An NFT gedankenexperiment

Imagine that you purchased a piece of physical art or other collectible which came with an accompanying NFT (or alternatively you bought an NFT which came with an accompanying physical object) - which one would be worth more money 20 years later? Would the pair be worth more than the sum of the two parts? Would one of them be totally worthless?

The impetus for this article

About a week ago I purchased an NFT from the patrn.me platform from the artist Lee Rubenstein on one condition: he had to send me the original drawing that was scanned to create the digital image. You can see the art on patrn or opensea, and a photo of the actual drawing below:

So now I own two things that point at the same image: a token on a blockchain that says I own "it", and a physical piece of paper that a human being actually drew on. Which one is the better bet to still be worth something when my hypothetical kids inherit it in 2042?

The case for the physical object

The boring answer, and I think the correct one, is the paper.

Physical art has a track record measured in millennia. Cave paintings, illuminated manuscripts, Old Masters. The medium degrades gracefully and the social consensus that "this object is valuable because a particular person made it" is one of the most durable agreements humans have. A drawing in a drawer requires no servers, no pinning service, no private keys, and no chain to keep producing blocks. It just has to not catch fire. The worst realistic 20-year outcome for the paper is that nobody cares about Lee Rubenstein in 2042 and it is worth what any nice drawing is worth, which is to say not zero.

The NFT has a much wider distribution of outcomes, and most of the mass is on the left side of zero.

Everything that has to survive for the NFT to be worth anything

This is the part people who bought jpegs in 2021 were not thinking about. An NFT is almost never the image. It is a token that contains, at best, a pointer to the image. For that pointer to still resolve to something in 20 years, a remarkable amount of infrastructure has to keep working:

• The chain has to still exist and be secured. Most NFTs do not live on Bitcoin or Ethereum mainnet, they live on whatever cheap chain or L2 was fashionable that quarter. Plenty of those will be abandoned. A token on a dead chain is a database row nobody is running the database for.
• The link can't rot. If the metadata is an HTTP URL pointing at patrn.me or OpenSea, then the day that company turns off its servers, your NFT points at a 404. This is not hypothetical, it is the default outcome for startups. Marketplaces shut down, domains lapse, S3 buckets get deleted.
• IPFS doesn't save you for free. "It's on IPFS" only means the content is addressable by hash, not that anyone is storing it. If the only node pinning that content is the platform's pinning service, then IPFS is just a more expensive 404 once the platform stops paying the bill. Decentralized storage that nobody bothers to keep is centralized storage with extra steps. I've ranted about adjacent versions of this in why NFTs.
• On-chain vs off-chain matters enormously. The small handful of fully on-chain NFTs, where the actual image bytes are stored in the contract, have a real shot at 20 years because they survive as long as the chain does. The 99% that store a URL or even a bare hash off-chain are betting on a stranger's hard drive.

And even if all of that survives, there's the thing nobody wants to talk about: owning the token is not owning the copyright. In most cases the NFT conveys no legal rights to the underlying work at all. You own a receipt. The artist can keep selling prints, the image keeps getting right-clicked, and what you actually hold is a blockchain entry asserting provenance. That can be worth a lot, but only if the social consensus that this particular receipt matters survives, which brings us right back to infrastructure and belief.

So which wins?

I can construct a scenario where the NFT wins. If the chain survives, the metadata was stored sanely, the artist becomes famous, and the culture decides that the on-chain provenance is the canonical record of ownership, then the token could end up worth far more than the drawing, because the token is the scarce, verifiable, transferable thing and the paper is just the source material. In that world the pairing is worth more than the sum: the physical object authenticates the token, and the token makes the physical object liquid and provable.

But that is the tail, not the body, of the distribution. The body of the distribution is: the platform is gone, the link is dead, the chain is a ghost town, and the only thing left with value is the piece of paper, precisely because it never depended on any of that. The NFT's value is entirely a function of whether the infrastructure and the social consensus both survive for two decades. Physical art only needs one of those, and a weaker version of it at that.

My bet: in 20 years the physical object is the safer store of value, the NFT is a lottery ticket on the survival of a software stack, and most of those stacks won't be around. The genuinely interesting case is the rare one where the pairing works and they reinforce each other. That's the experiment I actually wanted to run, which is why I made the artist mail me the paper. The paper is the hedge.

Update (2026)

We did not need to wait until 2042. I wrote this in early 2022, near the top, and the market answered quickly.

Both of mine are worthless now in any market sense. The patrn.me listing is gone, the token points at infrastructure nobody runs, and while I still have Lee Rubenstein's original drawing, there is no real market for it either. The pairing I was excited about netted me a dead link and a nice piece of paper. The paper is the better of the two, which was the bet.

The broader market played out as expected: volume collapsed, most collections went to roughly zero, marketplaces shut down or stopped paying for storage, and links rotted. "Decentralized storage" mostly meant "stored until the one company paying for it stopped." The on-chain provenance is, in many cases, a provenance record for a 404.

The Currency

The cleanest public test of this exact question was Damien Hirst's The Currency: 10,000 hand-painted spot artworks, each paired one-to-one with an NFT, sold in 2021. Buyers then had to choose, by 27 July 2022, to keep the physical or the NFT. Whichever they gave up would be destroyed.

The split was nearly even, with a slight edge to the physical: 5,149 kept the artwork, 4,851 kept the NFT. Hirst then burned the 4,851 physical works whose owners chose the token, starting with 1,000 at Newport Street Gallery on 12 October 2022 and continuing through the show.

Two things stand out. First, even a Hirst-branded, well-funded version produced roughly a coin flip, not a rout, so "physical always wins" is too strong. Second, the burn is the interesting mechanism: destroying the paper made the token the only surviving artifact for those 4,851 buyers, which is the one clean way to make an NFT genuinely scarce against a physical. My little patrn experiment couldn't do that, because nobody was ever going to burn anything.

What I'd still defend

The paper was the safer 20-year store of value, and the infrastructure failed on schedule. The nuance: a small slice of NFTs (fully on-chain work, and a few collections where the cultural consensus held) kept real value, the tail I said could happen. The idea of a chain as a provenance layer is not dead; it was bolted to the wrong things the first time. The version still worth running is the pairing, where a physical object and a credible on-chain record authenticate each other, ideally with a mechanism (like Hirst's burn) that forces the scarcity to mean something.

My longer-running decentralization thesis is in Monero is a DAO.