Rayyan vs. Aave (and Compound)
Aave and Compound are structurally unusable for a quarter of the world's population, and structurally more expensive for anyone who values cost certainty.
The one-table version
| Aave / Compound V3 | Rayyan | |
|---|---|---|
| Instrument | Interest-bearing loan | Murabaha (cost-plus sale) |
| Rate | Variable, changes every block | Locked at borrow for the life of the position |
| Accrual | Compounds continuously | Linear, never compounds |
| Maximum cost | Unknown until you repay | Contract price, disclosed before you confirm |
| Early exit | Repay principal + whatever accrued | Repay principal + accrued markup; unearned markup forgiven (ibra') |
| Term | Open-ended | 360-day settlement (repay anytime before) |
| Rate protection | Buy a rate cap elsewhere ($100K–300K on a $10M/3yr loan) | Embedded, the fixed rate IS the cap; ibra' is the floor leg |
| Rate discovery | Utilization curve (variable) | Same curve, sampled once, at your borrow |
| Rate slippage/MEV | None (curve-based), unlike AMM-based fixed-rate venues | None; curve-based discovery retained |
| End-of-term handling | n/a (open-ended) | Keeper settlement as the borrower's appointed agent, for 0.25% of the amount settled |
| Liquidation | Health-factor based; the liquidator keeps a bonus | Same health-factor trigger; the borrower is credited fair value less the buyer's discount and keeps the unearned-markup rebate; the protocol keeps nothing |
| LP yield | Floating supply APR | Streaming markup share, same balance-grows-every-block UX |
| Shariah status | Impermissible (riba) | Structured as murabaha/mudarabah end to end |
What a borrow costs
A $1M borrow, 360 days, both protocols quoting "5%" on day one:
| Scenario | Compound V3 (variable, compounding) | Rayyan (locked, linear) |
|---|---|---|
| Rates stay at 5% | $51,271 | $50,000 |
| Rates average 9.01% (Aave USDC's actual 2024 average) | $94,174 | $50,000 |
| Rates stay at 10% | $105,171 | $100,000 (if locked at 10%) |
The second row is the one that matters. The Rayyan borrower who locked 5% and lived through a 2024-style rate environment saved $44,174, not because of clever management, but because nothing about their position could change. On Aave or Compound, the number on the screen when you borrow is a spot price. On Rayyan, it's your contract.
The zero-cost collar, spelled out
Institutional borrowers on floating-rate venues who want cost certainty buy two products from a bank: a rate cap (protection if rates rise) and value the option to refinance (participation if rates fall). A structured collar on a $10M loan runs six figures in premium.
Every Rayyan position has both legs built in:
- Cap: your locked rate cannot rise. If utilization spikes and new borrowers pay 12%, you still pay what you locked.
- Floor participation: if rates fall, close your position (paying only accrued markup, ibra'), and re-borrow at the new lower rate. Cost of the round trip: gas.
This is not a promotional feature; it falls out of the murabaha mechanics. A sale price is fixed (cap), and Islamic law expects unearned profit to be rebated on early settlement (floor). Interest-based protocols cannot copy this without rebuilding their core accrual, since a variable-rate lender has no "unearned" component to forgive.
The LP side
Aave/Compound LPs earn a floating supply rate, 2.63% on Compound V3 USDC at recent readings, with $370M+ deposited at that level. Rayyan LPs earn a share of streaming, pre-contracted markup. Three differences that matter:
- Predictability of the book. Every dollar of future yield on Rayyan is a fixed claim written into an existing contract, not a projection of floating rates.
- Launch policy runs hotter. Protocol-owned liquidity holds utilization in a 65–72% band and LPs take 85% of markup under the mudarabah split, putting LP APY in the 3.9–4.7% range at the band (see Rate Model & Yield Policy).
- Nobody profits from non-payment. On Aave, liquidation bonuses go to liquidators and protocol fees to the treasury. On Rayyan a liquidation's surplus returns to the borrower, the settlement fee is the operator's agency wage, and neither is depositor income, so LP yield is markup alone.
One caveat in both directions: all pooled-lending LP yield, Rayyan's included, depends on borrowing demand, and bears bad-debt tail risk if collateral gaps through liquidation margins. Rayyan does not change that physics; see Security & Risk.
And the fixed-rate field: Midnight, Fira, Notional
Fixed-rate DeFi is no longer a niche, Morpho's Midnight (whitepaper May 2026) is the incumbent lending ecosystem betting that fixed rates are how DeFi moves beyond crypto. It's the strongest validation our secondary thesis could get, and the sharpest contrast to our design:
| Midnight (Morpho) | Rayyan | |
|---|---|---|
| Primitive | Tradeable zero-coupon credit/debt units; rate = the discount they trade at | Murabaha sale contracts in one pool |
| Execution | Off-chain maker offers, capital sourced at fill via callbacks | Always-on pool, any size, any block, posted curve rate |
| Early exit | Buy your debt back at market price, rates fell? Exiting costs you the difference, and needs a counterparty | Principal + accrued, always (ibra'), protocol-guaranteed, zero cost, no counterparty needed |
| Maturities | Fixed calendar dates; liquidity fragments across them | Rolling 360-day tenor; one pool is the whole market |
| Shariah | Discounted debt trading (bay' al-dayn), structurally non-compliant, unfixable | Compliant by architecture |
The early-exit row is the product difference that matters: on Midnight you exit at market; on Rayyan you exit at cost. That asymmetry, free exit when rates fall, locked cost when they rise, is the zero-cost collar, and a mark-to-market venue structurally cannot offer it. Notional (peaked $975M TVL) and Fira ($450M+ at launch, 2026) prove the demand; both, like Midnight, fail Shariah review, the entire fixed-rate field competes for our secondary market while leaving our primary market untouched.
What Rayyan kept from Compound V3
Collateral management, health factor, Chainlink oracles, the utilization rate curve, pause guardian, ERC-20 accounting, unmodified; absorb() liquidation, unmodified except that it credits the borrower against the amount owed after ibra'. The design bet is that ~240 changed lines on a five-year audited codebase carries less risk than 12,000 lines of novel fixed-rate machinery (the Notional lesson), and that curve-based rate discovery avoids the rate slippage, MEV, and liquidity fragmentation across maturities that AMM-based discovery introduces.