Where NFTs Found Real Utility After the Wave

Stripped of the speculation, a non-fungible token is a transferable record of ownership with programmable transfer rules, verifiable by anyone without asking the issuer. That is a narrow capability. It is genuinely useful in a small number of places and pointless in most of the places it was proposed.

What actually happened

The speculative phase attached enormous prices to assets whose only property was scarcity, and then the prices went away. What is worth examining is not the collapse — that was predictable and widely predicted — but the residue: the applications still running afterwards, built by people who were solving a problem rather than issuing a collection.

Those applications share a characteristic that is easy to state and was widely ignored at the time. They use tokenised ownership to solve a coordination problem between parties who do not fully trust each other. They do not use it to manufacture scarcity, and they do not depend on the asset appreciating.

Where it does real work

Ticketing

The clearest fit. Event ticketing has a long-standing problem: the issuer wants control over resale — price caps, transfer windows, identity binding — but loses that control the moment a ticket leaves their platform.

Encoding transfer rules in the asset itself means they travel with it. A ticket can be made non-transferable after a date, capped at face value on resale, or bound to a verified identity, and those constraints hold on any venue that respects them rather than only on the issuer’s own marketplace.

The secondary benefit is that the issuer retains a relationship with whoever currently holds the ticket, which is commercially significant and was previously impossible once a ticket changed hands.

Credentials and memberships

Certifications, course completions, professional licences and membership records share a property: the holder should control the record, and a verifier should be able to check it without contacting the issuer.

Traditional credentialing fails on both counts — you request a transcript, the institution issues it, the verifier trusts the institution’s current systems. Tokenised credentials, usually non-transferable, let a holder present something a verifier can check independently and that survives the issuer’s systems changing or the issuer disappearing.

The honest caveat is that this only matters where issuer-controlled verification is genuinely a problem. For many credentials it is not.

Creator royalties, with a caveat

The original pitch — creators earn on secondary sales automatically — turned out to depend entirely on marketplaces choosing to honour it, and many stopped. Royalties enforced only by marketplace convention are not enforced.

What survives is narrower: designs where transfer itself is restricted at the asset level so that a sale which does not pay the royalty cannot settle. That works, and it costs composability, because assets with restricted transfer do not interoperate with everything else. It is a real trade-off rather than a solved problem.

In-application ownership

Where a user acquires something inside an application and the value of portability is real — moving an item between applications built by different teams — tokenised ownership does something a database cannot. The condition is that a second application actually exists and wants to honour it. In practice that condition is met far less often than proposed.

Where it does not fit

Anything with a single trusted issuer and no transfer. If one organisation controls issuance, verification and revocation, and the record never moves, a database is simpler, cheaper and better.

Physical goods authentication, where the hard part is binding a token to an object. The token proves a token exists; it does not prove which physical item it corresponds to. That link remains a trust assumption and it is where the fraud lives.

Anything whose entire value proposition is artificial scarcity. Scarcity is easy to create and, on its own, generates no durable demand.

The test

One question separates the durable applications from the rest: does someone other than the issuer need to verify or enforce something, without the issuer’s cooperation?

If yes, tokenised ownership is doing genuine work. If no, it is a database with additional steps and worse tooling — which describes the substantial majority of what was proposed during the wave, and the reason so little of it survived it.