Investing in Pokémon and Trading Card Games

A game designed for nine-year-olds has become a collectibles market that outsells most segments of fine art. The Pokémon Company has printed more than 85 billion cards, about 10 billion of them in the year to March 2026, and still could not meet demand. PSA graded 19.26 million cards in 2025, more than 11 million of them game cards, overwhelmingly Pokémon. In February 2026 a single 1998 promo card sold for $16,492,000, the most ever paid for a trading card of any kind. And the headline return everyone quotes — +3,821% since 2004, against +483% for the S&P 500 — is a last-sold, price-weighted index that counts survivors, carries stale prices, and ignores a round trip that costs a quarter of your money. This is the whole picture: a single manufacturer that prints to demand, a grading oligopoly that manufactures the only scarcity that actually trades, a modern-product mania that looks a great deal like 2021 sports cards, the fee stack, the fakes, the 28% tax, and how an outsider buys one card well. Every figure is sourced at the end.

On the evening of February 16, 2026, Goldin Auctions closed a 41-day sale that had drawn 97 bids. The lot was a 1998 Japanese promotional card — a Pikachu holding a paintbrush, awarded to the winners of a magazine illustration contest, never sold in a shop, roughly 39 copies known, exactly one of them graded a perfect PSA 10. It closed at $16,492,000 with the buyer’s premium, on a hammer Guinness put at about $13.3 million; a Guinness World Records adjudicator appeared on the stream to certify it.

The seller, Logan Paul, had bought it in a 2021 trade valued at $5.275 million and worn it into a WrestleMania ring on a diamond chain. The buyer, A.J. Scaramucci, runs a venture fund, and went on CNBC nine days later to argue that collectibles are an asset class.

Six weeks later The Pokémon Company closed the fiscal year in which it had printed roughly ten billion cards — about 11.7% of every Pokémon card ever made, in one year — and its May 2026 disclosure said demand had still outrun its production capacity. Somewhere between those two facts, by the summer of 2026, a Prismatic Evolutions Umbreon that had traded at $1,600 was changing hands at $832.

That is the entire asset class in one paragraph. A tiny, genuinely fixed supply of old paper sets records, while the same company floods the world with new paper at a rate no other collectible has ever seen, and the buyers of the new paper discover that “scarce” was a marketing word. The job of this dossier is to show you exactly where the line between those two markets runs, who draws it, and what it costs to stand on the right side of it.

Nobody publishes a clean number for “the trading card market,” and the reason is structural: the primary market is one company’s private product line and the secondary market is a hundred million small transactions on eBay. What can be measured is the plumbing, and the plumbing is enormous.

The manufacturer

The Pokémon Company — a joint venture of Nintendo, Game Freak and Creatures, with Nintendo holding about 32% of the votes and none of the three breaking out card revenue — reports cumulative production of more than 85 billion cards as of March 31, 2026, in sixteen languages across more than ninety countries. The running total was 64.8 billion at the end of March 2024 and 75 billion a year later, which puts the last two fiscal years at roughly 10 billion cards each. The company’s own results for the year to February 2026, surfaced through a filing in Japan’s official gazette rather than an investor deck, showed net sales of ¥531.4B, up 29.3%, and net profit of ¥120.1B, up 70.7%.

Nintendo’s share of profit from equity-method entities, which is mostly this company, more than doubled from ¥35.1B to ¥82.7B in its year to March 2026. This is not a hobby company. It is one of the most profitable consumer franchises on earth, and cards are the part of it that grew.

The retail tape

Circana, which measures US toy sales at the register, reported that the US toy industry returned to growth in 2025, up 6%, that the games-and-puzzles category rose 37% “mainly due to Pokémon,” and that the Pokémon property alone did $2.5B in US toy sales, up 87% on the year. That figure covers the whole property, not only cards, but cards are why it moved. In Japan, Bandai Namco’s 2025 integrated report sizes the domestic trading card market at ¥319.7B for the year to March 2025, up 11.7%, and describes North America as the largest market in the world.

The secondary market

eBay bought TCGplayer, the dominant singles marketplace for Magic and Pokémon, for up to $295M in October 2022. By its Q1 2025 call it was telling investors that trading cards had grown at double-digit rates for a ninth straight quarter, with sports cards the largest single contributor to that growth. eBay also owns Goldin, the house that sold the Illustrator. Fanatics owns PWCC, rebranded Fanatics Collect. The venue layer, in other words, has consolidated into two companies. The sports side of the same machine has its own pages on this hub: Investing in Sports Cards, and beneath it Investing in Vintage Baseball Cards and Investing in Modern Basketball and Football Cards; Card Grading Explained covers the grading machine across both games, Investing in Sealed Product the boxes, and Investing in Magic: The Gathering the Reserved List.

The grading tape

This is the number we trust most, because it is a count of physical objects. GemRate, which aggregates the graders’ population reports, counted 26.8 million cards graded by the major services in 2025, up 32% from 20 million in 2024, and for the first time trading-card-game and non-sports submissions (16.8 million, up 95%) outran sports cards (10.0 million, down 12%). PSA alone graded 19.26 million, more than 11 million of them TCG cards, and most of those Pokémon. A grading tracker’s mid-2025 report put Pokémon at 97 of the 100 most-graded cards of the year.

Put a dollar figure on all that and you get vendor estimates — the TCG sector at roughly $6–8B a year, trading cards overall at $13–15B, and Market Decipher’s “$496B collectibles” umbrella, which double-counts almost everything. The honest statement is narrower: a ten-billion-card-a-year printing operation, nine straight quarters of double-digit growth at eBay, and 27 million cards a year encased in plastic so they can be sold as something other than a game piece.

Read the bars as a supply curve. Nearly 40% of every Pokémon card in existence was printed in the last three fiscal years, and about half in the last four. The fine-art market has no equivalent of that sentence, and every argument in this dossier about modern cards follows from it.

The figures below are the ones we would defend in a meeting. Each is dated, and each moves.

Three of those cells are ours and are labelled as such. Our Pokémon basket is five pinned PriceCharting IDs — the Base Set Charizard, Blastoise and Venusaur, the 1st Edition Pikachu, and the Evolving Skies Umbreon VMAX alternate art that the hobby calls Moonbreon — and we take the daily median of their PSA 10 and ungraded prices. It is a reference basket, deliberately narrow, built so that every print traces to an observable price. It is not the market.

The Collector Crypt figure is a count of live listings on one tokenized-card marketplace, pulled from its public API every morning. That is not the market either. Both are on the board because you should be able to see exactly what a number counted before you trust it, and most of the numbers in this hobby do not come with that courtesy.

The Pokémon Trading Card Game launched in Japan in October 1996, eight months after the Game Boy titles, published by Media Factory. Wizards of the Coast — the company that had invented the trading card game with Magic in 1993 — licensed it for the English market and released Base Set on January 9, 1999. Everything that trades for six figures today descends from the accidents of that first print run.

1st Edition, shadowless, unlimited

The first English cards carried a small “Edition 1” stamp under the art and were printed with no drop-shadow on the right side of the picture frame. When the stamp was dropped, a brief run kept the shadowless frame; then the frame gained its shadow and the “unlimited” print began, and ran for years. Three objects, one design, three wildly different populations: the hobby estimates the 1st Edition run in the low millions of cards across the set, the unlimited run in the billions. The Charizard is the same drawing in all three. The 1st Edition PSA 10 is a $550,000 card; the unlimited PSA 10 is a low-five-figure card. Nothing about the card differs except the count.

The lull

Pokémania peaked around 2000 and collapsed by 2002. Wizards lost the licence in 2003 and The Pokémon Company took the English game in-house. For a decade the cards were a children’s product again. Binders went into closets. That decade is why the 1999 supply is genuinely fixed and genuinely attrited: nobody was preserving it.

2016 and 2020

Pokémon Go put the brand on every phone in the summer of 2016, and the first cohort of Base Set children, now in their late twenties, started buying back their childhoods. Then the pandemic. In October 2020 Logan Paul opened a sealed Base Set booster box on a livestream; a 1st Edition Charizard PSA 10 that had sold for $18,900 in July 2017 made a reported $220,000 that month and $295,000 the next, and $311,800 at PWCC in early 2021. Stimulus money, a captive audience and a YouTube economy that monetised pack openings did the rest.

2021: the shelf fights

On May 14, 2021 Target suspended in-store sales of Pokémon and sports cards after a parking-lot fight over cards at a store in Brookfield, Wisconsin; Walmart said it would keep selling. Retail restocks were being cleared by adults with carts at opening time. The Pokémon Company printed 9.7 billion cards in the year to March 2023 to catch up, then a record 11.9 billion the year after. On April 2, 2022, Paul wore the Illustrator into WrestleMania 38 with a Guinness certificate for the $5.275 million private-sale record.

The top of the 2021–22 cycle is easy to date: the 1st Edition Charizard PSA 10 made $420,000 in March 2022 and a reported $250,000 that November, a 40% drawdown in eight months.

2024–2026: the modern mania

The 151 set (2023), Prismatic Evolutions (January 2025) and a run of illustration-rare chase cards produced a second scalping wave: local shops receiving a fraction of their allocations, Elite Trainer Boxes clearing at two and three times retail on the day of release, and a company statement on January 15, 2025 that it was “actively working to print more of the impacted Pokémon TCG products as quickly as possible and at maximum capacity.” Walmart, from late August 2026, was rationing releases through an online “Collectibles Drawing” lottery.

In the same window the vintage market did what it does when new money arrives. The 1st Edition Charizard PSA 10 made $550,000 at Heritage in December 2025 and $954,800 at Goldin in February 2026, a copy from a Logan Paul box break; the next public copy made $528,000. The Illustrator made $16.49 million the same week.

Thirty years, three cycles, one pattern. Every boom has been a demographic event — a cohort reaching the age at which it can afford its childhood — amplified by a new distribution technology: eBay in 1999, YouTube in 2020, TikTok and live-stream “rip” selling in 2025. Every bust has been a supply event. Keep that asymmetry in mind; it is the whole thesis.

Strip the nostalgia away and the Pokémon card market is three near-monopolies stacked on top of each other, and each one controls a different dimension of the price.

The printer controls quantity

One company decides how many of every modern card exist, when a set is reprinted, and whether a chase card gets reissued in a later product. There is no mint-report equivalent for Pokémon: the company publishes a cumulative total once a year and nothing per set. Every estimate of a set’s print run you will read is a community reconstruction from pull rates and grading populations, and the serious ones carry error bars of three to five times.

The grader controls quality, and therefore scarcity

A raw card has no price anyone can quote with confidence; a PSA 10 has one. Since the thing that sells for $550,000 is not “a 1st Edition Charizard” but “a 1st Edition Charizard that PSA has called a 10,” the supply of the investable object is set by a private company’s graders, one card at a time, at roughly 100,000 cards a business day in 2026.

PSA has about 72% of that volume. Its parent, Collectors, was taken private in February 2021 for about $850M by a group led by Nat Turner with D1 Capital and Steve Cohen’s family office. CGC, majority-owned by Blackstone’s Tactical Opportunities arm, has 18% and grew 121% in 2025. Beckett, once the premium name, is at 3%. None is listed. None publishes its gem-rate policy. The two companies that manufacture the market’s scarcity are owned by a hedge-fund founder and a private-equity firm.

The tape controls price discovery

eBay is the clearing house for singles and slabs; TCGplayer (eBay) sets the “market price” that every price-tracking app and every local shop imports; Goldin (eBay) and Heritage handle the six- and seven-figure lots; Fanatics Collect handles the rest of the catalogue business and runs a vault. Price guides — PriceCharting, Card Ladder, PSA’s own price guide — are downstream of eBay sold listings. When eBay reports that trading cards accelerated for a ninth straight quarter, it is describing the velocity of the entire secondary market.

Who is on the other side

Four kinds of seller. The ripper opens sealed product on camera; his cost basis is the box, so he sells a chase card at any price above zero. The flipper bought retail and is arbitraging the allocation shortage; he disappears when the reprint lands. The consignor is a long-term holder paying an auction house 0–15% to reach the top bidder, and usually knows more than you. The vault liquidator is a platform or fund unwinding inventory, the most price-insensitive seller of all. In modern cards you are buying from the first two; in vintage, from the third.

In every other collectible we cover, the maker is dead, defunct or constrained. Ferrari built 400 Enzos and stopped. Patek makes a few hundred of a reference. The artist dies. Pokémon is the only major collectible whose manufacturer is alive, profitable, growing 29% a year, and explicitly committed to printing as many of the things you own as people will buy.

Modern sets are print-to-demand

Since the Sword & Shield era the English game has had no announced print caps and multiple print waves per set. When Prismatic Evolutions sold out in January 2025, the company’s response was not a shrug; it was a public statement, two days before release, that it was printing more at maximum capacity, followed by reprint waves through the year.

Its printer, Millennium Print Group in North Carolina, signed the largest US manufacturing lease of 2025 in December: a 1.27 million-square-foot campus in Morrisville, with full operations expected in 2028. That is capacity being built for exactly this. A sealed Prismatic Evolutions box bought at $300 over MSRP in the shortage was a bet that the company would fail to do the one thing it had promised in writing to do.

Reissues erase “chase”

The 2021 Celebrations set reprinted the Base Set Charizard, Blastoise and Venusaur art in a modern holofoil for the 25th anniversary. The 2023 151 set reprinted the original 151 Pokémon with new art. A 2026 30th-anniversary product line is doing it again. None of these reprints touches the 1999 cards’ population, but every one of them satisfies some of the demand that would otherwise chase them, and every one of them creates a new “Charizard” for a buyer who wanted the character rather than the print run. The 151 Charizard ex has been graded roughly 96,000 times; PSA has handed out about 27,400 tens. That is the manufacturer converting a $500,000 scarcity into a $200 abundance without reprinting a single 1999 card.

Vintage is fixed but not closed

The Wizards-era sets (1999–2003) cannot be reprinted in their original form; the licence ended and the company has no incentive to reissue an “Edition 1” stamp. That is real. But the tradeable supply of vintage — the gem-mint population — is not fixed, because millions of raw 1999 cards are still in binders, and rising prices pull them out and into the grading queue.

Community trackers of PSA’s population report recorded the PSA 10 counts of Base Set starters such as Charmander, Bulbasaur and Squirtle rising by hundreds to thousands within single months in 2026; we could not reconstruct the monthly deltas ourselves. Some of that is the backlog clearing, some is crack-and-resubmit double counting. The direction is not in doubt. The Charizard’s 1st Edition gem population is small (about 125 as of December 2025, per Heritage) because the card is genuinely hard to find clean, not because the supply of candidates has run out.

The sports-card precedent is the whole warning

The 1987–1994 “junk wax” era of baseball cards was a manufacturer printing to demand until demand stopped; three decades later those cards are worth less than the cardboard, and the Sports Cards Dossier has the autopsy. Modern Pokémon at ten billion cards a year is junk wax with better art, and the only reason it is not priced that way yet is that the demand has not stopped. It will not stop because of a reprint; it will stop when the cohort buying it moves on, and the print run will still be there.

If you understand one mechanism in this market, make it this one. Grading is where a game piece becomes a financial instrument, and the grader is paid per card regardless of what grade it assigns. Card Grading Explained, on this hub, is the full guide to the graders, their fees and the population report.

What the grade is

PSA scores centering, corners, edges and surface on a 1–10 scale. A 10 (“Gem Mint”) requires front centering between 55/45 and 60/40, back centering no worse than 75/25, corners flawless under magnification, and a surface with no print lines, scratches or holofoil scuffs. A 9 is allowed one minor flaw. The difference between 9 and 10 is frequently invisible at arm’s length and worth a multiple of the price. CGC adds a “Pristine 10” above its 10 and BGS a “Black Label” 10 with four perfect sub-grades; both exist to sell an even thinner slice of the population, and both trade at premiums that have proven unstable when the grader’s market share does.

Market share

On GemRate’s 2025 count PSA had 72% of the volume, CGC 18% (up 121% year on year), BGS 3.1% and TAG 1.7%; the counts are in the chart below. The resale premium follows the share: a PSA 10 of a given card consistently sells for more than a CGC 10 of the same card, and a BGS 9.5 sits below both. That premium is not a statement about the card. It is a statement about which label the buyer pool trusts, and it moves. Beckett was the premium label in 2010 and is 3% of the market now.

Fees

PSA’s published price list in 2025 ran from about $19–$25 a card at the bulk and value tiers (65-day turnaround) through $75–$150 at Regular and Express, to $600 and up for Walk-Through service on high-declared-value cards, plus membership, a handling fee and return shipping with insurance. CGC undercut it until its March 24, 2026 tier update put Bulk at $17 and Economy at $20. By September 2026 PSA had paused its Value tiers, its $79.99 Regular tier (renamed Priority) ran at an estimated 70–80 business days, and a $59.99 Standard tier opened on September 14 at 90–100.

On June 2, 2026 PSA paused all four Value tiers (Value Bulk had been $24.99) after a 20% spike in submissions pushed its backlog toward 10 million cards. That made $79.99 Regular the entry point for a single card until the backlog falls to 5 million, and PSA published a public backlog tracker so submitters can watch the number. While that holds, the economics of grading a $60 modern card have simply ended, and the population growth in modern PSA 10s will slow for a reason that has nothing to do with the cards.

The gem rate is the product

GemRate’s count through mid-2025 put PSA’s gem rate at 43% of all submissions (45% in 2024), with the TCG category near 50% and sports cards at 34%. A grader that gave out fewer tens would be tougher, more prestigious, and less popular with submitters. A grader that gave out more would sell more slabs this year and destroy its label next year. Every grading company lives on that knife, and the population reports let you watch them walk it: when the monthly PSA 10 count on a modern card jumps while submissions are flat, either the cards got better or the grading did.

Half of all modern Pokémon cards submitted to PSA come back as a 10. About two per cent of 1st Edition Charizards do. That ratio, not the year on the card, is the actual thing you are paying for when you pay for vintage. It is also why a PSA 10 on a modern card is a grade, not a rarity: the population is whatever the submission volume decides it is.

The label is the asset, which is why the label is what gets counterfeited; the fraud section below covers the January 2026 federal conviction and the 2026 resealed-slab and acetone methods. The short version: a slab is only as good as the venue that re-authenticated it.

Here is our own tape, and it makes the point better than any argument. On September 8, 2026, the median PSA 10 price across our five-card Pokémon basket was $5,257. The median ungraded price of the same five cards was $90. Same cardboard, same art, same year: 58 times the money for a plastic case and a number.

A note on that series, because we said we would always tell you what a number counted. The PSA 10 median printed $4,665 on September 1, sat at $5,000 for three days, moved to $5,257 on September 5, printed $3,262 on September 7, and returned to exactly $5,257.31 on September 8. The board therefore shows a “+61%” one-day move. Nothing in the market moved 61%. On the 7th one or more of the five components failed to return a quote, the median of the remaining cards landed on a different card, and the next morning the full basket was back. The ungraded series shows the same dip and recovery ($107.50 to $73.66 to $90.12).

We are leaving the print on the tape rather than smoothing it, and telling you it is an artefact of a five-card basket rather than a price. A hobby index with a thousand components would hide the same thing; you would simply never know.

Now the cliff at the top of the market. The 1st Edition Base Set Charizard sold at $550,000 in a PSA 10 in December 2025 and $954,800 at Goldin two months later. A PSA 9 of the same card traded in the $20,000–$50,000 range across the same season, per the hobby’s price trackers. That is a gap of ten to twenty-five times for a single grade step, on a card whose 9-to-10 difference is a corner you need a loupe to see. Below that, the PSA 8 is a $15,000–$30,000 card and a raw, played copy a low-four-figure one. The value of a vintage Pokémon card is not a curve; it is a staircase with one enormous top step, and the grader decides who stands on it.

Three consequences follow. First, grading is a lottery ticket with a known price: a raw Charizard that looks clean is a ~2% shot at a $550,000 outcome and a ~98% chance of a $15,000–$50,000 one, and the market prices the raw card at roughly that expected value. Second, the cliff invites fraud: every method in the fraud section — trimming, pressing, recoloring, slab swapping — exists to move a card up one step. Third, the cliff is a policy, not a law of nature. If PSA tightened its 10 standard by one hair, or CGC’s Pristine label gained buyer trust, the staircase would re-price overnight without a single card changing.

The number the industry likes: Card Ladder’s Pokémon index is up 3,821% since 2004, against 483% for the S&P 500, a comparison the Wall Street Journal ran in September 2025 under the headline “The Hot Investment With a 3,000% Return? Pokémon Cards.” Fortune ran a version in July 2025 with 3,261% against 421% over twenty years. By 2026 Card Ladder’s own press release put the twenty-year gap at “nearly 8x.” In the twelve months to February 2026 the index rose 145% while the S&P returned 15.2%, per CNBC, and it was up about 28% year-to-date by mid-2026.

Now the number in context, and it needs a lot of context.

It is a last-sold, price-weighted index

Card Ladder publishes its method: every card in an index is carried at the price of its most recent sale, anchored to the index between sales, and the index is the sum of those prices under a stock-market-style divisor that keeps additions and removals from moving it. A card that last traded in 2019 is still in the index at its 2019 price. In a falling market owners stop selling, so the index does not fall; it stops updating. Because it sums prices, a $550,000 Charizard moves it more than a thousand $30 commons, so nobody could hold the index if they wanted to, and no fund tries.

Card Ladder’s own documentation walks through the case where one million-dollar sale lifts an entire sub-index by 50% while nothing else changes, calls it non-representative, and explains the divisor it uses to damp it. That candour is to their credit and has been read by almost nobody who quotes the 3,821%.

It is a survivor index

The cards in it are the cards that still trade. The billions of cards that lost value and stopped trading — which is nearly all of them — are not in the denominator. Goetzmann showed in 1993 that art-sale indices built this way show positive returns in thin periods even while the underlying market falls, because owners sell winners. Korteweg, Kräussl and Verwijmeren (RFS 2016), the paper the Art Dossier is built on, corrected for that selection in 20,538 repeat painting sales and cut the average return from 8.7% to 6.3% and the Sharpe ratio from 0.27 to 0.11. The card index is built from the same kind of data in a more retail, more sentiment-driven market. There is no reason to expect the correction to be smaller.

It is a “math crime”

That is analyst Bo Hanson’s phrase for stacking an average of thousands of collectibles against a diversified index of five hundred companies. The fair comparison for the best card is the best stock, and over the same twenty years a handful of stocks did as well. A July 2026 24/7 Wall St. piece titled “Pokémon Cards Beat the S&P 500 by 2.5x, But the Math Is a Lie” covers the same ground; it is unusual for a trade publication to say so out loud.

What the flagship card actually did

The single most tracked object in the hobby is the 1st Edition Base Set Charizard in a PSA 10, and its public sales tell the honest story.

From $18,900 in July 2017 to $550,000 in December 2025 is a 29-times gain, roughly 49% a year, and there is no stock index on earth that did that. But read the middle of the chart. From the March 2022 peak of $420,000 the card fell to a reported $250,000 in eight months, a 40% drawdown, and did not make a new high for nearly four years.

From that peak to December 2025 it returned about 31%; the S&P 500 with dividends returned roughly 60% over the same span (2022 −18.1%, 2023 +26.3%, 2024 +25.0%, 2025 +17.9%, prorated from March 31, 2022), with daily liquidity and no 20% buyer’s premium. Anyone who bought the best card in the hobby at the top of the last cycle trailed an index fund until the February 2026 print. Nobody quotes that window.

The “Charizard beat Apple” genre

Thirty dollars in the S&P 500 in January 1999, dividends reinvested, is worth about $290 today. Thirty dollars of 1999 Base Set packs that happened to contain a Charizard that happened to grade a 10 is worth $550,000. That comparison has been made in a thousand articles and it is survivorship in its purest form: the single most extreme outcome from billions of printed cards, against the average of the stock market. The honest version is: thirty dollars of 1999 packs, on average, is worth about thirty dollars, minus the cost of finding out.

The peer group says something different

Knight Frank’s Luxury Investment Index, which tracks ten passion assets with decades of history, fell 3.3% in 2023, 2.7% in 2024 and 0.4% in 2025. The Liv-ex 1000 fine-wine index fell 28.8% peak-to-trough on our tape, from October 2022 to August 2025. In the same window the Pokémon index rose 145% in a year. Either cards are a perfectly uncorrelated diversifier that marched through a three-year collectibles bear market untouched — which is an enormous claim — or the two indices are measuring different things. The methodology above says which.

Our position: the verifiable record for top-grade vintage Pokémon is extraordinary and beat equities over any window starting before 2020. It is also thinner, more survivor-selected, costlier to realise and more volatile than the index makes it look, it trailed stocks for three years after the last peak, and it says nothing about the modern cards most people actually buy.

The record books are short, recent and dominated by two kinds of object: Japanese promotional cards that were never sold, and English 1999 cards in a perfect grade. Dates and venues matter, so here they are.

Pikachu Illustrator

Awarded in 1998 to winners of CoroCoro Comic illustration contests; about 39 exist; drawn by Atsuko Nishida, who designed Pikachu. Exactly one is a PSA 10. Logan Paul traded a PSA 9 (valued at $1.275M) plus $4M for it in July 2021, took the Guinness private-sale certificate at WrestleMania 38 on April 2, 2022, and consigned it to Goldin’s 2026 Pokémon & TCG Auction, where it closed on February 16, 2026 at $16,492,000 including the buyer’s premium after 97 bids: it sat at $6.88M with a day left and went from there to a $13.3M hammer in the extended-bidding hour. The prior record for any trading card was the $12.932M Jordan–Bryant dual Logoman at Heritage in August 2025.

The buyer’s premium point matters: on a $13.3M hammer, Paul’s widely reported 3.1-times gain was nearer 2.5 times before any seller commission. Still the best trade in the hobby’s history. Still smaller than the headline.

Trophy cards

Below the Illustrator sits a tier of Japanese tournament prizes with populations in the single and low double digits: the No. 1, 2 and 3 Trainer cards, the Trophy Pikachus, the Kangaskhan family-event card, the Tropical Mega Battle promos, the Master’s Key. They trade privately or at Goldin and Heritage, in the high five to low seven figures, to a buyer pool small enough that the houses know the names. The 1998 Blastoise “presentation” prototype belongs to the same category: a blank-backed test print, one of two made to pitch the English game to Nintendo, graded CGC 8.5, it made $360,000 at Heritage on January 14, 2021. These are objects whose value is the story of the game’s manufacture, not the game.

Japanese promos in the mid-market

Our own notable-sales tape shows where the deep money sits below the trophies: five eBay sales in the first week of September 2026, three of them Japanese promotional cards.

Note what those five have in common. Three are promotional cards with fixed, small distributions (a 2010 campaign giveaway; two cards from a 2018 Munch exhibition tie-in at a Tokyo museum). One is a 2001 1st Edition holo from the last Wizards-era set. One is a 2015 full-art from a set that was printed in volume but graded a 10 far less often than modern product. None was printed in the last decade at a rate the manufacturer could dial up. That is the profile of the mid-market card that holds a five-figure price, and it is a very different profile from the modern chase card that reaches five figures for a month.

The best-documented segment of this market is also the one where the most money has been lost, because the data is a marketplace feed rather than an auction archive. Two serious independent datasets cover it, and they disagree in a way that is itself the lesson.

The archive that says “up”

An open, community-built price archive (pokestat, hosted on GitHub; we read it but could not independently audit it) of 108 Elite Trainer Boxes across 60 sets, built from TCGplayer market prices from March 2024 to September 2026, reports a chained index up 59.1% a year over the window, a 3.19-times multiple in two and a half years, with the median box going from $67.99 to $218.37.

Its own authors then do what almost nobody in the hobby does: they show that fifteen defensible ways of building the same index give answers from 58.8% to 90.5% a year, that the higher numbers come from fixed baskets that can only contain survivors, that booster boxes did the same (59.6%) and singles less (39.8%), and that the archive covers a single boom whose deepest drawdown was 3.4%.

The archive’s most useful finding is about friction: sales tax in, marketplace take out and shipping cost 20.2% of the box plus $12.30, charged once, so a three-month flip lost money 80% of the time during the boom, a twelve-month hold netted 23% a year, and flat prices for a year would produce a real loss.

The analysis that says “down”

A November 2025 analysis attributed to Apex Intelligence of 50,842 Pokémon sales from January to October 2025 (a document we could not retrieve directly; TCGplayer’s own 2026 price-trend reports corroborate the direction) found 2020–2024 sealed product down 40–60% from its late-2024 prices: Lost Origin booster boxes from $175 to $85, Paldea Evolved boxes by half, Crown Zenith Elite Trainer Boxes from $90 to $45. Its Evolving Skies row, a box falling from $220 to $95, is contradicted by three price trackers that put the same box at $2,600–2,700 in July 2026, so we omit it from the chart and treat that row of the study as doubtful; the sealed-product section returns to it. The Umbreon VMAX alternate art — the card in our own basket — went from $450 to $120, a 73% fall. By August 2026 a separate market summary had the Prismatic Evolutions Umbreon ex illustration rare at $832 from $1,600, the Obsidian Flames Charizard ex at $79 from $126, and modern singles generally 20–50% below their 2024–25 peaks.

The two datasets do not contradict each other. They measure different things, and the difference is the lesson. The archive’s index is chained across every product, including 37 that entered mid-window at pre-order quotes and rose from there; it measures the category’s momentum. The crash figures are specific boxes and cards that a person actually held from the 2024 peak; they measure a holder’s experience. If you own a particular box, the second number is yours. If you are asking whether the modern sealed category as a whole is still rising, the first is, and its trailing twelve-month rate (41.9%) is already lower than its first-half rate (65.8%). A category index that rises while its most-owned components fall is what the late stage of a product cycle looks like.

The 2021 sports-card rhyme

The pattern is not new, and the Sports Cards Dossier covers its last run in full. In 2020–21 modern sports cards ran on the same fuel — stimulus, streamed pack-breaking, a grading backlog, a retail shortage, Target and Walmart clearing shelves — and the best sales record for the sports hobby is still crowded with 2021 prints. Modern sports cards then gave back half or more of their gains through 2022, and many rookie cards never recovered while vintage set records again in 2024–26. Pokémon in 2024–26 has every element of that list plus one the sports market never had: a manufacturer that responds to a shortage by promising, in writing, to print more.

Where the modern money goes when it leaves

The 2025 sales study found the “middle class” of the hobby — cards between $50 and $200 — fell hardest, about 65% on average, while sub-$20 and $1,000-plus cards held. Graded vintage in PSA 7 and above rose strongly in the same window. That is the signature of a market rotating from speculation toward the objects with a real supply constraint, and it is why the Charizard and the Illustrator set records in the same quarter that Prismatic Evolutions fell by half.

“Sealed appreciates” is the hobby’s most repeated thesis, and it has a real mechanism behind it: every box that gets opened reduces the supply of unopened boxes forever, the buyer takes no grading risk, and a box is the only way to own a set’s entire chase-card lottery at once. A 1999 Base Set booster box, of which some unknowable number survive unopened, made $408,000 at Heritage in January 2021. A box from the 2003 Skyridge set, the last of the Wizards era, made $212,500 at Heritage on June 27, 2026 against a prior record of $90,000. The thesis is true for those objects.

It is true for them because three conditions held: the set is retired and cannot be reprinted, enough time has passed that attrition has done its work, and the set contains cards people still want. For modern product only the third condition can be checked, the second takes a decade, and the first is exactly what the manufacturer refuses to promise. A sealed Evolving Skies box is the sealed thesis’s best modern case — retired in 2023, beloved chase card — and three price trackers put it at $2,600–2,700 in July 2026, about 18 times its $144 list price (the November 2025 study’s $220-to-$95 path for the same box is contradicted by every tracker we read, and this guide treats it as doubtful). It is the exception the manufacturer permitted, not a rule: every set still in print is repriced by the next print run, and Investing in Sealed Product on this hub works both cases from list price to after-tax proceeds.

What a box costs to hold

Storage and insurance are trivial; the one-time friction of 20.2% plus $12.30 and the 28% tax are not. A $500 box must appreciate roughly 26% before the seller is whole, and the archive found a hold only paid when it ran past about eight months, in a market that rose the whole time.

The fraud surface is worse than for singles

A sealed box is a promise that nobody has touched the inside. Four documented ways they have. Pack weighing: vintage packs with a holofoil card weigh measurably more than packs without one, so loose vintage packs are assumed picked-over unless they come from a sealed box; the method is migrating to modern products with heavy foil hits. Mapping: print sheets are cut in a fixed order, so a careful observer can learn which pack positions in a box carry the chase cards, and shops and distributors with early product have been accused of pulling those packs before resale; the 2024–25 shortages produced a wave of such accusations around 151 and Prismatic Evolutions boxes.

Resealing: packs and boxes opened, cherry-picked and re-glued, sold as factory-fresh. Label fraud: in early 2026 collectors documented and reproduced a method that reprints the thermal shipping label of a high-demand case — an Evolving Skies label applied to a case of a cheap set — so that the “factory sealed case” is genuine cardboard around the wrong product.

The defences: buy sealed from the distributor chain or from a house that guarantees it; treat any “sealed vintage” box without a chain of custody as opened; and for a serious box, pay a grader to encapsulate it so the object you resell is verifiable.

Every card exists twice: a Japanese print that comes first and an English print (plus fourteen other languages) that follows by months. For a collector the two are different assets with different supply, different quality and different prices, and the gap between them has been one of the hobby’s most reliable trades.

Quality

Japanese cards are printed on different stock with tighter centering and fewer print lines, so they gem at materially higher rates than their English twins: good for a buyer of raw Japanese cards, bad for the scarcity of Japanese PSA 10s, whose populations run large. The English PSA 10 of a modern card is the rarer object and is priced accordingly.

Price

Through 2021–23 Japanese versions of modern chase cards traded at 30–50% discounts to English. By late 2025 that gap had narrowed to 5–15%, because Japanese cards fell less in the modern correction: the Umbreon VMAX alternate art went from $450 to $120 in English and from $280 to $105 in Japanese, by the same 2025 analysis cited above.

Three things pushed the same way: smaller print runs (hobby estimates put Japanese runs 40–60% below English, and nobody outside the company can verify them), a domestic collector culture that preserves rather than speculates, and a weak yen that made Japanese product cheap for dollar buyers. The arbitrage that paid for a decade is now a rounding error on modern cards; it still exists on older sets where English populations are thin.

The Japanese record book

Nearly every card in the top of the record books is Japanese: the Illustrator, the Trophy cards, the campaign promos, the museum tie-ins on our tape. The reason is distribution: Japan ran tournament-prize and promotional programmes with fixed, tiny print counts for thirty years, and English-market promos were rarely as scarce. If you want a card whose supply is genuinely known and genuinely closed, the odds are it has Japanese text.

Import mechanics

The deep liquid market for Japanese cards is domestic: Mercari Japan and Yahoo! Auctions Japan, reachable from abroad only through proxy services (Buyee, ZenMarket, FromJapan and others) that bid on your behalf, charge a service fee of roughly 5–10% plus domestic and international shipping, and consolidate. Grading in Japan runs through PSA’s Tokyo office (opened November 2018) or by shipping to the US. Two 2025 changes matter for a US buyer: the end of the $800 de-minimis duty exemption on August 29, 2025, so that card parcels from Japan now clear customs with duty and brokerage, and a US tariff schedule that moves; check the landed cost before the bid, because a 15% surprise on a $5,000 card is the whole margin.

Magic: The Gathering invented the category in 1993 and is the only trading card game whose manufacturer has ever made a binding promise about supply; Investing in Magic: The Gathering, on this hub, is the full guide. It is also, right now, the most commercially aggressive publisher in the hobby. Both facts matter to an investor, and they pull in opposite directions.

The Reserved List

In 1995 Wizards of the Coast reprinted early cards in a set called Chronicles and watched the collector market for the originals collapse. Its answer, in 1996, was a list of cards it would never reprint in a tournament-legal form; the list was finalised in 1999 at 572 cards, all from 1993–1999 sets, with no additions or removals since. It is the closest thing in any collectible to a Bitcoin-style hard cap, and it is why an Alpha Black Lotus is the most valuable game card after the Illustrator.

It is also not a floor. Reserved List prices fell sharply from their 2021–22 peak in the same liquidity drawdown that hit Pokémon and sports cards, because a fixed supply does nothing when the demand is what moves. The policy bars reprints and functionally identical cards under new names; it says nothing about demand, and every new set that does the same job for play erodes the demand under the lower tiers of the list. The Power Nine (Black Lotus, the five Moxen, Ancestral Recall, Time Walk, Timetwister) and the original dual lands are the tier that behaves like an asset; the rest of the list trades like thin over-the-counter stock, vulnerable to coordinated buyouts.

The records

An Alpha Black Lotus graded CGC Pristine 10 — the only one — sold privately in 2024 for $3 million, through Adam Cai of Pristine Collectibles, the Magic record. A PSA 10 Alpha Lotus, signed by the artist Christopher Rush, made $511,100 on eBay through PWCC in January 2021 and $540,000 at PWCC in March 2023. The serialised 001/001 One Ring from the 2023 Lord of the Rings set went to Post Malone in a private deal reported at more than $2 million, the first time a card printed that year became a top-tier trophy on release, and a signal of what the company was about to do.

Universes Beyond

Hasbro, Wizards’ owner, reported Magic revenue of $1.72B in 2025, up 59%, the franchise’s best year, on the strength of licensed crossover sets: Final Fantasy (June 2025) became the highest-selling set in Magic’s history by net revenue within months, per Hasbro’s FY2025 10-K; Lord of the Rings had crossed $200M in its first year, Hasbro said in 2024; Spider-Man and Avatar followed. Collector Boosters — premium packs sold at $200 and up a box with serialised, foil-treated chase cards — and the direct-to-consumer Secret Lair drops have turned Magic into a collectibles business that happens to have a game attached.

For the secondary market that means the same thing it means in Pokémon: a publisher that can produce a new “chase” every quarter is a publisher whose modern chase cards will not stay chased. The Reserved List is the only part of Magic where that logic does not apply, which is precisely why it trades at the prices it does.

We could not verify a 2025–26 Reserved List index level from a primary source for this build; dealer commentary describes a recovery from the 2022–23 trough alongside the broader card boom, and we would treat that as anecdote until a repeat-sales series says otherwise.

Three other games matter to the asset case, and each teaches something the Pokémon story does not.

One Piece (Bandai, July 2022)

The fastest-growing game of the decade, and one of the two categories, with Magic, that GemRate names alongside Pokémon when it counts PSA’s TCG volume in 2026. It launched in Japan in July 2022 and reached global scale by 2023, riding the manga’s anime audience and a supply model that, at least initially, was tighter than Pokémon’s: early sets sold out, English allocations were thin, and the first-print Manga-art Luffy and the tournament prize cards became five-figure objects within eighteen months of existing.

Our own tape recorded a 2022–23 Treasure Cup Roronoa Zoro promo in a BGS 10 Black Label at $40,000 and a 2025 third-anniversary gold Luffy in a BGS 10 Pristine at $16,900, both on eBay on August 27, 2026. The lesson: a new game can produce trophy prices in three years if the publisher controls early supply and the prize cards are genuinely few. The risk is the same publisher deciding, as it grows, that it would rather sell ten billion cards a year.

Yu-Gi-Oh! (Konami, 1999)

The oldest continuously published game after Magic and Pokémon, with a competitive scene that still drives most of its demand. Its record book is dominated by tournament prize cards with single-digit populations and by 1st Edition Legend of Blue Eyes cards in PSA 10, which trade in the five figures. The stainless-steel Tournament Black Luster Soldier, the 1999 Asian Championship prize, was listed at $10 million in 2013 and reportedly changed hands privately for about $2 million. Konami reprints aggressively and the game’s play formats churn, so the collectible value sits almost entirely in early prints and prizes. The lesson: a game can be enormous and its cards can still be a poor store of value, if the publisher treats reprints as a balancing tool.

Disney Lorcana (Ravensburger, August 2023)

The fastest product launch in Ravensburger’s history and, by 2025, past a billion cards sold. Its first-chapter “enchanted” cards briefly commanded four-figure prices as a new hobby chased a new scarcity. Then supply caught up: the publisher printed to demand, the collector premium normalised, and by 2025 group sales were reported down as the launch hype faded. The lesson is the clearest in the category: a new game’s first-print premium is a bet on the publisher under-printing, and no publisher owned by a toy company under-prints on purpose.

The venue you choose sets your round-trip cost more than any decision except the card itself, and the spread between venues is wider than in any other collectible we cover.

eBay

The default. The published final value fee for the Trading Cards category is 13.25% of the total the buyer pays (item, shipping and sales tax collected) up to $7,500 per item, then 2.35% on the portion above, plus a per-order fee of $0.30 (orders of $10 or less) or $0.40; an eBay Store subscription lowers the headline rate about a point. Promoted listings add 2–30% if you use them. The buyer pays no premium. At $200 and above (lowered from $250 in July 2026) a single card is routed through eBay’s Authenticity Guarantee, where PSA inspects the card (or the slab) before it reaches the buyer. It is free to both sides and it is the single most useful anti-fraud feature in the hobby.

TCGplayer

eBay’s other marketplace, and the one that sets the “market price” every app quotes. Its 10.75% seller commission plus 2.5% and $0.30 of payment processing sums to about the same 13.25% as eBay’s fee, and its Direct programme lets a buyer take cards from twenty sellers in one parcel. It is the venue for raw singles under $500; graded and vintage flow to eBay.

The auction houses

Heritage charges the buyer a 20% premium on the hammer price; Goldin (eBay-owned since 2024) charges 22% with a $19 minimum; both charge the consignor a commission that runs from zero, for a desirable slab the house wants, to 10–15%. Fanatics Collect (the former PWCC, rebranded July 2024) runs a comparable premier-auction structure alongside a weekly marketplace and a vault; verify its current schedule before consigning. Settlement runs 30–60 days after the sale. These venues exist for the top of the market — the Illustrator, the $550,000 Charizard — where a catalogue, an authenticity inspection and a room of bidders who have already wired deposits are worth the spread. Below about $10,000 they rarely are.

Japan

Mercari Japan and Yahoo! Auctions through a proxy, as above: 5–10% service fee, two legs of shipping, and, since August 2025, duty on entry.

Shops, shows and streams

A local shop buys at 60–80% of market and sells at market; the spread is the price of cash today. Card shows are where six-figure vintage changes hands privately and where the fake-slab losses of 2025–26 — one collector out $28,000 on a single purchase — happen. Live-stream selling (Whatnot, TikTok Shop, eBay Live) is the fastest-growing channel for modern product and the one with the least buyer protection; it is where the ripper sells.

Read the chart as a hurdle. Between what the buyer of a $6,000 card pays and what its seller keeps, 13% to 30% vanishes. A card must appreciate by that much before a holder is back to even, and it must do it again on the way out. That is why the honest annual return on a five-year hold is so much lower than the price chart, and why the worked example below is not encouraging.

Assume a US buyer in a 32% federal bracket, in a 5% income-tax state, buys two things on eBay in September 2026: a $5,000 PSA 10 vintage holo (a Neo-era or Jungle 1st Edition card, say) and a $500 sealed booster box from a retired 2021 set. Both appreciate at the rate the optimists use for good vintage — 8% a year for the card, 10% for the box — and are sold after five years.

Going in

eBay collects sales tax at the buyer’s rate, call it 7%: $350 on the card, $35 on the box. Insured shipping in, $30. Basis: $5,915. (Sales tax and shipping are part of basis; keep the invoices.)

Holding

A collectibles rider on a homeowner policy runs about 1% of scheduled value a year: roughly $55. A safe-deposit box or a proper safe, $60 a year. Five years: $575, not deductible for an individual holding as an investment since 2018.

Going out, route A: eBay

The card sells for $7,347 (8% compounded) and the box for $805. eBay’s 13.25% is charged on the buyer’s total including the 7% sales tax eBay collects, so it takes $1,042 on the card and $114 on the box; per-order fees, $0.80; insured shipping out, $55. Net proceeds: $6,940.

Going out, route B: auction house

If the room will pay the same $7,347 all-in, a 20% buyer’s premium (Heritage; 22% at Goldin) means the hammer is $6,122. With a 10% seller commission the consignor receives $5,510; with the commission waived, $6,122, paid 30–60 days later. Unless the house finds a bidder who pays more than eBay’s market, the auction route nets less on a $5,000 card. It earns its spread only at the top of the market.

Tax (route A)

Amount realised, net of selling costs, $6,940; basis $5,915; gain $1,025. The federal rate on a long-term collectibles gain is the taxpayer’s ordinary rate capped at 28%: here 28%, $287. Net investment income tax may add 3.8% ($39) above the income threshold; state, 5%, $51. After tax: $648. No 1099-K arrives — two transactions is far below the 200 — and the gain is fully reportable anyway.

The result

$648 after tax, minus $575 of carry, is $73 on $5,915 over five years. The price charts said 8% a year. The round trip paid about 0.25% a year. At 12% a year appreciation on both pieces the same arithmetic yields roughly 3% a year after everything; at 5% a year it loses about $540. The fee stack, the tax rate and the carry do not care what the card is, and they consume roughly the first eight points of annual appreciation on a five-year hold at this ticket size.

The newest venue in the category is a vault with a blockchain in front of it, and it has grown faster than any card marketplace since eBay. It deserves both the attention and the scepticism.

How it works

A platform takes delivery of a graded card (PSA, CGC or BGS — raw cards are not accepted), verifies the cert, photographs it, stores it in an insured vault, and mints a token on a public chain that represents that specific slab. The token trades on the platform’s marketplace and on open NFT venues; a holder can redeem at any time, which burns the token and ships the slab. Ownership transfers in seconds with no shipping, no grading risk and no eBay fee. Custody, insurance and the honesty of the vault are the whole counterparty question. The NFT Dossier covers the chain side; here the token is treated as what it is, a warehouse receipt.

Collector Crypt (Solana)

The largest. By August 2026 it had tokenized more than 130,000 graded cards and crossed $1.6B in lifetime volume, about 64% of Solana’s tokenized-card market, and by June 2026 it had overtaken Courtyard in weekly revenue. Its most popular product is not the marketplace but a “gacha” machine: randomised digital packs of tokenized cards bought with crypto, with an instant buy-back button for whatever you pull.

Redemption burns the token and ships from a vault in Delaware, for a 2% vault-withdrawal fee on the card’s insured value plus shipping, per CoinGecko’s platform guide; we could not read the platform’s own terms page directly. Its native token, CARDS, trades separately, which means the platform’s economics are partly a crypto-token story rather than a card story.

Courtyard (Polygon)

The other. Custody with Brink’s, about $37M of venture funding, and monthly volume that reached about $50M by mid-2025 and $78M in August 2025 — when tokenized Pokémon volume across all platforms reached $124.5M in a month, per Crypto Briefing. For stretches of 2025–26 trackers such as DappRadar ranked Courtyard the largest NFT collection in the world by weekly volume, ahead of Bored Ape Yacht Club and CryptoPunks.

Then the other half of the story: Courtyard’s floor price fell about 96% from a February 2026 peak, as Forbes reported in June 2026. A vault of real cards does not fall 96%; a marketplace of tokens priced in a speculative asset can, and did. Phygitals, a third platform, passed 100,000 cards and $250M in volume and tied up with Fanatics Collect in April 2026.

Our tape

We pull Collector Crypt’s public marketplace feed every morning. On September 8, 2026 it showed 142,884 live listings, up 9.4% from 130,614 on August 28 — twelve thousand more cards entering a vault in eleven days — at a median asking price of $85.98. That median is a list price, not a sale, and it is noisy: it ranged from $75 to $115 over the same eleven days and printed a 20% one-day fall on the 8th from $108. What it tells you is the shape of the inventory: this is a venue for $50–$500 modern slabs and gacha fodder, with a thin tail of expensive cards used as marketing. The trophy market is not here.

The custody model, honestly

You own a token that is a claim on a company’s promise to ship you a card. Ask what you would ask a gold vault: who is the custodian, whose name is on the insurance, what happens in a bankruptcy, is redemption a right or a policy. Courtyard’s Brink’s arrangement answers the first; nobody in the sector has been tested on the third.

The gacha products are pack-opening with a house edge and instant liquidity — across the platforms, tokenized Pokémon drew about $324M of on-chain spending in June 2026 alone, per a KuCoin market note, run on private odds — and are entertainment spend. As a venue for a specific graded card, tokenization is cheaper and faster than eBay. As an asset class, it is a leveraged bet on cards and on a chain at once.

There is no clean way to own this market without owning cards, and the attempts to build one have a poor record.

Fractional

Otis sold shares of a 1st Edition Charizard valued at $236,800; the company was bought by Public.com in March 2022 and wound down, with investors in some series taking a loss. Collectable, structured under Regulation A, has run no new offering since 2022 and is tied up in Delaware Chancery Court litigation. Rally still lists a few cards among its cars and memorabilia, at minimums of $50 and with thin secondary trading. The pattern: fractional platforms attract money at the top of a cycle, freeze when the cycle turns, and charge fees in both directions. The one durable product in the space is the vault-and-lend model Alt runs — store the card, borrow against it — and that is a financing business, not a fund.

Funds

There is no listed or registered fund holding trading cards. Several private vehicles raised money in 2021–22 and 2025–26; none publishes audited returns. Treat any “card fund” pitch as a private-equity-style commitment with a manager risk you cannot diligence.

Listed proxies

The infrastructure is partly public and the manufacturer is one-third public. Nintendo books its share of The Pokémon Company as equity-method income (¥82.7B in the year to March 2026, up from ¥35.1B) but blends it with a console launch. Hasbro carries Magic ($1.72B in 2025) inside a toy company. eBay owns TCGplayer and Goldin and has reported nine straight quarters of double-digit trading-card growth. Bandai Namco publishes One Piece. GameStop has become a PSA drop-off point and a card retailer.

The graders, which are where the margin in this market actually sits, are private: Collectors (PSA) under Nat Turner’s group with D1 Capital and Steve Cohen’s family office, CGC under Blackstone. Naming those companies is not a recommendation; it is a map of who captures the economics of a ten-billion-card year regardless of whether any particular card goes up.

A card is a laminated piece of paper, and its enemies are the same as a photograph’s: humidity, light, heat and handling. The good news is that storage is the cheapest part of this asset class. The bad news is that a slab does less than people think.

Humidity and heat

Cardboard is hygroscopic; above about 55% relative humidity it swells, warps and, over years, fosters mould, and a warped card is a PSA 7 whatever it was before. The target is 40–50% humidity and a steady room temperature; a sealed container with a hygrometer and a silica packet does most of the work. Attics and basements are where binders go to lose a grade.

Light

Holofoil and ink fade under UV. A PSA holder is not UV-proof, and a display case in a sunny room will visibly dull a holo in a few years. Store in the dark; display, if you must, behind UV-filtering acrylic and away from windows.

The slab

A holder protects against handling and makes the card fungible; it does not stop humidity or light, and a cracked holder means re-grading at your expense and risk. Sleeved, upright, in a box.

Insurance

A standard homeowner or renter policy caps collectibles at a few thousand dollars and excludes most causes of loss that matter. A scheduled collectibles rider or a specialist policy (Collectibles Insurance Services and similar underwriters) runs about 1–1.5% of scheduled value a year, covers transit, and requires an inventory with photographs and cert numbers. For a collection above $25,000 it is not optional. Keep the inventory current: the cert-number list is also your theft report and your tax basis file.

Vaults

For a serious position, a third-party vault: Fanatics Collect (ex-PWCC) offers storage, insurance, photography and consignment straight from the vault; the tokenized platforms are vaults with a token in front. A vault removes shipping risk on every trade and adds a counterparty. Read the custody terms as a brokerage agreement, because that is what they are.

Every risk so far costs you percentage points. This one costs you the position. The fraud in this market is industrial, it targets exactly the cards worth owning, and its methods have improved faster than the defences.

Counterfeit cards

Fake 1st Edition Charizards were common enough in the 2020 boom that spotting guides became a genre, and PSA’s 2025 Fraud Report says counterfeit Pokémon submissions rose 125% in 2025 and altered cards of all kinds 108.7%. A Saratoga Springs, Utah woman charged in January 2026 and arrested again on February 3 faces 25 felony counts across three cases for selling counterfeit Pokémon cards to local collectors for thousands of dollars each; the search of her home turned up printers, laminators, heat presses and 3D-printing tools.

The tells are physical: wrong card stock, wrong holofoil pattern, wrong font weight, colours a shade off, and the light test (a genuine Wizards-era card has a black core layer that stops light). Modern fakes are good enough that a raw purchase above a few hundred dollars from anyone but a known dealer is a mistake.

Altered cards

The grade cliff makes a card doctor’s trade worth a fortune. Trimming shaves a worn edge to fake a sharp one, leaving a card fractionally undersized. Pressing flattens creases. Recoloring touches up edge whitening. Rebacking marries a clean front to a different back. The graders catch most of it, which is why the scam has moved downstream to the slab.

Counterfeit slabs

A federal jury in the Southern District of New York convicted Anthony Curcio on January 28, 2026 of wire fraud and conspiracy for a scheme that, from 2022 to May 2024, sold sports and Pokémon cards in forged PSA holders and labels through online marketplaces and auction sites, using a front man to face the victims, and sought more than $2 million from them.

Community reports in 2026 describe a European ring that reseals genuine cases with worse cards and an acetone method that strips a real label’s ink while keeping its UV features; we have not verified either independently. One collector documented a $28,000 loss on fake slabs at a single card show. The secondary market for PSA 10s is, at the margin, a market in labels, and labels can be printed.

Shill bidding and manufactured comps

Because every price guide is built from sold listings, a seller who buys his own card at a high price on a second account creates a “comp” that every tracker imports. eBay’s enforcement is real but slow, and the auction houses’ bidder deposits exist partly for this reason. A sudden price step on a thin card with two or three sales behind it is a comp to be distrusted until a third party pays it.

Theft, non-delivery and dirty money

The 2025–26 catalogue runs from a Florida man who, some 75 times from July 2025, ran boxed Pokémon cards through Target self-checkouts while scanning 99-cent taco-seasoning packets in their place and resold the cards on eBay for about $39,000, to the ordinary run of pre-order non-deliveries; and because high-end cards are now bought with crypto and wires from strangers, the seller of a six-figure card carries clawback risk on the buyer’s funds. PSA’s 2025 Fraud Report put the market value of the counterfeit and altered material it intercepted in 2025 above $200M, up 45% on 2024, and its president called that conservative.

The defence stack

Raw: buy from a dealer with a reputation to lose, and grade it yourself. Slab: verify the cert and match the image to the card, check holder geometry and label print against a known-genuine example, and on eBay buy at $200 and above so PSA handles it. Sealed: chain of custody. Above $10,000: a house that inspects in hand and holds the money in escrow. None of it is expensive. All of it is slower than the scammer wants you to be.

Trading cards are collectibles under the US code, and the treatment is precise, unfriendly, and frequently ignored by people who sell on eBay.

The 28% rate

A gain on a collectible held more than a year is taxed at the holder’s ordinary income rate capped at 28% (IRC §1(h)(5)), not the 15–20% that applies to stocks; short-term gains are ordinary income. The 3.8% net investment income tax applies above the thresholds, and state tax on top. On a $100,000 gain the difference between the collectibles rate and the ordinary long-term rate is roughly $8,000–$13,000 of federal tax. It is the same rate every tangible collectible in this library carries; the Art, Watches and Cars Dossiers work the same arithmetic on their own objects.

Basis

Purchase price plus sales tax, shipping, grading fees and authentication costs. Every PSA invoice raises your basis. Storage and insurance do not; since 2018 an individual investor cannot deduct them at all. Amount realised is net of selling fees and commissions. Keep the records for as long as you hold the card plus three years.

Reporting

The 1099-K threshold for marketplaces was restored, permanently, to more than $20,000 in gross payments and more than 200 transactions in a calendar year by the One Big Beautiful Bill Act (P.L. 119-21, §70432, signed July 4, 2025), which repealed the $600 rule as if it had never been enacted; the IRS confirmed in October 2025 that the $20,000 threshold applies from calendar year 2025 onward, ending a phase-in that had put 2024 at $5,000 and would have put 2025 at $2,500. Both conditions must be met: a seller who moves one $50,000 card gets no form.

Several states set lower thresholds for their own reporting — $600 in Massachusetts, Maryland, Vermont, Virginia and the District of Columbia, $1,000 in New Jersey and Illinois — and eBay files accordingly. None of this changes what is taxable: a gain is reportable from the first dollar, form or no form, and the IRS matches marketplace data when it wants to.

Dealer or investor

Sell enough, often enough, with enough intent to profit, and the IRS can treat you as a dealer: gains become ordinary income subject to self-employment tax, but expenses (grading, storage, travel to shows, supplies) become deductible against them. A hobbyist gets neither. The line is facts and circumstances; a person flipping fifty modern boxes a year is closer to it than they think.

Losses and wash sales

A loss on a collectible held for investment is a capital loss, deductible against gains and $3,000 a year of ordinary income. A loss on a card you held for personal enjoyment is not deductible at all, and the distinction is argued after the fact. The wash-sale rule (§1091) applies to stock and securities; it does not, as written, reach a trading card, which means selling a losing card and buying an identical one is not prohibited — though the substance-over-form doctrine is available to an examiner who thinks you are being cute.

Sales tax and structure

Marketplaces collect sales tax at the buyer’s rate on every sale; it goes into basis, and vault storage in a no-sales-tax state (Delaware, Oregon, New Hampshire) is one reason the vaults sit where they do. An IRA cannot hold a collectible — the purchase is a distribution (§408(m)). An LLC changes nothing about the rate. The useful structure is the boring one: an inventory with cert numbers, a basis schedule, a periodic appraisal, and an estate plan that says who gets the Charizard, because the step-up in basis at death is the single most valuable tax feature this asset has.

Our answer for the category is the same answer we give for any no-yield, high-friction, sentiment-priced asset with a private custodian in the middle: small, and only in the part of the market with a closed supply.

The number

For a household with a real alternatives sleeve, trading cards belong inside the collectibles line, and the collectibles line inside a 5–10% alternatives allocation. That puts cards at 1–3% of investable assets at the very top, sized so that a 50% drawdown — which the flagship card delivered in 2022 and modern cards delivered in 2025 — is an annoyance and not an event. The Greer test is the honest one: an asset that pays nothing and cannot be valued except by the next buyer is a store of value at best, and store-of-value assets are position-sized for loss, not for return.

The split

Within the allocation, the only cards with a supply curve that cannot move against you are pre-2004 (Wizards-era and early e-Reader) English cards in PSA 8 and above, Japanese promotional and tournament cards with documented distributions, and the top tier of Magic’s Reserved List. That is where the entire investment case lives. Modern cards and modern sealed are a consumption budget with an option attached; fund them from the entertainment line, not the portfolio.

The consumption dividend

The thing that makes this category work for the people it works for is that they would own the cards anyway. A slab on a shelf pays a dividend in enjoyment that no index captures, and for a collector that dividend is real. For a pure investor it is zero, and without it the arithmetic in the worked example says the honest expected return on a mid-size vintage position, after fees, tax and carry, is a low single digit with fat tails in both directions. Size to that, not to the chart.

If you are starting from zero, the sequence that loses the least money is the one the good collectors already follow, and it is slower than the hobby wants you to be.

One set, one language

Pick a single Wizards-era set (Base, Jungle, Fossil, Team Rocket, the Neo sets, Skyridge) or a single Japanese promotional programme, in one language, and learn it completely: every card, every variant (1st Edition, shadowless, unlimited, error prints), every known fake, every price for six months. Breadth is how new money gets picked off.

Pop reports before prices

Before you look at what a card sells for, look at how many exist in each grade on the PSA, CGC and BGS population reports, and how fast the top grade is growing. A card whose PSA 10 count doubled in a year is not scarce whatever its chart says. A card with 40 tens and 2,000 nines has a grade cliff you can see.

Buy the grade, not the card

For a first serious purchase, buy an already-graded PSA 9 or 10 through eBay’s Authenticity Guarantee or a house that inspects in hand. Do not buy raw and gamble on the grade until you have handled a hundred cards and can call a 9 from a 10 with a loupe.

Learn the fee stack before the first bid

Compute your all-in cost (price, tax, shipping) and your all-in exit (venue fee, shipping, tax on the gain) before you bid. If the card needs 30% to break even, you know that going in.

Ignore anything printed after 2020 as an investment

Buy modern product if you enjoy opening it. The moment you find yourself computing a return on an Elite Trainer Box, re-read the section on the manufacturer.

Insure at $25,000, vault at $100,000

Below the first figure a safe and a rider suffice; above the second, a professional vault with consignment access removes more risk than it adds.

Trading cards are a real asset class by the only definition that matters — a group of things with a shared economic character, priced by a deep and observable market — and Pokémon is its largest and most liquid segment. The tape is genuine: 26 million cards a year graded, a record trading card at $16.49M, a manufacturer earning ¥144B of operating profit, nine straight quarters of double-digit growth at eBay. Nothing here is a fad in the sense of being about to vanish.

It is also the only collectible we cover whose maker prints ten billion new units a year and the only one whose scarcity is assigned, one object at a time, by two private companies paid per unit. That makes the market two markets. The vintage-and-promo market has a closed supply, a record book that keeps setting records, and an honest long-run return that is extraordinary even after you apply the Korteweg haircut, subtract a 25–40% round trip and pay 28% on the rest.

The modern market is a consumer product with a hype premium, and the premium is currently being repriced: 50–75% drawdowns on the chase cards of 2021–25 while the category index still reads “up.” That divergence is the sports-card cycle of 2021 replaying with a publisher that has promised, in writing, to keep printing.

Our position: the investable case is confined to pre-2004 English cards in PSA 8 and above, documented Japanese promotional and tournament cards, and the top of Magic’s Reserved List, sized at a low single-digit share of assets and held by people who would own them anyway. Everything else is entertainment with resale value, which is a fine thing to buy and a poor thing to call an allocation.

What we are watching, with thresholds

Six gauges, each with the level that would change our view.

PSA’s monthly Pokémon volume

GemRate publishes it. PSA graded 2.07M cards in May 2026 with TCG at 67%. Three consecutive months of year-on-year decline in TCG submissions would be the first hard evidence that the grading wave — and the modern-card demand behind it — has turned.

Pop growth on the key cards

The 1st Edition Charizard PSA 10 count was about 125 in December 2025. If it passes 150 within twelve months, the price at the top of the staircase has a supply problem that the record book does not show yet. On the modern side, a monthly PSA 10 increase above 10% of the existing population on a card whose price is rising is the mark of a grade being handed out to meet demand.

The manufacturer’s print announcements

The next “Pokémon in Figures” update lands around May 2027. A cumulative total above 95B (another 10B year) with shelves still empty confirms the demand story; a total near 92B with product on shelves says the print run finally caught the cycle, and the modern market will find out what that means.

Sealed premiums over MSRP

The premium on a current-set Elite Trainer Box over its roughly $50–$70 retail price is the cleanest live gauge of speculative demand. Above 100% at release is mania; a current set trading below MSRP six months after release, as several 2025–26 sets now do, is the unwind.

TCGplayer market prices for the chase cards

The Prismatic Evolutions Umbreon ex illustration rare at $832 (August 2026) is the modern bellwether; a print below $500 says the 2025 cohort has left. The Umbreon VMAX alternate art in our own basket, at about $120 in the 2025 analysis above, is the 2021 cohort’s bellwether; a print back above its late-2024 level would say the speculative bid has returned.

Our own tape

The TCG seat in the IA Composite is “awaiting” at a 2.1 weight while the basket accrues the history our cadence rules require; when it joins, the PSA 10 median and the ungraded median will be published daily with the basket stated, and the day the ratio between them moves from 58 times toward 30 is the day we will say, in print, that the grade cliff is eroding.