TL;DR
- What it is: ERC-4626 Morpho Vault V2 (
0xbeef…0f51,steakUSDC), curated by Steakhouse Financial, lending USDC across 8 Morpho Blue markets — but almost the entire book sits in one cbBTC/USDC market. ~$78.7M TVL, fees 0, no emissions. - Yield source: fully organic borrower interest — a credit-risk premium on overcollateralized cbBTC loans; with no reward stream, the net headline is what a redemption pays.
- Biggest risk: that cbBTC market prices cbBTC off Chainlink BTC/USD with no cbBTC-specific feed, so a cbBTC depeg is invisible to liquidations and socializes as bad debt to suppliers (Risk #1).
How it works
USDC deposits mint steakUSDC (18-dec, non-rebasing); NAV grows as borrower interest accrues. A Vault V2 routes through one adapter (0x2b96…8290) into the immutable Morpho Blue singleton (0xBBBB…FFCb). 8 markets are funded, but the exposure is one: cbBTC 92.9% ($73.1M), then weETH 3.5%, WETH 1.5%, wstETH 2.1%; idle USDC = 0. Named “Prime USDC,” it reads diversified; on-chain it is a single-market cbBTC-credit position (~93% BTC-collateral) with a multi-market wrapper.
Oracle. The vault only lends, so it is never liquidated — its NAV absorbs bad debt if a borrower’s collateral gaps below the loan. The cbBTC market (0x64d6…cc64, 86% LLTV) prices cbBTC off Chainlink BTC/USD (0xF403…E88c) — no cbBTC leg, and read with no staleness check; the ETH sleeves use exchange-rate feeds. Unlike the Base sibling, this vault is only ~26% of a $281M market shared across ~130 suppliers, so it neither sets nor sole-funds it.
Yield. One leg enters NAV — borrower interest, cbBTC supply averaging ~3.7% (spot ~4.3% today) at ~90% utilization; with fees at 0, net = gross. There is no emission stream (rewards empty — yield fully organic), and both fees are 0, so all borrower interest reaches depositors.
Risks
Verified on-chain (2026-07-23, block 25,594,758): allocation, LLTVs, utilization, the cbBTC oracle feed, fees (=0), owner/curator thresholds, per-selector timelocks, the Supervisor TIMELOCK (14d) + guardian, sentinel/allocator sets, idle=0, instant capacity, no steakUSDC AMM pool. Taken on trust: cbBTC 1:1 custody, Steakhouse’s zero-bad-debt record, the fee/penalty ceilings (5%/50%/2%, getters revert), and the guardian’s/extra sentinel’s internal control (contracts whose signer sets aren’t readable).
| # | Risk | Mechanism | Source + our delta |
|---|---|---|---|
| 1 | cbBTC-as-BTC oracle blind spot | cbBTC oracle 0xA6D6…182a reads Chainlink BTC/USD only (no cbBTC feed); a cbBTC depeg is invisible to liquidations → bad debt to suppliers. 92.9% of the vault; markets immutable, so unpatchable. Exit: cbBTC↔BTC spread >200bps >1h. | On-chain feed read; NOCA-original. Same blind spot as the Base/PYUSD siblings. |
| 2 | ~93% single-collateral concentration | $73.1M against cbBTC in one market; the ETH sleeves (~7%) are the only diversification. Any cbBTC/BTC drawdown hits nearly the whole book at once. | On-chain 92.9% cbBTC (block 25,594,758); NOCA-original — a single-collateral share this high needs no external benchmark. |
| 3 | Instant yield seizure by the 2/7 curator | setManagementFee/setPerformanceFee are curator-submitted with 0 timelock → the 2/7 curator can turn on mgmt ≤5%/yr + perf ≤50% instantly (a fee needs its recipient set first — both curator, 0-timelock). Yield-only; fees 0 today. Exit: any fee raise. | On-chain timelock + Morpho source. Milder than PYUSD (1/1 Safe) but bypasses the owner’s 14d Supervisor. |
| 4 | Liquidity-gated exit on contested depth | Instant ≈ $27.8M (~35%) is almost all the cbBTC market’s unborrowed liquidity — shared across the market’s ~130 suppliers, not reserved; peer redemptions compete for it. Exit: instant ratio <15% or cbBTC util >98%. | On-chain. NOCA delta: unlike a sole-supplier vault, our accessible depth shrinks when peers exit. |
| 5 | Bad-debt gap-down tail at 86% LLTV | A fast BTC gap-down can outrun permissionless liquidations; no loss reserve, and 90% utilization leaves a thin repayment buffer. | Standard Morpho Blue tail. No Steakhouse bad debt to date and pps strictly monotonic — directional, not a guarantee. |
| 6 | Young, un-crisis-tested capital | TVL sat ~$10M through mid-Apr 2026, ramped to ~$54M by early May, plateaued, then hit $78.7M this week — most is ~3 months old, top $13M days old. Exit: >15% single-day TVL drop with cbBTC util >95%. | On-chain historical totalAssets; NOCA-original. Isolated-market architecture is the mitigant, not a retention record. |
Key-compromise threshold
Morpho Blue’s core is immutable, the adapter registry (0x3696c5eA…) constrains adapters to vetted types, and all four ERC-4626 gates are 0x0 (imposing a withdrawal gate carries a 7-day timelock).
- Funds (theft / principal loss): no instant path. Diverting principal needs the 2/7 curator (
0x827e…ECdB) to raise a cap or add a vetted adapter on the 7-day vault timelock — which the sentinels (0x7630…b715+ Supervisor) can revoke — then an allocator to move funds; realistic harm is concentration, not theft. Owner-level changes route through the VaultV2Supervisor’s 14-day timelock + guardian (0x0272…79ec) veto, driven by the 5/10 Safe (0x0A0e…f8DD). All 7 curator signers are a subset of the 10 owner signers — the Safes are not independent. - Liveness / economic: 2/7 curator, instant — set fees to the 5%/50% ceiling (yield seizure). Or 1 allocator key — incl. a hot EOA bot (
0xfeed46…) — reallocates into full markets to stall instant redemptions (soft freeze;redeem()is otherwise permissionless, no pause). - Single point of failure: economic — yes (2/7 curator seizes yield; one allocator degrades instant liquidity). Funds — no.
Liquidity & exit
Primary path is ERC-4626 redeem at NAV. totalAssets() = 78,702,037 USDC, idle = 0, convertToAssets(1e18) = 1.025878 (block 25,594,758). Instant capacity ≈ $27.8M (~35% of TVL), essentially all the cbBTC market’s $27.8M unborrowed liquidity — but that depth is shared across the market’s ~130 suppliers, so under peer stress our realizable share is a fraction of it. The remaining ~$51M needs borrower repayment, fresh supply, or forceDeallocate (penalty 0.001%, itself needing market liquidity). No secondary AMM for steakUSDC (Uniswap V3 all tiers + V2 empty).
Defensible size at current liquidity (post-haircut): $7M — ~25% of the $27.8M instant depth, clears same-day at NAV; under a 3× stress haircut (→ ~$9.3M) $7M still fits a 1-day window. We size to instant liquidity, held conservative given contested depth and untested retention. Stress (“exit 50% in 7d while peers redeem”): $3.5M against ~$9.3M is comfortable; a full $7M with peer redemptions and cbBTC util >98% queues the residual for repayment cycles. The binding constraint is Risk #1, not depth. Halve if not refreshed by 2026-10-23.
External reviews
Routed sources checked (per rubric.md): Exponential — no grade for this vault; DeFiScan — rates the Morpho protocol at Stage 0 (protocol-level, not this vault). No independent vault-specific rating is relied on here — LlamaRisk / Bluechip have none, and Steakhouse (curator) is conflicted (discount per rubric.md). Per rubric.md, >90d (⚠️) is stale.
- DeFiScan — Morpho protocol Stage 0 (governance/permissions; protocol-level, not this vault) · accessed 2026-07-23
- Morpho — Vault V2 concept & security model · accessed 2026-07-23
- Morpho — protocol & adapter audits · audit library
- Siblings:
steakhouse-usdc-prime-instant(Base V1),morpho-paypal-usd-main(Vault V2) · internal