Zap DeFi

Three lending products on QOM, each handling a class of asset the other two can't. Pawn prices NFTs by lender judgement, no oracle. Forge is a kinked-IRM money market with a TWAP read from a permanent (LP-burned) ZAP/QOM market, plus a self-deleveraging Bootstrap Loan that auto-routes QNS Staking emissions to retire debt — a primitive that exists nowhere else. Loaded NFT Pool prices staked .qom collateral by the staking contract's own deterministic emission schedule, sidestepping the floor-price-oracle problem that has broken every prior NFT pool.

Three products, one stack

	Pawn — live	Forge — in design	Loaded NFT Pool — planned
Borrow against	Any QOM-chain NFT	ZAP collateral (Direct) / a fresh .qom (Bootstrap)	Staked, reward-bearing .qom
Pricing model	Bilateral lender judgement	30-min TWAP from a permanent on-chain market	Read directly from QNS Staking emission schedule
Liquidation	None — forfeit on default	Permissionless, partial, with grace window	Permissionless, bounded by future emissions
Counterparty	A specific lender	Pool of QOM lenders	Pool of QOM lenders
Closest external analog	NFTfi / Arcade	Aave v3 (Direct), Alchemix SRENS (Bootstrap)	Ajna (oracle-less), nothing for self-pricing

These aren't three takes on the same idea — they're three lending models, each chosen because the others can't underwrite that asset class.

• Pawn handles the long tail — random NFTs without oracles, by making each loan a bilateral price-discovery event.
• Forge handles the liquid base — ZAP, with a TWAP whose underlying market is structurally permanent.
• Loaded NFT Pool handles structured collateral — .qom domains carrying a deterministic future ZAP-emission stream the contract can value without any external feed.

What's structurally different about this stack

1. The TWAP reads a permanent market

The ZAP/QOM pool on QomX has its LP tokens burned. The pool can never be unwound; its liquidity can only grow. The oracle has no off-chain feed, no signer committee, no rug surface above the contract level — as long as QOM L1 produces blocks, the oracle produces prices.

Single-source / shallow-liquidity oracles are the dominant 2024–25 attack vector — price manipulation alone accounts for ≈17% of major DeFi exploits and over USD 165M of losses since 2023. A TWAP whose underlying liquidity cannot be removed by anyone, including the team, eliminates the LP-removal class entirely.

2. A loan that pays itself back

Forge's Bootstrap Loan composes pool lending with QNS Staking lien hooks. The protocol mints a fresh .qom, stakes it into QNS Staking with a lien, and auto-routes the resulting ZAP emissions back to retire the borrower's debt. The closest live analog at scale is Alchemix's self-repaying loans (yield-source-deleveraged CDPs); the closest NFT analog is Alchemix SRENS, which auto-renews ENS names — but pays from a generic Alchemix CDP, not from the domain's own cash flow.

Domain-collateral pool lending whose debt service comes from the collateral's own intrinsic emission stream does not exist anywhere else at production scale. It only works on QOM because no other chain has QNS Staking's deterministic emission schedule + registry-level lien hooks together.

3. NFT lending without an oracle problem

Every prior NFT pool — BendDAO, Blur Blend, NFTfi pools — has had to make a choice: noisy floor-price oracle, or peer-to-peer with no shared liquidity. The Loaded NFT Pool sidesteps this by reading the staked position's known future emissions directly from the staking contract:

$$V_{\text{collateral}} = \underbrace{Z_{\text{locked}}}_{\text{already-earned}} \;+\; \underbrace{r \cdot w \cdot (T_{\text{end}} - t)}_{\text{deterministic future emissions}}$$

Both terms are state on QNS Staking. Neither involves a price feed for the NFT itself.

4. Lenders are protected by design

A reserve-funded backstop absorbs any bad debt liquidators can't recover. The model is closer to Aave's Umbrella (first-loss capital absorbs before depositors) than to Morpho Vault V2's pro-rata socialization — except our first-loss capital is funded by the protocol's reserve factor, not by a separate staked-token slashing pool. Combined with the grace period and per-block velocity caps, the lender promise is "principal plus accrued interest at withdrawal," enforced mechanically.

Where we sit in the lending landscape

Mechanism	Where it's been done	What Forge does
Kinked two-slope IRM	Aave, Compound, ~all	Same, calibrated for ZAP volatility
Isolated immutable markets	Morpho Blue, Compound v3	Single pool at launch; isolation for Phase 3
Curator vault layer	MetaMorpho, Kamino Vaults	Out of scope — direct pool, no curators
Self-repaying via yield routing	Alchemix v3	Bootstrap Loan, but NFT-collateral & emission-driven
Oracle-less collateral pricing	Ajna (bid-bucket pricing)	Loaded NFT Pool reads emissions, not bids
Stability-pool absorptive liquidation	Liquity v2	Reserve-funded backstop fills the same role
Hub-and-spoke with native settlement	Aave v4 (March 2026)	Phase 3 multi-asset is the analog

The unoccupied square — NFT-collateral pool lending whose debt is auto-serviced from the collateral's own emission stream — is the one we ship.

Where to start

• Use the live product → Pawn
• See the lending model → Forge — Introduction
• All numbers in one place → Rate Card
• Stress scenarios and guarantees → Resilience
• New to lending math? → Glossary and FAQ

Status

Protocol	Status	Capital model
Pawn (P2P NFT lending)	Live	Bilateral
Forge — Direct Loan	In design	Pooled
Forge — Bootstrap Loan	In design	Pooled
Loaded NFT Pool	Planned	Pooled

Heads-up

Forge and the Loaded NFT Pool are documented as designed — they have not yet launched. Every parameter quoted on this site is proposed for discussion across three calibration tiers (conservative / base / aggressive); final values are fixed at audit and announced at the official launch. Pawn is live with the parameters documented on its page.