Market notes · November 21, 2026
What sells.
What the numbers
leave out.
A source-led audit behind five new studio collections.
Demand is segmented. Physical-backed collectibles trade on underlying goods and transaction mechanics; established character brands trade on distribution and identity; specialist art buyers respond to an artist, a coherent series and a credible process. None establishes demand for an unknown new studio.
Desk audit of provider snapshots, auction-house results, marketplace-labelled sales, official product documentation and collection pages. This is not a wallet-level certification of organic sales. Sources were retrieved on November 21, 2026; their measurement periods differ and cached figures are not synchronized.
Different products.
Different buyers.
| Segment | Observed evidence | What it means here |
|---|---|---|
| Physical-backed collectibles | Courtyard: 12,789 displayed 24-hour sales; 259.91 ETH volume. CoinGecko — Courtyard.io | Trading, packs and redemption refer to underlying goods. We are not creating vaulted-card substitutes. |
| Established character IP | Pudgy Penguins: 4 displayed 24-hour sales; 9.26 ETH volume. CoinGecko — Pudgy Penguins | Even a recognized brand can have thin daily turnover. New characters need an audience. |
| Generative art | Chromie Squiggle #9871 marked sold for USD 11,430 in February 2025. Sotheby's — Digital Art Day Auction, February 2025 | A specialist collecting market exists. This is an established-artist historical result, not our price anchor. |
| Curated AI art | Four visible Botto sales range from 1 to 3.025 ETH. SuperRare — Botto 2026 collection | AI creation can be disclosed and collected; artist, process and curation matter. |
A closer look
at the evidence.
The Courtyard snapshot implies about 0.0203 ETH per sale, while the Pudgy snapshot implies 2.315 ETH. These are arithmetic means of provider-reported turnover, not valuations, independent-buyer counts or synchronized comparisons. Repeated trading and product mechanics may shape both.
growthepie separately shows 365.2K Courtyard application transactions against 1.6K active addresses in its displayed seven-day window. Those are application actions and addresses, not sales and people. Adding them to NFT-sale counts would be an error. growthepie — Courtyard application analytics
Courtyard's own documentation describes vaulting, physical redemption, packs and buyback offers. That is a custody and fulfilment business. Courtyard — How it works
Completed sales, with their limits
| Work / artist | Reported result | Evidence class |
|---|---|---|
| Chromie Squiggle #9871 Snowfro | 11,430 USD | auction-house sold result Sotheby's — Digital Art Day Auction, February 2025 |
| Digital Cascade Bisects Organic Decay Botto | 1.2 ETH | marketplace labelled sold SuperRare — Botto 2026 collection |
| Hourglass Evacuates Upward Botto | 3.025 ETH | marketplace labelled sold SuperRare — Botto 2026 collection |
| Striped Metamorphosis Botto | 1 ETH | marketplace labelled sold SuperRare — Botto 2026 collection |
| Tongue Supports the Brutalist Fortress Botto | 1.1 ETH | marketplace labelled sold SuperRare — Botto 2026 collection |
| Machine Hallucinations — ISS Dreams — A Refik Anadol | 277,200 USD | auction-house realized price Christie's — Augmented Intelligence results |
The four Botto entries are a convenience sample, not an AI-art index. Their median is 1.150 ETH, but individual settlement dates and buyer-funding relationships were not verified. Cached dollar conversions are excluded. The larger Sotheby's and Christie's results are historical and belong to established artists.
Christie's USD 728,784 AI-art sale included mixed mediums, not just NFTs. Its USD 277,200 top lot contributed about 38.0% of the total. The catalogue distinguishes estimates from results including buyer's premium. A headline total can conceal concentration. Christie's — Augmented Intelligence sale catalogue
Five directions.
Five different tests.
Priority 1 · Generative fine art
Tidal Mathematics
Generative art has documented auction results and a specialist collecting ecosystem. Its visual discipline, reproducibility and inspectable process fit a small studio. Established-artist prices do not set the value of these new works.
Observed specialist art market; demand for this studio untested.
Priority 2 · Surreal AI stills
Impossible Herbarium
Disclosed AI art has completed sales on specialist platforms. The test here is a tightly edited visual series and a clear creation record, rather than AI novelty or implied physical ownership.
Observed AI-art sales; botanical theme and new-artist demand untested.
Priority 3 · Original character worlds
Night Custodians
Established character brands combine recognizable designs with sustained storytelling and distribution. These original characters test attachment to a new world; they do not inherit the audience, licensing or liquidity of established collections.
Observed established-brand activity; building a new audience is the main constraint.
Priority 4 · Playable digital objects
Orbit Relics
Game-related NFT activity depends on usable products and existing economies. This is a modest playable-art test, not an item in another game, a promised metaverse, or an earning mechanism.
Adjacent game-asset activity observed; transfer to standalone playable art is speculative.
Priority 5 · Participatory art for people and agents
Common Signals
Participatory curation has precedents, but reliable purchasing demand from independent AI agents was not established in this audit. This collection tests real use and completed creative exchanges before any commercial claim.
Experimental. No dependable independent-agent buyer market verified.
What we will
not infer.
- No floor-price movement is treated as realized profit.
- No claim that a wallet address equals a unique person or independent agent.
- No market-wide wash-adjusted volume is asserted: CryptoSlam's direct dashboard returned 403, and wallet-level funding/beneficial ownership were not audited.
- No claim that a tiny visible sale sample establishes a median for all AI art or a valuation for Steve Bishop's work.
- Physical toy sales, application transactions, fungible tokens, liquidity-position NFTs and unsold asks are not added to art-sale totals.
Art Blocks' phase page shows an edition of 1,024 and 23 available listings from 0.0899 ETH. That is evidence of a collecting format and asking supply, not sell-through. Art Blocks — phase by Loren Bednar Pudgy's official recap describes a broader brand and distribution system; those are company statements, not independently audited NFT revenues. Pudgy Penguins — September 2026 recap
Historical sector contraction also matters: Binance Research records a difficult 2025 for NFT prices and GameFi investment. A functioning puzzle is more credible than a promised game, but it is not proof of a market for playable art. Binance Research — Full-Year 2025 and Themes for 2026
How to test
without manufacturing demand.
The current catalogue contains six tidal drawings, nine botanical works and nine character scenes, alongside two three-work studies and the original triptych. Test generative and surreal stills first; use the character stories and free interactive pieces to learn what people return to. Mint only after platform access and a concrete buyer path exist.
Track qualified enquiries, independently completed creative exchanges and actual paid settlements separately. No trackers, enquiries or sales counters have been invented or installed.
Success means: A verified independent buyer and completed payment, with net proceeds after actual fees. Views, downloads, demo agents and wallet transfers alone do not qualify.
What a deeper transaction audit would require
Resolve each sale to a contract, token and settlement transaction; distinguish mints, secondary sales, transfers and bundles; inspect reciprocal trading and funding relationships; examine buyer concentration and the time between trades. A wallet-relationship flag is a reason to investigate, not automatic proof of wash trading. NFT Wash Trading: Quantifying suspicious behaviour in NFT markets
The direct CryptoSlam dashboard returned 403 during this research. No complete market-wide wash-adjusted data set or beneficial-owner identity was obtained.
Source ledger.
CoinGecko — Courtyard.io
Period: Provider-displayed rolling 24 hours; fetched November 21, 2026. Cached page freshness is not synchronized with other sources.
Class: provider snapshot
12,789 sales and 259.91 ETH volume; 459,732 assets. The opened page showed 139,882 owner addresses and a 0.00320 ETH floor in its marketplace table. A search excerpt displayed a different owner count, so owner counts are not used for cross-source calculations.
Limit: Floor is an asking price. Sales are transactions, not unique buyers. Buybacks, packs and repeated trading may affect turnover.
CoinGecko — Pudgy Penguins
Period: Provider-displayed rolling 24 hours; fetched November 21, 2026.
Class: provider snapshot
4 sales and 9.26 ETH volume; 8,888 assets and 4,998 owner addresses. Displayed floor: approximately 2.25 ETH.
Limit: An established collection's floor and turnover do not predict a new character collection's demand. Owners are addresses, not verified people.
growthepie — Courtyard application analytics
Period: Latest displayed seven-day window; data timestamp November 20, 2026, 05:35:03 UTC in the indexed snapshot.
Class: indexed onchain application metrics
365.2K transactions, 1.6K active addresses, and $4.6K in gas fees. Transactions were down 6.3% against the prior seven days.
Limit: Application transactions are not NFT sales. Active addresses are not unique buyers. Contract labels and coverage can change.
Courtyard — How it works
Period: Official product page fetched November 21, 2026.
Class: operator documentation
The product includes authenticated vaulted collectibles, physical redemption, packs and buyback offers.
Limit: These mechanics require inventory, custody and fulfilment. Art depicting a card is not an equivalent product. Operator statements are not an independent custody audit.
Sotheby's — Digital Art Day Auction, February 2025
Period: Auction held February 12–19, 2025.
Class: auction-house completed-sale result
Chromie Squiggle #9871 by Snowfro is marked sold at USD 11,430.
Limit: Historical result for an established artist. Use the house's displayed result; do not compare it with an ETH ask as if fees, dates and currencies were identical.
SuperRare — Botto 2026 collection
Period: 2026 collection; indexed marketplace snapshot retrieved November 21, 2026. Individual settlement dates were not exposed in the accessible snapshot.
Class: marketplace-labelled completed sales
Digital Cascade Bisects Organic Decay: 1.200 ETH; Hourglass Evacuates Upward: 3.025 ETH; Striped Metamorphosis: 1.000 ETH; Tongue Supports the Brutalist Fortress: 1.100 ETH. These four entries were labelled Sold for, separate from list prices.
Limit: Convenience sample of visible entries, not a representative sample or an independent wallet-funding audit. Cached dollar conversions are excluded.
Christie's — Augmented Intelligence results
Period: Sale closed March 5, 2025.
Class: auction-house result
Sale total USD 728,784. Refik Anadol's Machine Hallucinations — ISS Dreams — A realized USD 277,200.
Limit: The auction mixed mediums, including physical works. The total is not pure NFT volume. Auction results include buyer's premium, as the sale catalogue specifies.
Christie's — Augmented Intelligence sale catalogue
Period: February 20–March 5, 2025.
Class: auction catalogue
The catalogue reports the USD 728,784 sale total and states that sales totals include buyer's premium, while estimates exclude it.
Limit: The difference between estimates and realized prices matters; estimates are not sales.
Art Blocks — phase by Loren Bednar
Period: Collection released August 20, 2021; current page fetched November 21, 2026.
Class: specialist collection listing page
Fixed edition of 1,024 artworks; 23 listings starting at 0.0899 ETH on the fetched page.
Limit: Listing supply and starting ask are not completed sales or sell-through. This page is evidence of a specialist collecting format, not current demand at that ask.
Pudgy Penguins — September 2026 recap
Period: Published September 30, 2026.
Class: official brand account
The brand describes wholesale retail access, product partnerships, merchandise and collector events.
Limit: Brand-reported distribution is not independently audited NFT revenue. A new design does not inherit this installed audience or its licenses.
Binance Research — Full-Year 2025 and Themes for 2026
Period: Year ending December 31, 2025.
Class: research report
The report describes broad NFT-price contraction and a 55% year-over-year decline in GameFi investment during 2025.
Limit: Historical sector context, not a 2026 year-end result. Game activity does not demonstrate demand for a standalone browser puzzle.
NFT Wash Trading: Quantifying suspicious behaviour in NFT markets
Period: 2022 research on an earlier market sample.
Class: primary research paper
The paper demonstrates that suspicious trading patterns can be investigated through transaction relationships.
Limit: Its historical findings are not used as a present-day wash-trading percentage. Proving independent demand requires more than aggregate volume.
The collections are original studio tests, unminted and unlisted. Comparator artists, marketplaces and brands are not affiliated with these collections.