Investing in Magic: The Gathering

On a night in April 2024 a Magic card changed hands for $3,000,000. It was a Black Lotus from the game’s first print run, the 1993 set collectors call Alpha, and it carried a grade no other Alpha Lotus had: CGC Pristine 10, the only one in that grader’s census. The seller was Adam Cai of Pristine Collectibles, the buyer was not named, and there is no auction record, no bid count and no hammer. CGC announced it, ICv2 and cllct reported it, and a good part of the hobby refused to believe it. That is the Magic market at the top: a handful of objects from a print run of about 1,100, priced by two or three people at a time, in transactions you will never see.

Eight months earlier, in August 2023 and in a very different room, Hasbro’s chief executive Chris Cocks had told analysts that Magic’s 2023 licensed set, The Lord of the Rings: Tales of Middle-earth, was on its way to $200 million in revenue within seven months. In July 2025 he reported that the next big crossover, Final Fantasy, had taken $200 million on its launch day, June 13, 2025, after Wizards raised the print run four times and still, in his words, left the market wanting more. The company then shipped Spider-Man in September 2025, Avatar in November, Teenage Mutant Ninja Turtles and Marvel Super Heroes in 2026, and booked the largest quarter in Magic’s history, $545 million, in the three months to June 2026.

Those are the two Magics. One is a fixed and shrinking pool of 1993–1999 cardboard, protected by a promise the publisher made in 1996 and has kept since. The other is the fastest-growing product line at a public company, printing as much as the market will take. The second pays for the first, competes with it for the same buyers, and, as the 2024–25 record shows, can push the middle of the vintage market down while the top makes headlines. This guide is about telling the two apart and pricing each honestly.

The print run of one summer in 1993 decides which Magic cards are assets and which are cardboard, so the supply comes first.

Richard Garfield designed Magic for Wizards of the Coast, a small Seattle-area publisher, and the game went on sale in August 1993. It was the first trading card game, a game whose pieces were sold in random packs, so the game and the collecting market were born together. Wizards became part of Hasbro on September 30, 1999, for about $325 million, per Hasbro’s 10-Q for that quarter. The games that followed, Pokémon in 1996 and the rest, copied the model, and the sister guide on this hub, Investing in Pokémon and Trading Card Games, covers them; this guide stays inside Magic.

Alpha, Beta, Unlimited

The first set was printed three times in a year, and the three printings are three different assets. Alpha, released in August 1993, had 295 cards and a print run of about 2.6 million cards. Its corners were cut on a rounder die than anything printed since, which is how you tell an Alpha card at a glance. Beta, in October 1993, added the seven cards Alpha had left out, corrected the die, and ran about 7.3 million cards.

Unlimited, in December 1993, was the same card list with a white border instead of black, and the run was 35 million by the count in The Duelist’s 1995 card list and the 1996 Official Encyclopedia; some later compilations say 40 million, and we show both because nobody outside Wizards has the ledger. Revised, in April 1994, was larger again and dropped the cards that mattered most.

Alpha’s rarity sheets put about 1,100 copies of each rare into the world, 4,500 of each uncommon and 15,900 of each common. So there were about 1,100 Alpha Black Lotuses, about 3,300 Beta copies and something over 18,000 Unlimited ones, before thirty years of play, loss and children. Nobody knows how many survive; the grading census, covered later, counts a little over a hundred Alpha Lotuses that PSA has ever seen.

The Power Nine

Nine cards from those first printings are so much stronger than anything printed since that the game’s oldest tournament format restricts each to one copy per deck, and the hobby calls them the Power Nine: Black Lotus, the five Moxen (Pearl, Sapphire, Jet, Ruby, Emerald), Ancestral Recall, Time Walk and Timetwister. They are the reference assets of this market as the T206 Wagner and the 1952 Topps Mantle are for baseball, and all are on the Reserved List. Below them sit the ten original dual lands (Underground Sea, Volcanic Island, Tundra and the rest), printed through Revised and still the most-played cards on the list, and then a long tail of 1994–1999 rares whose prices depend on whether anyone still wants to play them.

One rule makes Magic different from every other card game: a published promise never to reprint 572 cards. Wizards has tested it twice, and an investor should treat it as a constraint on supply and nothing more.

In 1995 Wizards published Chronicles, 116 cards that were all reprints of early rares, and the collector market for the originals fell apart. Its answer, published on March 4, 1996, was a list of cards it promised never to print again. The list was closed at the end of the decade and has stood since at 572 cards, all from sets released between 1993 and 1999, with nothing added and nothing removed. No other publisher of a living game has done this. The Pokémon Company reprints Charizard every few years; Wizards cannot, by its own word, print another Black Lotus that is legal to play.

The two tests

The promise has been stressed twice. In the 2000s Wizards read the policy as allowing premium foil versions of reserved cards, on the argument that a foil was a different object; collectors disagreed loudly enough that in 2010 the company closed what the hobby calls the foil loophole and tightened the policy so that no future reprint of a reserved card would be tournament-legal in any form.

The second test came on November 28, 2022, when Wizards sold the 30th Anniversary Edition: four fifteen-card packs of reprinted Beta cards, Black Lotus and the Power Nine included, printed with a different back so they could not be played in any sanctioned event, for $999. The sale was declared over after about forty minutes. It angered both camps at once, the collectors who saw a Lotus reprinted and the players who saw a thousand-dollar box of proxies, and Wizards has not repeated it. The list itself was not touched, and Mark Rosewater, Magic’s head designer, has said publicly, in remarks reported by ICv2, that he spent years trying to abolish it and does not think it is going away.

What the promise does and does not do

The Reserved List fixes the supply of tournament-legal copies. It does nothing about demand, and demand is what moves prices: the 2022–23 drawdown and the 2024–25 slump in mid-grade vintage, documented in the next section, happened with the list intact. Every new set that prints a card doing the same job as a reserved card erodes the play demand under it, and the Commander format, the casual multiplayer game that drives most of Magic’s singles volume, has been the main source of demand for the dual lands and the cheaper reserved rares for a decade.

The list is why Alpha and Beta power behaves like a fixed-supply collectible, and why the bottom three-quarters of the list behaves like thinly traded stock, where a buyout by a handful of dealers can double a card in a month and the absence of a reprint cannot stop it falling back.

What Magic has actually returned can be read from three price series and no index, and the compounding story belongs to a few dozen cards, not to the category.

Magic has no repeat-sales index that we would put beside Card Ladder’s for sports cards. TCGplayer publishes a market price for every card, built from its own sales, and the trackers republish it, but nobody publishes a dated, methodology-stated index of vintage Magic, and we could not verify one for this build any more than the Pokémon guide could. What exists is a ladder of public sales for the Black Lotus, ICv2’s readings of the auction record, and TCGplayer’s price history for the played tiers. We use all three and label each.

The Black Lotus ladder

The Alpha Black Lotus in top grade is the market’s reference series. A BGS 9.5 sold through PWCC on eBay for $87,672 in July 2018 and another BGS 9.5 for $166,100 in February 2019. A PSA 10 signed by the artist, Christopher Rush, made $511,100 on eBay through PWCC in January 2021, and a second signed PSA 10, from the same collector, David Edwards, made $540,000 at PWCC on March 16, 2023.

Eight days later, on March 24, 2023, a different object sold at Heritage for $615,000, the auction record for any Magic card: a Beta artist-proof Black Lotus, one of the blank-backed copies given to the illustrator, signed by Rush, graded CGC 8.5 and consigned from the collection of Rush’s former agent Jeff Ferreira. It sits beside the ladder as a marker of what the market will pay, not as a rung on it, because it is neither an Alpha card nor a production card. Then the CGC Pristine 10 Alpha at $3,000,000, privately, in April 2024.

From the 2018 BGS 9.5 to the March 2023 PSA 10 is a 6.2-fold rise in under five years, about 48% a year. It is also two grades, two graders, and a period in which every collectible on earth went up.

The honest comparison is the one inside a single grade. A signed PSA 10 sold for $511,100 in January 2021 and a second signed PSA 10, a different card from the same collection, for $540,000 in March 2023: 5.7% over twenty-six months, about 2.6% a year, across a period in which the S&P 500 fell 18.1% in 2022 and recovered 26.3% in 2023. The top of the Magic market did not crash in 2022 and did not compound either; it sat. The $3 million sale a year later is a different grade, a different grader and a private transaction, and we would not draw a line from $540,000 to $3,000,000 and call it a return.

The middle of the market

Below the trophies the record since 2022 is negative, and it is the most important fact in this guide. ICv2, the trade paper for the game and hobby channel, tracked Heritage results through 2025 and found collectors of pre-2000 singles moving to chase cards from new sets regardless of print run. Its marker was the Unlimited Black Lotus, the most-traded real Lotus: a CGC 5 sold at Heritage on July 10, 2025 for $10,625, at the low end of the $10,000–12,000 range ICv2 saw through 2025 and well below the $15,000–20,000 the same card brought in 2022–23. That is a drawdown of a third to a half in the one Lotus an ordinary collector can buy, over three years in which Hasbro’s Magic revenue rose by roughly two-thirds, from $1.07 billion in 2022 to $1.72 billion in 2025.

After Heritage’s December 12–13, 2025 sale ICv2 described high-end hammer prices as rebalancing rather than falling: a CGC 8 Alpha Lotus made $68,750, an unsealed box of Beta boosters $237,500 and a sealed Arabian Nights box $218,750.

The played tier tells a third story. Underground Sea, the most-wanted of the original dual lands, was a $360 card in Near Mint Revised condition in August 2020, by Quiet Speculation’s survey; it reached about $680 in 2022, listed at $799.95 in August 2025, and stood at a market price of about $1,045 on MTGStocks in September 2026 (low $900, average listing about $1,200), per the trackers that republish TCGplayer’s data. That is roughly 19% a year over six years, in a card with a fixed supply and a format, Commander, that keeps needing it. It is the best return in this guide and it belongs to a $1,000 card, not a $3 million one.

What the record leaves out

Three things, all cutting the same way. Every figure above is a price with no costs against it; a round trip takes 13% to 30% off, as the fee section shows. Every auction figure is a card that found a buyer; the lots that passed are not in ICv2’s summary. And the ladder is drawn through the cards that went up; nobody publishes the history of the 1997 reserved rare that was $40 in 2021 and is $18 in 2026. Set beside the S&P 500’s total return of 25.0% in 2024 and 17.9% in 2025, vintage Magic as a category lost money in 2024–25 while its best-known object set a record. Both are true, and the second one is the marketing.

In a market whose record was set in private, a price without a venue and a date is a rumour, so here is the canon of Magic’s largest sales with both attached.

The Lotus

Black Lotuses fill the top of that chart, with the one exception covered next, and what the chart adds to the ladder is the venue. Two of the three largest Magic sales ever were private. The largest public sale, the signed Beta artist proof at Heritage in March 2023, is a fifth of the private record, and the largest public sale of a production Alpha Lotus is the $540,000 PSA 10 of the same month.

The next public test is on the calendar. A PSA 10 Alpha Black Lotus, reholdered in 2026 and one of three PSA 10s known per Star City Games and ICv2, headlined Goldin’s 2026 Summer Goldin 100 auction; it stood at $535,000 on thirteen bids with twenty-six days to run in late August 2026, and regular bidding closes on September 19, 2026, ten days after this guide’s as-of date. Whether it clears the $540,000 of March 2023 is the first entry in What to watch.

The One Ring

The second-largest sale is not a 1993 card at all. The Lord of the Rings: Tales of Middle-earth, released in June 2023, included one serialised copy of The One Ring numbered 001/001; a Canadian collector, Brook Trafton, opened it, and in August 2023 the musician Post Malone bought it from him in a private deal reported by NPR, Forbes and Star City Games at $2 million, a figure from the parties rather than a venue. It was the first time a card printed in the same year became a top-tier trophy, and it was a product decision: a publisher that can manufacture a one-of-one can manufacture a record. Wizards repeated it with 77 serialised golden Traveling Chocobo cards in Final Fantasy in 2025.

Sealed product and the artist proofs

Below the singles sit the boxes. An unsealed box of Beta boosters, with its original packs, made $237,500 at Heritage on December 12–13, 2025, and a sealed Arabian Nights box, from the first expansion of late 1993, $218,750 in the same sale, a little over $6,000 above a 2024 Heritage result.

The artist proofs are the smallest category: CGC’s census counts four Beta Black Lotus proofs, the Rush-signed one made the $615,000 auction record in March 2023, and an unsigned CGC 8.5 sold for $106,250 at Heritage on October 21, 2023. These are the categories in which the record was still rising through 2025, and they are the categories in which counting the supply is hardest: nobody knows how many sealed Beta boxes exist, and an artist proof’s provenance is a signature. Investing in Sealed Product, on this hub, covers the boxes across every game.

Magic is the only collectible on this hub whose supplier is a listed company that reports the line, and Hasbro’s 2023–2026 figures explain why vintage prices fell while the game’s revenue rose.

Hasbro reports Magic inside its Wizards of the Coast and Digital Gaming segment and has reported the franchise’s growth every quarter since it became, in 2022, the company’s first billion-dollar brand. Magic did $1.065 billion in 2022 and $1.085 billion in 2023, when it was about 52% of Hasbro’s gaming revenue. In 2025 it reached $1.72B, up 59%, which puts 2024 at roughly $1.08 billion by arithmetic, and Hasbro’s fourth quarter of 2025 alone was up 141% on the year before, to about $502 million by the same arithmetic, on Avatar: The Last Airbender and the continuing Final Fantasy demand.

The first quarter of 2026 brought Magic revenue of $469.6M, up 36%, on Lorwyn Eclipsed and the Teenage Mutant Ninja Turtles set; the second quarter, reported in July 2026, was $545.3M, up 32%, on Marvel Super Heroes. Hasbro’s release called it the first quarter above half a billion dollars; on the company’s own quarterly growth rates the fourth quarter of 2025 was already at about that level, so we call it the record quarter and leave the threshold to Hasbro. The Wizards segment did $663.8 million in that quarter, up 27%, and Hasbro raised its 2026 revenue guidance to 5–7% growth in constant currency, from 3–5%, on the strength of the card game.

Where the growth came from

Hasbro’s own explanation has three parts: Universes Beyond, backlist and Secret Lair. Universes Beyond is the licensed-crossover line, cards that use other companies’ characters under the Magic rules; it became the main event with The Lord of the Rings in June 2023, which Hasbro said on its August 2023 call was already the second-best-selling set in Magic’s history and tracking to $200M within seven months, a figure Modern Horizons 2 had taken two years to reach. Final Fantasy, released June 13, 2025, did $200 million on its first day, by Cocks’s account on the July 2025 call, after four increases to the print run, and was called the best-selling set in the game’s history on that call and again in Hasbro’s FY2025 reporting.

Marvel’s Spider-Man followed on September 26, 2025, Avatar on November 21, 2025, Teenage Mutant Ninja Turtles early in 2026 and Marvel Super Heroes on June 26, 2026; Star Trek is scheduled for November 13, 2026. Backlist is the catalogue of older sets that keep selling; Secret Lair is the direct-to-consumer store where Wizards sells small themed drops without a retailer in between.

The cadence

The cost of that growth is the calendar. Wizards released seven main sets in 2026, ten products counting supplemental releases, four of the seven licensed crossovers; one of the game’s senior designers called the seven-set year an outlier. At MagicCon Amsterdam in July 2026 the company announced 2027 as six sets, three from Magic’s own settings and three Universes Beyond, and framed 2026 as the transition year; the first, Nauctis: The Sunken Realm, is dated February 5, 2027. Whether six is a slowdown or a plateau against the roughly four-set years of the 2010s is the second entry in What to watch.

Why this matters to a vintage buyer

A collectible’s price is set by the marginal buyer, and Magic’s marginal buyer has, since 2023, been offered a new chase every quarter with a licence that buyer already loves. ICv2’s 2025 reading of the Heritage results, collectors valuing rare vintage less and new chase cards more regardless of print run, is the demand side of Hasbro’s revenue chart. The Reserved List protects the supply of a 1994 card; it cannot protect its share of a collector’s budget when Final Fantasy prints 77 serialised Chocobos and Marvel prints Spider-Man. The parallel is the sports-card hobby’s 1980s, when Topps, Fleer and Donruss printed as much as the market would take. The difference is that Wizards owns the only vintage tier and has an incentive to keep it credible, which is why the 30th Anniversary Edition was a one-off and not a policy.

Collector Boosters and the modern chase

The product where the modern chase lives is the Collector Booster, a premium pack introduced in 2019 that guarantees foil and alternate-art rares. Wizards resumed printing a suggested retail price with Foundations in late 2024, so the numbers are public: a standard Collector Booster listed at $24.99, raised to $26.99 in 2026, about $324 for a twelve-pack box. Final Fantasy’s listed at $37.99, about $455 a box, and because the set seeded 77 serialised golden Traveling Chocobo cards across its collector packs, sealed boxes traded at roughly $1,000–1,200 in the months after release, per Draftsim’s and MTG Rocks’s surveys, and MTG Rocks put them at about $1,900 by July 2026, a year on.

A Foundations Collector box, with no such lottery ticket, had a latest sale of $335 on StockX in 2026 against a twelve-month range of $382–429: at or below its own list price.

That pair of numbers is the whole modern-product thesis. A box’s premium over MSRP depends on whether the set contains a lottery ticket the market cannot get elsewhere, and Wizards decides that at the printing stage. The One Ring, the Chocobos and the serialised Marvel cards are manufactured scarcity inside an unlimited print run, and the Pokémon guide’s lesson applies without change: a publisher that can produce a new chase every quarter is a publisher whose last chase will not stay chased.

Sealed Collector boxes of the licensed sets have been a good trade in the first weeks after release and, Final Fantasy’s serialised exception aside, a poor hold afterward, because Hasbro reprints what sells; Cocks said in July 2025 that Final Fantasy production was continuing rather than ending with the launch window. Keep modern Magic on the consumption side of the ledger, and never pay more for a modern chase card than its Collector Booster box costs at MSRP. Nothing on the Reserved List can be reprinted. Everything in a Collector Booster can.

In vintage Magic the holder is most of the price, so who grades the card, what the grade measures on 1993 stock, what the census says about the Power Nine and what the fee is all come before the card itself.

Three companies matter. PSA is the volume leader and the label that sells for the most in a slab: GemRate’s 2025 review counted 19.26 million cards graded by PSA, of a 26.8 million industry total, with trading-card games and non-sport at 16.8 million against 10.0 million for sports.

CGC Cards, majority-held by Blackstone’s Tactical Opportunities arm, graded 4.92 million, up 121%, and is the grader most vintage Magic specialists prefer for its strict centering and surface standards and its Pristine 10 label above the ordinary 10; the $3 million Lotus is a CGC card. Beckett, whose BGS 9.5 was the vintage Magic standard through the 2010s and holds the 2018–19 Lotus records, graded about 824,000 cards in 2025.

What the grade measures on an Alpha card

A grade on the 1-to-10 scale is a judgment on centering, corners, edges and surface, and on a 1993 card two of the four work differently. Alpha’s rounded corners are a property of the die, not wear, and every grader treats them as such; a sharp Alpha corner is a rounded corner. Centering tolerances are looser for pre-1990 printing on every grader’s published scale, because the 1993 sheets were cut with the precision of a small company in a hurry. Surface is where Alpha and Beta cards fail: print lines, ink dimples and the scuffing that a black border shows and a white one hides. PSA’s published policy is that it will not holder a card that has been trimmed, recoloured or restored, and the fraud section explains why that clause carries most of the weight in this market.

The census on the Power Nine

The population report is the grader’s running count of every copy it has holdered, by grade, and for a 1993 card it is close to a census of the survivors that were worth the fee. On the Alpha Black Lotus, the trade reading in August 2026, when Star City Games and ICv2 covered the Goldin lot, was three PSA 10s known; an older reading of PSA’s report on the Magic Librarities collector forum put it at six, out of about 105 copies graded. We could not reconcile the two from PSA’s own report this session, the snapshot dates were not captured, and the count moves slowly. CGC’s Pristine 10 was unique as of April 2024; the company has since announced further examples, the latest of which it described as the third.

Against a print run of about 1,100, a hundred-odd graded Alpha Lotuses means most surviving copies are still raw, played and unaccounted for, the opposite of the 1952 Topps Mantle, where the graded census is close to the whole surviving supply.

On the Alpha Power Nine the top grade is therefore not merely rare but almost untraded. The blue cards, Ancestral Recall, Time Walk and Mox Sapphire, change hands in Alpha only a handful of times a year worldwide by the count of the price guide tcgTalk, about eight for Ancestral Recall and six for Time Walk; those are a vendor’s estimates, and they match what the auction record shows. A PSA 10 Alpha Ancestral Recall has no reliable price because there is no series of sales to build one from.

What it costs

PSA’s fee for a card declared up to $1,500 was $79.99 at its Regular tier after February 10, 2026, when it raised every service level by $5, with an estimated turnaround of 70–80 business days; the same change folded the old TCG Bulk service into a $24.99 Value Bulk tier for Collectors Club members, and on June 2, 2026 PSA paused all its Value tiers, Value Bulk included, while it worked through a backlog it put at about ten million cards, so $79.99 is the practical entry point for a single card.

High-value cards pay tiers priced on declared value, so the fee on a $20,000 Alpha card is in the hundreds of dollars plus insured shipping both ways; the Pokémon guide’s grading section covers the tiers; CGC’s cheapest tiers are $17 Bulk and $20 Economy from March 24, 2026; and Card Grading Explained, on this hub, covers what the digit on the label does and does not measure. On a card that trades at $20,000 in a PSA 8 and $45,000 in a 9 the fee is noise and the grade is everything, which is why every vintage Magic card worth grading has been, and why a raw Alpha Lotus offered to you on a marketplace is a warning before it is an opportunity.

One edition and one grade point can move the price of the same Black Lotus by a multiple, and that spread is where every dollar in this market is made or lost.

The three 1993 printings of Black Lotus are the same card in the game and three different assets in the market. The price guides that republish TCGplayer and eBay sales put an Unlimited Lotus in PSA 8 at $7,000–12,000 in 2026 and a PSA 10 at $18,000–25,000; a heavily played raw Unlimited copy sold on eBay for $8,400 in March 2024, and the CGC 5 at Heritage in July 2025 made $10,625. A Beta Lotus in PSA 8 runs $24,000–45,000.

The Alpha runs from a CGC 8 at $68,750 in December 2025 to a PSA 10 bid to $535,000 at Goldin in late August 2026, with the sale open to September 19, and the CGC Pristine 10 at $3,000,000. Alpha to Unlimited, grade for grade, is roughly seven to one; Alpha CGC 8 to Alpha CGC 10 is more than forty to one.

Why the cliff sits where it does

In modern cards the cliff sits between a 9 and a 10, because tens exist in the hundreds and nines in the thousands. On an Alpha card it sits wherever the census runs out, which for a Black Lotus is between the 8 and the 9: enough 7s and 8s trade to build a price, and so few 9s and 10s exist that each sale is its own market. The CGC 8 at $68,750 in December 2025 is a price. The CGC Pristine 10 at $3 million is a transaction. Between them, the signed PSA 10s at $511,100 and $540,000 are the only top grade with two public sales in the same holder, which is why we use them as the reference, and why the September 2026 Goldin result on a third PSA 10 matters more than its size suggests.

The dual lands and the played tier

Below the Power Nine the cliff flattens, because the buyer is a player. A Revised Underground Sea is bought to be shuffled, and its price runs on the dealer’s four condition tiers (Near Mint, Lightly, Moderately and Heavily Played) rather than on a grade; the spread from Heavily Played to Near Mint on a $1,000 card is a few hundred dollars, not a multiple. Alpha and Beta dual lands are graded and collected and behave like the Power Nine in miniature. The rule is the one the vintage baseball guide reaches from the other side: buy the highest grade that still trades at least twice a year in public, which for an Alpha Lotus is a 7 or an 8, and treat the grade above it as art, priced by whoever is in the room.

A Magic price comes from a marketplace, not an auction room, and the person selling to you usually knows more than you do; the section ends on what our own tape said about listing supply in the week to September 8, 2026.

Magic has a price on every card every day, which is unusual for a collectible and the reason the market feels more like a stock exchange than an art fair. TCGplayer, the marketplace eBay bought on October 31, 2022 for up to $295 million, publishes a market price for every printing of every card from its own completed sales, and every tracker, dealer and buylist keys off it.

Card Kingdom, the largest single dealer, publishes its retail price and its buylist, the price at which it will buy the card from you, with a 30% bonus if you take store credit instead of cash. The spread between TCGplayer’s market price and Card Kingdom’s cash buylist is the hobby’s bid-ask; on played Reserved List cards the hobby’s rule of thumb puts it at 30–40% of retail, a rule of thumb and not a published figure, and that spread, not the fee, is the true cost of liquidity for a card under a few thousand dollars.

The participants

Four groups are on the other side of your trade. The roughly 9,600 stores in Wizards’ Play Network by the third quarter of 2025, per the programme’s figures as compiled by Wikipedia, buy singles from players below the buylist and sell them at retail; a local shop buying at about 70% of market is the Pokémon guide’s figure for the trade. The large dealers, Card Kingdom, Star City Games and their peers, arbitrage between the shops, TCGplayer and their own buylists and hold the deepest vintage inventory.

The auction houses, Heritage and Goldin, handle the Alpha and Beta power and the sealed boxes. And the players, by far the largest group, are the only one that buys a card for a reason other than resale, which is why the play formats, Commander above all, are the demand data to read. The dealer selling you an Alpha Mox knows its buylist, its census and its last five sales; you are the least-informed party at the table unless you have read the same three documents.

The cycle, dated

Vintage Magic rose through the 2010s on the Reserved List thesis, then sharply from 2020 into early 2022 with every other collectible: the Lotus ladder’s jump from $166,100 in February 2019 to $511,100 in January 2021 is the boom in one card. The 2022–23 drawdown that cut Card Ladder’s sports index 23% and 9% in consecutive years hit the middle of the Magic market too, but the top held, and the March 2023 Lotus sales were records.

Then the two Magics diverged: from 2023 through 2025 the licensed sets pulled the marginal collector into new chase cards, vintage mid-grades drifted down through ICv2’s July 2025 reading, and the December 2025 results read as a rebalancing at the top. As of September 2026 the market stands with a private record two years old, an auction record three and a half years old, and a PSA 10 on the block.

Our tape

Invest Alternative keeps its own data on the tokenised card market, where a physical card sits in a vault and a token representing it trades on-chain, the one venue with a public, machine-readable order book. Our series from Collector Crypt, read on September 8, 2026, counted 142,884 cards listed, up from 130,614 when we began storing the series on August 28, 2026, a 9.4% rise in listed supply in eleven days, while the median asking price fell from $90.00 to $85.98.

That venue is mostly Pokémon, and we have no Magic-only series in our store; we present the reading as a proxy for retail listing supply and nothing more. The trading-card-game sales our eBay notable-sales tape logged in the first week of September 2026, five of them, all above $12,000, were likewise all Pokémon, the largest a 2010 Rayquaza promo at $43,000 on September 2.

Our IA Composite, the provisional cross-asset index we keep across every market we cover, stood at 100.271 on September 8, 2026, up 5.74% over thirty days and 0.29% over a year. None of that is a Magic price; all of it says that in the week this guide was written, card supply on the retail venues was rising faster than prices.

On anything held less than five years the venue decides more of your return than the card does, so every way in and out is priced here on the published 2026 fee schedules.

eBay

eBay’s fee on trading cards is 13.25% of the buyer’s total (item, shipping and sales tax) up to $7,500 per item, then 2.35% above, plus $0.40 an order. Cards sold for $200 and above pass through the Authenticity Guarantee, in which PSA inspects the card or slab before it reaches the buyer, a threshold cut from $250 in July 2026.

On a $20,000 Alpha card the fee is $993.75 on the first $7,500 plus $293.75 on the rest, about 6.4% all-in, and the Authenticity Guarantee is the only free authentication in the hobby. eBay is where the 2018–2021 Lotus records were set, through PWCC, and it remains the deepest venue for a single card between $500 and $50,000.

TCGplayer

TCGplayer, also eBay’s, charges standard sellers a 10.75% commission plus a 2.5% and $0.30 payment fee, after a February 10, 2026 change that raised the commission from 10.25%; the commission is capped at $75 an item, Pro sellers pay 9.25%, and Direct sellers, who ship through TCGplayer’s warehouse, pay 8.95% plus a per-item Direct fee introduced in June 2026. For a $100 card that sums to 13.25% plus $0.30, or $13.55, the same as eBay; for a $20,000 card the $75 cap makes TCGplayer the cheapest venue in the hobby at under 3%, though its buyer pool at that price is thin and there is no Authenticity Guarantee.

The dealers

Card Kingdom and its peers pay their buylist price by cheque, PayPal or store credit, with Card Kingdom’s 30% credit bonus, and grade the cards on arrival to their own four-tier standard, so a card you sent as Near Mint can be paid as Lightly Played. There is no fee; the cost is the spread, roughly 30–40% below retail in cash on played vintage and less on the Power Nine, where dealers compete. The buylist is the fastest exit in the hobby, a cheque within days, and the most expensive.

The auction houses

Heritage charges the buyer a 22% premium with a $29 minimum for auctions closing after January 1, 2026, up from 20%, and the seller a commission negotiable to zero on high-value consignments and around 10% on ordinary ones. Goldin charges 22% with a $19 minimum. Neither is where you sell a $2,000 dual land; both are where the Alpha power sells, and because the premium is paid by the buyer, the hammer price in a headline is 82% of what the buyer paid and more than the seller received.

The calendar

Heritage runs trading-card-game signature auctions several times a year, and its December sales set most of the 2025 vintage results; Goldin’s Summer and Winter 100 auctions carry the trophies, and the Summer 2026 edition, with regular bidding closing September 19, 2026, carries the PSA 10 Lotus. The rest of the year the market is continuous: TCGplayer and eBay trade every hour, and a dealer’s bid on a Tuesday in March is not much different from December’s, which is the one real advantage Magic has over every other collectible on this hub.

The Reserved List made 1993 cardboard valuable enough to forge, and the forgeries have kept pace with the prices; these are the ways a Magic buyer loses everything, with the cases the hobby has recorded.

Counterfeits

Wizards has warned about counterfeit cards since at least April 26, 2004, when it published a buyer-beware notice on its own site about boxes of fakes from an operation in Southeast Asia, and the problem has grown with the prices. Modern counterfeits are good enough to fool a photograph and bad enough to fail a jeweller’s loupe: the rosette pattern of the ink dots under magnification, the layered core that makes a real card opaque under a torch, and the way the card bends all differ.

The playing side of the market has blurred the line by calling counterfeits proxies, cards bought knowingly as stand-ins for casual play, and the effect is a steady supply of fakes whose sellers can claim they never said the card was real. Red Points, a brand-protection vendor, reports a single fake Black Lotus selling for more than $25,000; we could not verify that case and cite it as the vendor’s claim.

Rebacks

The fraud specific to Magic is the reback. In 1993 Wizards also printed Collectors’ Edition and International Edition, boxed sets of the Beta card list with square corners and gold-bordered backs, not legal for play and worth a fraction of the real thing. A rebacker separates the front of a Collectors’ Edition Black Lotus from its back, glues it to the back of a common Beta card, and offers a Beta Black Lotus. The front is genuine, the back is genuine, and the card is worthless.

The dealer Vintage Magic published a buyer’s warning on December 18, 2018 after rebacked cards were found inside BGS and PSA holders, and at the Grand Prix in Las Vegas in 2018 a major vendor was found offering a rebacked Beta Black Lotus in played condition, per the same account. Rebacks fail a weight test nearly every time and a light test always; none of that helps a buyer looking at a photograph of a slab.

The private-sale problem

The third risk is not forgery but opacity. The two largest Magic sales in history were private, and when the $3 million Lotus was announced in April 2024 Kotaku’s headline was that fans did not believe it. A private sale reported by the seller is a claim, and this market’s record price rests on one; that is not an accusation but a statement about evidence, the same standard the art market applies to a dealer’s reported price. The comparable you are being shown may not be a comparable at all, and the only prices you should underwrite against are the ones with a venue, a date and a buyer’s premium attached.

The carry on a Magic card is small, the tail risk is total, and the rate on the gain is not the rate on a stock; the costs and the US federal treatment follow, as precisely as the Code allows.

Carry and custody

Scheduled collectibles insurance, a rider or standalone policy that names the card and its appraised value, runs roughly $0.50–$1.00 per $100 of value a year, so a $20,000 card costs $100–200 a year to insure; a homeowner’s policy will not cover it at anything like its value. Storage is a UV-blocking case in a dry safe and costs nothing after the safe.

The alternative is a vault: eBay’s, operated by PSA, charges 0.37% a year on cards above $50 and 1% to withdraw, capped at $50, and a card bought and sold inside the same vault never ships, never risks a cracked holder and, under the structures the vaults have adopted in Oregon and Delaware, is not charged sales tax on the way in. The flagship sports-cards guide covers the vault trade-off; the short version is that you own a claim on a card held by a company, and the largest balance sheet is the safest custodian.

For an Alpha card there is one carry cost the other hubs do not have: its black borders and softer 1993 stock show every scuff, and a dropped holder or a card taken out to be played moves down the scale in a way a modern card does not. Ten years of carry on a $20,000 card is about $1,500 in insurance; one crack of the slab can be $20,000.

The 28% rate

A Magic card is tangible personal property, and the prevailing practitioner position is that it is a collectible within the meaning of Internal Revenue Code §408(m), which names art, rugs, antiques, metals, gems, stamps, coins and alcoholic beverages and lets the Treasury add other tangible property; trading cards are not expressly listed and neither the IRS nor a court has squarely ruled on them, and every adviser we have read treats them as collectibles. Under §1(h)(4) the maximum federal rate on a long-term gain on a collectible, held more than a year, is 28%, against 20% for stock. It is a ceiling: a taxpayer in the 24% bracket pays 24%.

The 3.8% net investment income tax of §1411 stacks on top above $200,000 of income for a single filer and $250,000 for a couple, for a federal maximum of 31.8%, and state income tax sits on top of that. A short-term gain is ordinary income. Grading fees and insurance on a card held for investment are best treated as additions to basis and carry respectively; keep the invoices.

Dealer, investor, hobbyist

A dealer holds cards as inventory and pays ordinary income and self-employment tax on every sale, with full deduction of expenses. An investor holds for appreciation and pays the 28% ceiling on gains, with capital losses that offset gains normally. A hobbyist owes tax on gains and deducts nothing, and the IRS decides which you were by how you behaved; the worst outcome is to be a hobbyist on the losses and an investor on the gains.

Two mechanical points. The One Big Beautiful Bill Act, signed July 4, 2025, restored the Form 1099-K threshold to more than $20,000 in gross payments and more than 200 transactions, both conditions, for tax year 2025 onward, so eBay and TCGplayer report most individual sellers to nobody; the tax is owed regardless. And the wash-sale rule of §1091 applies to stock and securities, not cards, so a card sold at a loss to an unrelated buyer at a market price and repurchased later is a real loss.

Do not put a card in an IRA: §408(m) treats the purchase as a distribution of its value. Sales tax is collected by the marketplace at the buyer’s rate in nearly every state since the Supreme Court’s Wayfair decision of 2018, and the vaults are the practical lawful way for an investor to avoid it.

The price ladder never does this arithmetic: buy an Alpha card, hold it five years, sell it, and count every fee, carry and tax, so that you know what a 40% rise is worth and what rise you need just to get your money back.

Buying

You buy an Alpha Power Nine card, say a Mox in a PSA 8 holder, on eBay through the Authenticity Guarantee at a listed price of $20,000. Sales tax at 7%, collected by eBay: $1,400. Insured shipping: $50. Cost basis: $21,450. You have paid 7.25% over the price before the card is in your hand; the only way to avoid it was a vault in Oregon.

Holding

Scheduled insurance at $0.75 per $100: $150 a year, $750 over five years. Storage in a safe: nothing. Regrade risk: none if the slab is never opened. Income: none. Carry over the period, $750.

Selling, three ways

Suppose the card rises 40% to $28,000, generous against the 2021–2023 PSA 10 Lotus record of 5.7% and the 2022–2025 Unlimited drawdown.

On eBay the fee is 13.25% of the first $7,500 ($993.75), 2.35% of the remaining $20,500 ($481.75) and $0.40: $1,475.90, before the buyer’s sales tax and shipping, on which eBay also charges the fee. Net proceeds $26,524.10. Taxable gain $26,524.10 less $21,450, or $5,074.10; federal tax at 28% is $1,420.75, plus $192.82 of net investment income tax above the threshold, $1,613.57. After tax and the $750 of carry you keep $24,160.53 against $21,450 laid out: a profit of $2,710.53, 12.6% over five years, 2.4% a year.

Through Heritage, with the buyer paying $28,000 all-in, the hammer is $22,951 after the 22% premium and you receive $20,656 after a 10% seller’s commission: a loss of $794 before carry, no tax to pay, and the premium in the house’s hands rather than yours. To Card Kingdom’s cash buylist at about 65% of a $28,000 retail price, $18,200 and a loss of $3,250; in store credit with the 30% bonus, $23,660 to spend on cards. If the card is flat at $20,000, the eBay exit returns $18,712 and the round trip costs you $3,488, 16% of your outlay, for five years of owning an Alpha Mox.

Break-even

Net eBay proceeds are 0.9765 times the sale price less $817.90. To recover the $21,450 basis and $750 of carry before any tax you need a sale at $23,572, a rise of 17.9%: 3.3% a year of appreciation, compounded, just to stand still, before tax, on the cheapest deep venue for a card of this size. Every dollar above it is taxed at up to 31.8%.

An outsider with real money can own vintage Magic without paying for the education twice by taking these steps in this order.

1. Read the three documents first. For the card you want: TCGplayer’s market price and history, the grader’s population report at your grade and the grade above, and the last five public sales with venues and dates. Heritage’s and Goldin’s archives are free; so is PSA’s CardFacts.
2. Choose the tier by budget, not by name. Under $2,000, played Revised dual lands and the Reserved List cards that Commander play data ranks highest; they are the only tier in this guide with a 19%-a-year record. From $5,000 to $50,000, graded Beta and Unlimited Power Nine in the highest grade that trades at least twice a year in public. Above $50,000, Alpha, and expect to be one of a few dozen buyers on earth.
3. Buy graded, on eBay through the Authenticity Guarantee or on TCGplayer from a top-rated seller, and never raw above $1,000. Check the certificate number on the grader’s site before paying, and if the card comes in a CGC or BGS holder and you plan to sell in a PSA market, price the crossover risk: a 9 in one holder is not a 9 in another.
4. On arrival, verify. Weigh it, light it, loupe it, even in the slab. Keep the listing, the certificate lookup and the invoice with the card; that file is your provenance and your basis.
5. Insure it and put it away. Scheduled coverage at $0.50–$1.00 per $100, a UV case, a dry safe, or a vault if you expect to sell to a buyer who will leave it there.
6. Decide the exit before the entry. eBay for anything under $50,000; a house only above it, with a negotiated seller’s commission of zero. Write down the price at which you sell and the date by which you sell regardless.
7. Ignore the new sets. Buy Collector Boosters to open, not to hold; if a serialised card falls into your hands, sell it in the first month while the licence is loud.
8. Keep records for the 28% and the 1099-K. Basis, fees, insurance, sale price, venue, date. The form may never come; the tax is owed either way.

These readings, with thresholds and dates, would change the view in this guide; the levels are as of September 9, 2026.

• The Goldin PSA 10 Alpha Black Lotus, closing September 19, 2026. The high bid was $535,000 on thirteen bids with twenty-six days to run, read in late August 2026; the lot was open at this guide’s as-of date. A result above the $540,000 of March 2023 says the top of the market is intact after three and a half years; a result under $500,000 says the private $3 million was a different market from the public one, and the reference price for the grade is the lower number.
• Hasbro’s third-quarter 2026 results, due in late October 2026. Magic grew 36% in Q1 and 32% in Q2 2026. Growth below 20% would be the first deceleration since the Universes Beyond run began; above 30%, with Star Trek still to come in November, the flood continues into 2027 and the mid-grade vintage slump has another year in it.
• The Unlimited Black Lotus in mid-grade at Heritage. A CGC 5 made $10,625 in July 2025 against $15,000–20,000 in 2022–23. A CGC 5 above $12,000 at a 2027 sale says the middle of the vintage market has turned; under $10,000 says the marginal collector is still elsewhere.
• The 2027 cadence. Wizards announced six main sets for 2027 in July 2026. A seventh, or a second 30th Anniversary-style product, changes the supply reading in the Hasbro section; a held line at six is the first year of restraint since 2023.
• The Reserved List itself. Any Wizards statement that narrows the reprint promise, or any non-tournament-legal reprint of a reserved card sold at retail, is the event this guide is priced against. The list has not changed since 1999, and the Rosewater remarks reported by ICv2 are the last word from inside the company.
• The census. Three PSA 10 Alpha Black Lotuses on the August 2026 trade reading (six on an older forum count) of about 105 graded, and three CGC Pristine 10s by CGC’s latest announcement, from one in April 2024. A PSA 10 count above six, or a fourth Pristine 10, is new supply at the very top and a reason to mark the reference price down.
• The fee schedules. eBay’s 13.25%, TCGplayer’s 10.75% and $75 cap, PSA’s $79.99 Regular tier and paused Value tiers, Heritage’s 22%. A resumption of PSA’s $24.99 Value Bulk tier lowers the cost of grading the played tier and raises the census.
• Underground Sea, Revised, Near Mint. About $1,045 on MTGStocks in September 2026. Above $1,300 the played Reserved List tier is compounding faster than its 2020–2026 record; below $750 the Commander demand that carried it has moved on, and the dual lands go on the list of things that fell with the list intact.

Here is the position, stated so that a reader in 2027 can check it against what happened.

Magic is two markets sharing a name. The first is a fixed supply of 1993–1999 cards under a thirty-year promise that has held through two tests, and its top tier, the Alpha and Beta Power Nine and the original dual lands, is the only part of any trading card game we would call an asset with a straight face: a known census, public sales with venues and dates, a play demand that does not depend on a licence, and a publisher with a commercial interest in keeping the promise.

The second is the fastest-growing product line at Hasbro, $1.72 billion in 2025 and a record $545 million in a single 2026 quarter, a consumer business whose serialised chase cards are manufactured scarcity inside unlimited print runs. The second competes with the first for the same collector, and on the 2024–25 record it won.

The honest record is not the $3 million. It is a signed PSA 10 that compounded at 2.6% a year from 2021 to 2023, an Unlimited Lotus that lost a third or more into 2025, and a Revised Underground Sea that did about 19% a year for six years because people kept wanting to play it. Add the cost stack, 13% to sell on the cheap venue and about a quarter of the buyer’s money through a house (a 22% premium plus a 10% commission), a 28% ceiling on the gain and $150 a year to insure a $20,000 card, and a five-year hold needs an 18% rise to return your money. That is the bar, and most of the Reserved List does not clear it.