY cleared 100%+ at 2.4x to 6.2x in both drawdown windows of the last six months, against 10–20% when volatility compressed. That figure is unhedged, yield-only and struck at higher multipliers than this position runs, so it evidences direction rather than return: the strategy’s own measured carry is +9.32% over the trailing year.
This is the most profitable structure in the book when the regime turns, and the least interesting when it has not. Today it has not: traders are not net losing, the Kamino reserve sits at 92.5% utilisation and USDC borrows at 6.41%. So the position is built, measured and idle, ready to open on short notice when the market breaks. And the standby return is not the weak one in the book: at 8.53% the hedged carry sits 62bp below Strategy 03, which runs eight turns of leverage to get there. What pays the return matters more here than how much is borrowed.
Those buffers are what set the leverage. This is the one structure in the book levered on two venues at once, and the legs liquidate in opposite directions — a 49.6% fall in JLP closes the Kamino loop, a 37.0% basket rally closes the Hyperliquid short — while the gain that would cure either margin call sits trapped on the other chain. So the position runs a deliberately low gear: 1.75x against the 6.2x the venue permits and the 2.88x the cohort averages, accepting a slightly lower return to keep both buffers wide.
Timing is what blocks this today. Four conditions gate funding and none is met today — the regime, reserve utilisation, the borrow rate and a proven operational cycle, each set out on the fact sheet and in the memo. None is a forecast, and the first three reverse in the same event: a sell-off pushes traders into losses, lifts fee volume, and typically drains leverage demand from the borrow reserve at the same time. The fourth is the one inside our control: a $500k pilot to prove the operational cycle now, so that when the regime turns the only remaining decision is size.
03 · Strategy
PT-sUSDe · Aave V3 · Plasma · PT issued by Pendle. A recursive fixed-rate carry. Buy the principal token, post it as collateral, borrow USDe against it, stake that back into sUSDe, buy more PT, and repeat to 8.00x. Entry runs through sUSDe rather than USDe because sUSDe is the Pendle pool’s base asset, so the PT is bought with minimal slippage. The return is the spread between the PT’s fixed 3.87% yield and the floating 2.80% USDe borrow rate: $16.0M of PT earning 3.87% against $14.0M of debt costing 2.80% gives 11.36% gross on $2.0M of equity. The asset leg is locked to 22 October; the funding leg floats, and that is the position’s largest single exposure. Aave marks the collateral through a protocol-owned linear discount adapter rather than Pendle spot, so the mark accretes faster than the debt accrues and the health factor improves every day the position is held.
The oracle is what makes the leverage tolerable: Aave reads a protocol-owned PendlePriceCapAdapter that discounts to par at 3.97% a year, marking the PT 2.4bp above market. Execution is evidenced at size, with a live quote absorbing $12.4M of PT for 1.9bp of fee and 4bp of price impact; the $16.0M position is worked in clips above that single fill. Public Pendle depth caps practical sizing near $2.8M of equity, well before either Aave limit binds — the remaining PT supply cap would fund about $5.56M at maximum leverage. Entry is the slow leg. Staking USDe into sUSDe is instant, but the PT is accumulated over several clips and every turn stakes, buys and re-deposits before the next. The unwind is comparatively quick: the matured PT is redeemed into USDe, which repays the Aave debt directly, so closing the position takes a few interactions rather than a timed process. The position runs at 8.00x rather than the 10.70x the E-Mode bucket permits: maximum leverage would add 2.2 points of return for 3.4 points of liquidation buffer, on collateral that cannot be sold quickly.
04 · Deliverables
Each position ships with a fact sheet and a due diligence memo carrying the full risk register and dimension scoring.
PT-ONyc on Loopscale. Paid by insurance underwriting. Five principal risks, all scored 4; capacity binds at $1.65M of equity.
Levered JLP hedged on Hyperliquid. Paid by trader losses and fees, both of which rise in a downturn. Measured against live venue state; on hold at $1.0M until the four funding conditions are met; the fact sheet and memo carry the same four.
PT-sUSDe on Aave V3 Plasma. Paid by the spread over USDe funding. Six principal risks; oracle read on-chain, execution evidenced by a live quote at size.