Morpho Paypal USD Main

approvedethereumreviewed 2026-07-27

TL;DR

How it works

PYUSD deposits mint senPYUSDmain; NAV grows as borrower interest accrues. Vault V2 routes through one adapter (0xc2A1…34d4) into the immutable Morpho Blue singleton (0xBBBB…FFCb) — 10 markets hold caps, 8 a funded position. Allocation (2026-07-27): kBTC 32.0% · syrupUSDC 21.4% · LBTC 14.8% · weETH 14.6% · sUSDe 9.9% · cbBTC/WBTC/wstETH 1.3% · idle 6.1%. Labelled a PayPal USD product; the exposure is a leveraged credit book — 48% BTC-wrappers, 31% yield-bearing stables, 15% ETH-LST.

Oracle setup. The vault only lends and is never itself liquidated; its NAV absorbs bad debt when collateral gaps below a loan. All eight markets quote collateral in PYUSD/USD (Chainlink 0x8f1d…8FB1) — 16h stale at our snapshot, so a PYUSD depeg reaches every liquidation engine late. kBTC prices off bare BTC/USD; syrupUSDC and sUSDe off ERC-4626 convertToAssets × a peg feed; the LSTs off exchange rates; LBTC off Redstone.

Yield. Borrower interest is the only leg entering NAV — 2.7–4.1% by market, 1.89% after the 1% fee. Demand is utility-driven but reflexive: AdaptiveCurve rates track utilisation, not risk, so kBTC pays below syrupUSDC. The 4.98% headline adds 3.09pp of off-NAV Merkl distribution — 62% of the total, none of it share price. Size on the 1.89%. Capital is rented: April 2026 redeemed 63% of shares in 82 hours (Risk R2).

Risks

Verified on-chain (2026-07-27, block 25,623,650): positions, caps, LLTVs, utilisation, all eight oracle feeds, fees, owner/curator getOwners/getThreshold + codesize per signer, 23 per-selector timelocks plus gate/registry abdications, sentinel/allocator sets, the full Accept config history, totalAssets/totalSupply across the April 2026 drawdown, the Merkl funding tx and funder balances, top-holder claims, Sky’s IlkRegistry, Spark’s PYUSD holdings. Taken on trust: Paxos attestation, Sentora’s allocation intent, audit scope, the Vault V2 fee/penalty ceilings (5% / 50% / 2% — constants whose getters revert), collateral-issuer solvency, and the reward funder’s legal identity.

IDRiskMechanismSource + our delta
R1kBTC oracle blind spot + 32% concentrationLargest market (0xe51f…eef5, 32% of TVL) prices kBTC off Chainlink BTC/USD with no kBTC leg. A Kraken-custody depeg is invisible to liquidations at 86% LLTV → bad debt, unpatchable (markets immutable). Exit: kBTC↔BTC spread >200bps >1h.On-chain oracle read; NOCA-original. Sharper this quarter: cbBTC and WBTC both carry wrapper-specific feed legs in the same vault, so kBTC’s omission is a choice, not a market-wide norm. Concentration up from 29% on 2026-07-21.
R262% of the yield is a weekly subsidy, and the capital it rents has already run3.09 of 4.98pp is an off-NAV Merkl distribution pre-funded 7 days at a time; targetAPR cut 6.0%→5.7%→5.5% since 2026-07-02. April 2026 shed 63% of shares in 82 hours at a rising share price. Exit: two weeks with no campaign, or targetAPR ≤5%.On-chain funding tx, 32-campaign history, historical totalAssets/totalSupply (study) — NOCA-original. Funder is a 1-of-1 Safe whose budget traces to primary PYUSD issuance; visible balance $7.01M ≈ 19 weeks at current spend. Rewards do pay — top holders have claimed 98–99.7% of accrual within hours. The defect is tenor, not delivery.
R3Recursive yield-collateral credit at 91.5% LLTV — and it loops back to a position we may hold directlysyrupUSDC (21.4%) + sUSDe (9.9%) = 31% of TVL: PYUSD lent against Maple’s mark-to-model NAV and Ethena’s peg. This vault is 61.5% of all syrupUSDC-collateralised Morpho lending, priced off Pool.convertToAssets(). Exit: syrupUSDC/USDe deviation >2% for >24h.Maple credit risk per our maple-syrupusdc strategy; Ethena short-vol carry. Surface re-traced on-chain today (study) — share unchanged at 61.5%. Treat a Maple markdown as an immediate bad-debt event here; a loss hits this and any direct syrupUSDC holding at once.
R4⚠ FIRED — single-key curator/owner, instant fee seizure, and a veto that barely existsOwner (0xe8C9…8008) and curator (0x9e39…aac0) are each 1-of-1 Safes of a bare EOA; fee setters carry 0 timelock → 5% + 50% instantly. Fired: the curator raised the kBTC cap to $120M on 2026-07-13. Exit: any fee raise or new Accept.On-chain event log, isSentinel, per-selector timelock/abdication reads. The 2-sentinel barrier does not hold: one sentinel is the curator Safe, the other (0x74D1…1618) has nonce 0, and setIsSentinel/setCurator are both 0-timelock. New since the 2026-07-21 review: an adapter to Sentora’s own MetaMorpho V1 vault removed 2026-07-02; forceDeallocatePenalty set to 0.01% on 2026-07-20 — the lever is live and raisable.
R5Liquidity-gated exit, and the buffer is shrinkingInstant capacity $48.3M = 15.3% of TVL (idle $19.1M + $29.2M unborrowed), down from 18.4% on 2026-07-21 while TVL fell only 2.4%. Markets run ~90% utilised; the other 85% needs borrower repayment or forceDeallocate. Exit: instant ratio <10%.On-chain 2026-07-27. Tighter than a queue-based MetaMorpho — the vault is ~100% of supply in every market. Counterweight: April 2026 proved $268M could clear in 82 hours at NAV.
R6PYUSD has no Sky peg backstop — and one depositor is a third of the vaultSky’s IlkRegistry (35 ilks) holds no PYUSD collateral type; the Paxos↔Maker PSM on record is PSM-PAX-A, gem USDP, at a zero debt ceiling. The closest thing that ever existed — Spark’s own $100M PYUSD↔USDS Curve pool (A=10000, 0.001% fee, Spark sole LP) — was withdrawn to zero on 2026-06-25; its Savings PYUSD vault drained −96% in July. The remaining $237M inventory is discretionary allocation, not convertibility, and Spark has now twice shown it exits. Exit: SparkLend PYUSD supply −50% in 30d.IlkRegistry enumeration + Spark contract reads (study) — NOCA-original, and it corrects a widely repeated claim. The 1:1 LitePSM route often cited for PYUSD is gemmed in USDC, so par access runs PYUSD→USDC→USDS and is bounded by PYUSD/USDC market depth — ~1bp at $10M, no guarantee at any size. One holder is 36.4% of TVL, the top two 53.9%. Also pre-authorised: an sUSDS collateral cap at uint128 max since 2026-05-18, with no market cap to fund it yet.

Key-compromise threshold

Morpho Blue’s core is immutable and withdrawals cannot be frozen: three of four gates read abdicated() = true, the survivor covers only deposits, and setAdapterRegistry is abdicated so the registry is frozen to Morpho-vetted types. Privilege sits in two single-signer Sentora Safes, re-verified today: owner 0xe8C9…8008 is 1-of-1 over EOA 0x295D…d7E9, curator 0x9e39…aac0 is 1-of-1 over EOA 0x9925…0ABE, both signers codesize 0.

Liquidity & exit

Primary path is ERC-4626 redeem at NAV. totalAssets() = 315,672,691 PYUSD, idle PYUSD.balanceOf(vault) = 19,098,732 (2026-07-27, block 25,623,650). Instant capacity = idle + Σ min(vault position, market unborrowed) across the 8 funded markets = $48.3M (15.3% of TVL); the remaining ~$267M needs borrower repayment, fresh supply, or forceDeallocate — an in-kind exit that itself needs market liquidity. No secondary market for senPYUSDmain.

Defensible size at current liquidity (post-haircut): $12M — 25% of the $48.3M instant depth, clearing same-day at NAV; under a 3× stress haircut (→ ~$16.1M available) $12M still fits a one-day window. Cut from $15M because instant depth fell 19% in six days while TVL fell 2.4%. We size to instant liquidity, never the $316M TVL. Stress scenario (“NOCA exits 50% in 7d while peers redeem”): $6M against ~$16.1M post-haircut is comfortable. April 2026 is the honest counterweight — $268M cleared in 82 hours at a rising share price — but that was an orderly unwind of loops, not a credit event, and every market now runs ~90% utilised. Aggregate with any direct Maple position (Risk R3). Halve if not refreshed by 2026-10-27.

External reviews

No independent risk rating (LlamaRisk / Chaos Labs / Bluechip / DeFiScan / Exponential) of this Vault V2 was located — it is new (deployed 2025-12-16), and Sentora, the curator, is conflicted (discount per rubric.md). Anything >90d is stale (⚠️).