An NFT lending platform lets holders borrow crypto against NFTs they lock as collateral. The central design choice is whether lenders price each loan themselves (peer-to-peer) or deposit into a shared pool that prices loans with an oracle (peer-to-pool). That choice determines your risk, your liquidity and how you liquidate.
This page covers protocol design. For the borrower and lender product, meaning loan offers, the loan lifecycle and operations, see the companion guide on NFT loan platform development.
Why NFT lending is harder than token lending
In a protocol like Aave, collateral is a liquid token with a reliable price feed, and liquidators can sell it instantly on deep markets. NFTs have none of that. Each item is different, prices update only when something trades, and selling a seized NFT can take days. Every NFT lending design is a way of coping with illiquid, hard-to-price collateral. If you are used to fungible lending, compare with Aave-style lending protocols.
The main protocol models
| Model | Pricing | Liquidation | Examples of the pattern |
|---|---|---|---|
| Peer-to-peer, fixed term | Lender makes an offer per NFT or per collection | Lender takes the NFT if the borrower misses the deadline | NFTfi, Arcade |
| Peer-to-pool | Oracle floor price × loan-to-value ratio | Health factor triggers an auction | BendDAO, early ParaSpace |
| Perpetual callable loans | Lender offers; rate set by market | Lender calls the loan, a refinancing auction runs, NFT goes to the lender if nobody refinances | Blend (Blur) |
| NFT vault-backed | Price of vault token on a DEX | Sell the vault token | Lending against fractionalized collections |
Peer-to-peer
No oracle is needed, because a human lender decides how much a specific NFT is worth. Defaults are simple: after the due date, the lender can claim the collateral from escrow. The trade-off is slower matching; borrowers wait for an offer they like. This model has survived market crashes best because there is no automated liquidation cascade.
Peer-to-pool
Lenders deposit ETH into a pool and earn interest; borrowers draw instantly against whitelisted collections. The protocol needs a floor-price oracle and a liquidation process. In August 2022, BendDAO showed the risk: as blue-chip floor prices fell, many loans became liquidatable at once, auctions found few bidders above the debt, and depositors rushed to withdraw. The protocol survived by changing parameters under pressure, but the episode is the standard case study for this model.
Perpetual callable
Blend, launched by Blur in 2023, removed fixed terms and oracles. Loans run until the lender calls them; a call starts a Dutch auction in which other lenders can refinance the loan. This keeps the market-pricing benefit of peer-to-peer while making loans liquid for lenders.
Protocol components
- Escrow or lien contract. Holds collateral, or records a lien that blocks transfer. Must handle ERC-721 and often ERC-1155 safely, including
onERC721Receivedcallbacks. - Loan terms engine. Principal, APR or fixed repayment, duration, origination fee. Use integer math with explicit rounding.
- Oracle (pool model). A floor-price feed should use time-weighted averages, ignore wash trades and outliers, and pause when data is stale. Single-marketplace floors are trivial to manipulate.
- Liquidation module. Auctions with sensible durations, a grace period for the borrower to repay, and a plan for bad debt when no bid covers the loan.
- Collection allow-list and risk parameters. Maximum LTV, liquidation threshold and exposure caps per collection, under governance with a timelock.
- Indexer and front end. Shows offers, health factors and auction status clearly.
Risk parameters that matter
For a pool model, the numbers you choose matter more than the code. Keep loan-to-value conservative (well below what token lending uses), cap total borrowing per collection relative to its real daily trading depth, and model a scenario in which the floor halves in a week and no buyers appear. If the pool cannot survive that, the parameters are wrong. Delisting a collection should be possible quickly, and the process for doing so should be public.
Security and audits
NFT lending contracts handle valuable collateral and complex state transitions, which makes them prime targets. Common bugs include re-entrancy through token receiver hooks, incorrect interest accrual, liquidation logic that can be front-run, and signature replay on off-chain offers. Budget for at least one independent audit and a bug bounty; see smart contract audit.
Is the market there?
NFT lending volume shrank along with NFT trading after 2022 and is concentrated in a handful of protocols and blue-chip collections. A new protocol needs a distinct angle: an underserved chain, a specific asset class such as tokenized physical goods, or integration into a marketplace that already has users. General background on pooled lending is in DeFi lending and borrowing platforms.
Frequently asked questions
Which NFT lending model is safest for lenders?
Peer-to-peer with conservative offers, because each lender sets terms and there is no shared pool exposed to oracle failures. Lenders still risk ending up with an NFT worth less than the loan.
How do NFT lending platforms price collateral?
Peer-to-peer platforms let lenders decide. Pool platforms use oracles built from recent sales and floor listings across marketplaces, usually averaged over time to resist manipulation.
Can rare NFTs be valued above the floor?
In peer-to-peer markets, yes, if a lender agrees. Pool models generally lend against the floor only, since trait-level pricing is too thin to automate safely.
How long does it take to build an NFT lending protocol?
A focused peer-to-peer protocol with a front end might take a small senior team three to five months, plus audit time. Pool models with oracles and auctions take longer and need more risk work.