How NFT Markets Really Work: Illiquidity, Floor Prices and the Actual Risk Profile

NFT markets are described using the vocabulary of financial markets — floor price, volume, market cap, blue chip — and that vocabulary quietly implies things that are not true. Those words carry assumptions built up around assets that trade continuously, in fungible units, with many buyers at any moment. Almost none of that applies here. Understanding where the analogy breaks is most of what separates a considered decision from an expensive one.

This is not an argument that non-fungible tokens are worthless or that the technology is fraudulent. Both claims are too broad. It is an argument that the market’s structure creates specific, knowable risks that the standard presentation obscures.

Illiquidity is the defining property

The single most important fact about NFT markets is that most tokens, most of the time, cannot be sold at anything close to their displayed value.

A liquid market has continuous two-sided interest: standing bids and offers, and a spread between them narrow enough that transacting does not itself destroy much value. NFT collections rarely have that. What they typically have is a set of asking prices from people who want to sell, and a much thinner set of bids from people willing to buy. The gap between the two can be enormous, and it is not visible on the charts that collections advertise.

Trading activity is also concentrated in bursts. A collection may see intense turnover during a launch or a news cycle, then near-silence. Selling into silence means either waiting an unknown period or accepting whatever the highest standing bid happens to be — frequently far below the lowest asking price. Anyone modelling an exit should assume the exit takes time they do not control, at a price they do not set.

Within a collection, liquidity is uneven again. The cheapest, most numerous items trade most often. Rare items may have no comparable recent sale at all, which makes them simultaneously the most valuable-sounding and the hardest to price or sell.

Floor price is the cheapest listing, not a valuation

The floor price is the lowest price at which someone is currently offering to sell an item from a collection. That is the entire definition. It is one seller’s ask.

Several consequences follow, and they are consistently under-appreciated. A floor price is not a bid, so it is not evidence that anyone will pay it — it is evidence that someone hopes to receive it. It reflects a single listing, so if that seller cancels, the floor moves without any transaction occurring. It can therefore rise or fall on zero volume. It says nothing about depth: a floor with one item behind it and a floor with two hundred items behind it look identical on a dashboard and mean completely different things, because attempting to sell into the second one pushes straight through a wall of competing supply.

The derived figure that follows is worse. Multiplying a floor price by the total supply of a collection produces a “market cap” that assumes every single item could be sold at the price of the cheapest current listing, simultaneously, with no effect on price. There is no market anywhere in which that assumption holds. Treat collection market cap as a ranking convenience rather than a quantity of money.

Reported volume deserves scepticism too. Because a wallet can trade with another wallet the same person controls, transaction volume can be manufactured. Wash trading has been an observable pattern on public chains, and it is particularly attractive where a marketplace or protocol rewards activity. The countermeasure is to look at how many distinct holders a collection has, how concentrated ownership is, and whether sales look like genuine dispersal or circulation between a small cluster of addresses. All of that is checkable on-chain, and the fact that it is checkable is one of the genuine advantages of these markets.

Royalties: a norm, not a guarantee

Early NFT marketing leaned heavily on creator royalties — a percentage of each secondary sale flowing back to the original creator, apparently forever.

The technical reality is more limited than the pitch. On most chains, royalty payment is generally not enforced by the token contract itself. It is honoured by marketplaces as a matter of policy. That means a royalty is a convention among the venues where trading happens rather than a property of the asset, and conventions change when competition pressures them. Marketplaces have variously enforced royalties, made them optional, or set them aside entirely, and creators have responded with technical measures like restricting which contracts may transfer their tokens — which introduces its own trade-offs around censorship and future compatibility.

For a buyer, the practical implications are: royalties are a real cost on resale, they may differ between venues for the same item, and they should be included in any mental calculation of what a round trip costs. For anyone buying partly because a project promises creator funding from perpetual royalties, the correct question is whether that revenue stream is contractually enforced, policy-dependent, or simply hoped for.

The actual risk profile

Set aside price speculation and the risks fall into distinct buckets, each worth naming separately.

Structural risk. Illiquidity means the position may be unsellable at the moment you want out. This is not a tail risk in this market; it is the base case for most collections after attention moves on.

Reference risk. Many NFTs do not contain the artwork. They contain a pointer to it. If that pointer resolves to a conventional web server or an unpinned distributed file, the referenced asset can disappear while the token persists. Where and how the asset is stored is a factual property you can verify before buying, and it is one of the few things about an NFT that is genuinely objective.

Rights risk. Owning a token is not automatically owning copyright, commercial rights, or anything beyond the token. What you get depends on a licence the project publishes — or does not. Assumptions here are frequently wrong.

Counterparty and contract risk. Smart contracts can contain upgrade mechanisms, privileged administrative functions, or outright flaws. Marketplace approvals granted to a contract can persist long after you stop using it, and revoking stale approvals is basic hygiene.

Operational risk. Phishing aimed at NFT holders is relentless and well-crafted: fake mint pages, fake support accounts, malicious signature requests that transfer assets rather than logging you in. Read what you are signing. And, without exception, no legitimate marketplace, wallet, project team, or support channel ever asks for a recovery phrase.

Concentration risk. Prices in a collection can be substantially supported by a handful of large holders. On-chain data lets you see that concentration, which is more than can be said for most private markets.

A fairer summary

The honest position is that NFTs are a genuinely novel primitive — verifiable, transferable ownership records on a public ledger — wrapped in a market structure that has repeatedly behaved badly. The primitive has uses: provenance, ticketing, credentials, identifiers for digital goods inside applications. The market that formed around collectible images optimised for attention and turnover, and its reported metrics were built to flatter rather than inform.

Nothing here is a recommendation to buy or avoid anything. It is a set of questions worth answering before committing money you would rather keep: how would I sell this, to whom, over what period, at what realistic discount to the number I am looking at, and what exactly do I own if the project stops existing tomorrow? Our glossary defines the terminology precisely, our learn section covers underlying mechanics, and the wider NFTs and Web3 coverage looks at how these markets are put together.