Thirteen

Lot One

On 1 October 2021, at Christie’s in New York, sale 29238 — Post-War to Present — opened with Curio Cards.

Not included in. Not featured alongside. Lot one. A two-hundred-lot sale of twentieth and twenty-first century art, containing Frankenthaler and Guston, began with thirty-one pictures that had sold for twenty-five cents each in 2017 and that the people who made them had been unable to give away on a Discord server three years earlier.

The lot title, verbatim: “CURIO CARDS (EST. 2017) / Full set of 30 cards plus 17b”. The medium line reads “non-fungible tokens, ERC-1155” — which means what stood on the rostrum was not the 2017 tokens but their 2021 sleeves, the wrapper of chapter twelve. The auctioneer was Rahul Kadakia; the head of sale was Isabella Lauria; the lot’s wallet was 0xD035a780DECCF7808875c6a555937B7c44299F45.

And Christie’s took live bids in ether, which it had never done before. CoinDesk, two weeks ahead: “The Oct. 1 auction also marks the first time Christie’s live bidding will be denominated in cryptocurrency instead of the usual local currency.”

Noah Davis, Christie’s head of digital art, told CoinDesk on 17 September what that felt like from inside the firm:

“I’ve been bidding in auctions on behalf of clients for almost a decade,” said Davis, the Christie’s digital art lead. “I’ve placed bids in USD, GBP and EUR – always fiat. To think I’m going to hear the auctioneer calling out my bids in ETH, that’s just incredible. It doesn’t happen.”

He also gave the sentence that did more than anything else to fix the project’s status:

“A true full set of Curio Cards is one of the few holy grails as far as art projects on Ethereum go,” said Noah Davis, head of digital art at Christie’s. “It’s the first of its kind, predating Punks, even.”

Predating Punks, even. Chapter fourteen takes that claim apart. What belongs here is that it was said by Christie’s, in print, a fortnight before the hammer, and that the authority of the institution saying it is why the claim stuck.

The three numbers, kept apart

This is where almost every retelling of Curio Cards goes wrong, so the figures are set out separately and the arithmetic shown.

The estimate was 250–350 ETH. That is the fact. The dollar equivalents vary by source because ether moved between 17 September and 1 October; CoinDesk gave $870,000 to $1.3 million. The ETH band is the stable number; any dollar band is date-dependent and should be reported as such.

320 ETH was the hammer price — what the winning bid was called at.

393 ETH was the price realised, which includes the buyer’s premium Christie’s charges on top of the hammer.

$1,267,320 is that same 393 ETH expressed in dollars at the day’s rate.

These are not three results. They are two numbers describing one sale plus one currency conversion, and the arithmetic confirms it: $1,267,320 ÷ 393 = $3,224.73 per ether, and $1,034,377 ÷ 320 = $3,232.43 per ether. The same exchange rate to within a rounding error, which is what you would expect if the two ETH figures are hammer and hammer-plus-premium on a single lot. Applying Christie’s 2021 New York premium tiers to a hammer of about $1.035 million yields roughly $1.278 million — again within rounding.

Artnet News, 4 October 2021, states its own convention plainly — “as always, final prices include auction-house fees; presale estimates do not” — and reports:

“The sale kicked off with a set of Curio Cards, said to be the first art NFTs minted on the Etherium blockchain. Christie’s accepted live bids in Ether cryptocurrency for the set of 30 cards plus 17b (a reference to a rare misprint that was included with the set, making a total of 31 NFTs). The group sold for ETH393 ($1.3 million), exceeding the high estimate of ETH350.”

“Etherium” is Artnet’s spelling, reproduced as published.

One source disagrees and it is the project’s own. Curio DAO’s retrospective of 14 January 2022 calls 393 ETH a bid rather than a price realised:

“The winning bidder was none other than Ethereum co-founder Taylor Gerring with a bid of 393 ETH, exceeding Christie’s estimate of 250–350 ETH and nearly quadrupling the previous Full Set sale price of 107 ETH.”

The weight of evidence is against that wording — a bid of 393 against a hammer of 320 does not reconcile, whereas premium does. The project’s own post is reported here as conflicting rather than silently corrected. It also supplies a useful comparator that appears nowhere else: the previous full-set sale was 107 ETH.

Hunt himself has given both figures at different times, and chapter five of the quote bank keeps them apart for exactly this reason. Live on air he said “the final USD was one million, uh, 35,000 dollars” — which is the hammer in dollars. Years later he said the set “sold for 393 Ethereum worth 1.2 million sixty seven thousand three hundred and twenty dollars.” Both are right. They are not the same measurement, and letting them sit next to each other unlabelled is how a single sale becomes two contradictory sales in the retelling.

One thing not to cite. The live Christie’s lot page currently renders “Estimate ETH 805,548 – ETH 1,127,767” against “Price realised ETH 393” — mutually impossible figures, produced by a currency-toggle artefact in Christie’s present template. The estimate was 250–350 ETH.

What the lot essay said, and what it got wrong

Christie’s lot essay is a good piece of writing and it is also the single most consequential overreach in the Curio Cards record.

On the cards themselves it is excellent:

“The Cards are a convoluted and motley array of surrealist-tinged kitsch, anti-corporate/pro-decentralization satire, slapstick cartoons, high-minded abstraction and hijinks.”

On the Robek cards of chapter six it is accurate and oddly careful — it never once uses the word “founder” and gives no surnames:

“Card nos. 21 through 23, by Robek World, depict three of the IRL members of the Curio team in a loose and campy cartoon style—Tom (The Wizard (who is also very obviously MadBitcoins)), Travis (The Bard) and Rhett (The Barbarian).”

The nested parentheses are Christie’s own. Note what that sentence does: it identifies the Wizard as MadBitcoins — Hunt — which is correct, and which is also precisely the collision chapter five flags, because card 20 is the card titled MadBitcoins, by Phneep, and card 21 is the Wizard, by Robek. Two Mad Bitcoins cards, adjacent, by different artists. Christie’s gets it right. Other accounts do not.

On totals: “There are a total of almost 30,000 Curio Cards in existence, with an estimated 4,000 currently lost or effectively destroyed.” That is consistent with Christie’s own card-by-card figures, which sum to 29,796 including 17b — and inconsistent with the project’s 29,700 and with curio.cards’ 29,185. Three published totals, none reconciled. The card table in the apparatus keeps the columns separate rather than averaging them.

And then the overreach, which chapter fourteen has to deal with:

“On May 9th 2017, the first ever NFT-based art project launched, billed as a kind of virtual gallery opening.”

No “on Ethereum”. No hedge. That is a bigger claim than any founder ever made — bigger than Travis Uhrig, who said “I never felt comfortable saying that before the first because you know you never know”; bigger than Hunt, who said he would not fight an earlier one if it were on the blockchain. The auction house was less careful than the people selling through it, and the auction house is who the press quoted.

Two errors for the record. Christie’s gives the same contract address for both card 17 and card 17b — chapter seven covers this; the address given is card 17’s. And Christie’s describes Marisol Vengas as “an anonymous artist collective comprised of various members who have never met in the physical world”, which chapter five sets against Max Infeld saying on air that Marisol Vengas is his own pseudonym.

The buyer

Status: credibly reported by two independent sources and self-identified, but not confirmed by any primary record this research read.

Castor and Gerard, June 2022, state that “Taylor Gerring, an early contributor to the Ethereum project, was the buyer”, and hyperlink a tweet of his dated 2 October 2021, the day after the sale. That tweet was not read, so nothing is said here about its contents. Curio DAO’s January 2022 post calls him “Ethereum co-founder Taylor Gerring” — a promotion this book does not follow, preferring “early contributor to the Ethereum project”, which is the more conservative of the two descriptions available.

Wikipedia, for its part, names no buyer at all and describes the set as “sold by an anonymous seller.” Christie’s does not publish buyer identities.

So: Gerring publicly identified himself as the buyer the day after the sale, and two independent write-ups report it. That is where the evidence stops, and this chapter stops there with it. Hunt’s own recollection of the name, incidentally, is characteristically hedged — “a mess his name, Taylor Garing, I think, I’m pretty close” — which is a man who has not met the person who bought his life’s most valuable object.

Afterwards, in three cities and a mountain

Sotheby’s, two weeks later — confirmed to have happened, price not established. A full set sold in Natively Digital 1.2: The Collectors, Sotheby’s Metaverse, running 14–21 October 2021, a sale of 53 NFT lots from 19 collectors. The lot URL is corroborated by three independent Wayback captures. Curio DAO’s account:

“Curio Cards at Sotheby’s — Closely following this event was Curio Cards’ appearance at Sotheby’s. A Full Set was sold at their week-long Metaverse auction. This second sale by a mainstream auction house confirmed what the Christie’s auction had demonstrated — the legitimacy of Curio Cards in the art world.”

The price is an honest gap. The live URL now redirects and the archived snapshots are an empty JavaScript shell. A figure of “more than 100 ETH ($260,000)” surfaced in search results, could not be traced to any source, and is therefore not printed as a figure here.

Svalbard — confirmed, but not the programme usually credited. Curio Cards was named as a partner in the Non-Fungible Vault, a programme of TopDog Studios that deposits into the Arctic World Archive. TopDog’s announcement of 2 December 2021 lists Curio Cards among the partners forming “COHORT ONE”, describing it — inaccurately by chapter ten’s standard — as “First NFTs on Ethereum that helped define the ERC721 and ERC1155 standards we know today.”

The mechanism is film, not code, and the programme page is precise about it:

“We’re etching NFTs (including the art and provenance information) onto special purpose-built silver-halide film that is designed to last for over 1,000 years. The films are then stored in a decommissioned coal mine in the Svalbard archipelago, known as the Arctic World Archive.”

Paul Price, founder and managing director of TopDog Studios, supplies the line that justifies the whole exercise:

“Traditional art has stood the test of time with pieces created some 45,000 years ago still in existence, yet NFTs minted 45 days ago are already missing.”

The Arctic World Archive’s own post of 16 March 2022 confirms the deposit and dates the ceremony to 17 March 2022 in Longyearbyen. It quotes Price, and Rune Bjerkestrand of Piql, and gives a total of over seventy thousand works — and it does not name Curio Cards. Both TopDog pages, as they still stand, describe the first deposit in the future tense. So Curio’s presence in that mountain is attested by TopDog’s partner listing and by the project itself, and by nobody else.

The project’s own claim is at least dated after the event rather than before it. From the Curio Cards account on 14 May 2022, two months after the ceremony:

“Curio Cards’ history and art was sealed for 1000+ years in the Non-Fungible Vault at the Arctic World Archive! Ever hold a Card before the snapshot? You are in the vault and eligible for the Snowy Curio @POAPxyz!”

Ever hold a Card before the snapshot? You are in the vault. Which means the thing committed to film was not only the artwork but a register of who held what on a particular day — provenance, in the sense TopDog uses the word, being the part they thought worth saving for a thousand years.

One thing this is not: the GitHub Arctic Code Vault. That is a separate programme, which deposited a bulk snapshot of public source code into the same facility — the Arctic World Archive’s own post about the TopDog deposit carries a link to it in the margin, which is presumably how the two came to be confused. One is reels of code; the other is film of pictures. The 1n2.org wiki asserted the GitHub version for some time, and also said “Card #21 (Mad Bitcoins)” when card 20 is MadBitcoins; both were corrected on 1 October 2026.

Taschen — confirmed. Robert Alice, On NFTs, TASCHEN revised edition, 22 October 2025, pages 46–47. This is Wikipedia’s first reference and carries real weight for its opening factual sentences. The spread was photographed in a project tweet of 7 March 2024. The book itself was not obtained, so what it says is not quoted. A widely repeated claim that cryptograffiti’s cards 11–13 were excluded for permissions reasons is unverified — plausible, since those are the cards altering corporate marks, but plausibility is not evidence.

What the encyclopedia knows

It is worth ending on how thin the public record still is, because it is a measure of how much of this book came out of one man’s hard drive.

Wikipedia’s Curio Cards article is four paragraphs and seven references. It has the founders, the launch date, the launch price, the artists-were-paid claim, the “colloquially considered to be the first collection of NFT artworks on the Ethereum blockchain” hedge, and the Christie’s sale at $1.2 million with the 17b misprint named.

What it does not contain: the rediscovery. @DieAping. Adam McBride or the word archaeology. The wrapper. The locked cards. Sotheby’s. The Arctic archive. The DAO. The royalty. The 111-to-106-to-33 set arithmetic. The buyer. Travis Uhrig removing the shop links. And — remarkably, given that it is the project’s own proudest fact — no mention of ERC-721 citing Curio Cards at all, in either direction.

The most-read account of Curio Cards in the world omits both the reason the project survived and the footnote that this book is named after.