Pop Mart

The mechanism is the blind box. A series of twelve figures sells for roughly RMB 59 to 99 each, in sealed identical packaging, with one 'secret' figure inserted at odds the company prints on the box — typically one in 144. The buyer pays the same price for a common figure and for the rare one, which means the secondary market does the pricing Pop Mart would be embarrassed to do itself. A completed set requires, on average, far more than twelve purchases, and the duplicates end up on trading apps that Pop Mart does not run and does not need to. Every purchase is a small lottery with a floor: you always get a figure, you just do not get to choose it.

Upstream of the box is the part I find more interesting, because it is where the durability lives. Pop Mart does not draw the characters. It signs artists — Kenny Wong for Molly in 2016, Kasing Lung for The Monsters in 2019 — and takes exclusive rights to produce and sell their work, paying royalties whose terms the company has never disclosed. What the artist gets is a factory, a store network and a fandom. What Pop Mart gets is a portfolio: in 2024 it reported thirteen separate character lines each earning over RMB 100 million, and the four largest — The Monsters, Molly, Skullpanda and Crybaby — brought in roughly RMB 7.6 billion between them. The company calls this 'proprietary IP' and it accounted for about 85% of revenue.

Downstream is a retail machine built to keep the lottery in front of people. About 400 stores in mainland China and over 2,300 'robo shops' — vending machines in malls and stations that sell a blind box for exact change and no staff — plus Tmall, Douyin and a drawing app of its own called Pop Draw. Overseas, which meant almost nothing in 2020, was 39% of revenue in 2024 and around 40% in the first half of 2025, with the Americas the fastest-growing region. The pieces fit together as a flywheel: the artist supplies novelty, the box supplies compulsion, the store supplies the encounter, and the fan supplies the marketing by posting the pull.

I want to be careful here, because the easiest mistake with Pop Mart is to confuse Labubu's moat with Pop Mart's. Labubu has no moat; it is a fashion. The question is whether the machine that made Labubu has one.

The case for yes rests on three things. First, the artist pipeline: Pop Mart has become the place an independent designer takes a character if she wants it manufactured and sold in 500 stores, which is a network effect in the loose sense — more artists bring more fans, more fans attract more artists. Second, physical distribution at a scale no rival in the category has: hundreds of stores and thousands of vending machines in locations where a fourteen-year-old actually is. Third, the membership data. Over 46 million registered members in mainland China at the end of 2024, contributing 93% of sales with a repurchase rate near 50%, which means the company can see which character is peaking before the secondary market does.

The case for no is that every layer of this has been copied. 52Toys, Miniso's Top Toy and a dozen smaller players sell blind boxes on the same terms; Miniso has more stores than Pop Mart; Bandai, Sanrio and Disney own characters with sixty years of memory against Labubu's decade. And the mechanism itself attracts regulators — China has already restricted blind-box sales to under-eights and required odds disclosure. So I score the moat as contested: real at the platform level, and genuinely absent at the level of any single character, which is what the share price has mostly been trading on.

Wang Ning's allocation record has one great decision, one expensive detour and a large open question.

The great decision was 2016: cutting the variety-store assortment down to designer toys and spending the company's scarce cash on exclusive artist rights and a factory relationship rather than on more shelves of stationery. Everything since sits on top of that. The IPO proceeds in 2020 went where you would want — stores, overseas subsidiaries, the Pop Draw app, and a run of artist signings that produced Skullpanda and Crybaby. Store capital is cheap and returns quickly; the robo-shop programme in particular is close to a textbook case of capital-light distribution.

The detour is Pop Land, the theme park the company opened in Beijing's Chaoyang Park in 2023, and the surrounding ambitions in animation, games and a jewellery line called Popop. None of these is large yet. But they are the moves of a company that wants to be Disney, and Disney's economics come from characters that have been loved for decades, not from characters that trended last summer. I would rather see that money go into signing the next twenty artists.

The open question is what a company with this much cash and this much volatility should do with the surplus. Dividends have risen each year and buybacks have been small. My honest view is that the surplus is insurance against the next 2022, and that management knows it.