China's 'Consumer Trinity' Loses Its Crown as Valuation Reality Bites
Pop Mart, Lao Feng Xiang, and Mixue Have All Shed More Than Half Their Peak Market Value — and the Rotation Into AI Tells Only Part of the Story
A toy, a gold pendant, and a cup of bubble tea once collectively justified a combined Hong Kong market capitalization exceeding HK$730 billion — a valuation that the underlying businesses can no longer credibly defend.
Pop Mart International Group, Lao Feng Xiang, and Mixue Group peaked almost simultaneously in June 2025, riding a wave of structural consumption shifts in mainland China and a voracious appetite for high-multiple growth stories in Hong Kong equities. By the third quarter of 2026, all three had surrendered more than half their peak values, with Lao Feng Xiang nursing a drawdown approaching 75%. The proximate catalyst differs by company, but the underlying diagnosis is the same: terminal growth rates were repriced, and the premium evaporated faster than management could replace the narrative.
The most recent data point arrived on October 27, 2026, when Pop Mart disclosed that Q3 2026 consolidated revenue had declined 45%–50% year-on-year, with its Americas segment — once the crown jewel of its globalisation story — falling 70%–75%. Twelve months earlier, the identical reporting period had produced growth of 245%–250%. Shares gapped lower by more than 6% at the open the following day. That single data point captures the central tension facing all three companies: when growth decelerates from hyper-velocity to contraction, a triple-digit price-to-earnings multiple offers no floor whatsoever.
Growth Deceleration Exposes the Fragility Baked Into Triple-Digit Multiples
To understand the magnitude of the re-rating, it is worth anchoring the peak valuations. At its August 2025 high of approximately HK$340 per share, Pop Mart carried a price-to-earnings ratio exceeding 105x. Lao Feng Xiang at its zenith traded at roughly 140x trailing earnings. Even Mixue Group, the most operationally straightforward of the three, commanded above 45x. At those levels, investors were not buying current earnings — they were buying a discounted cash flow model populated with heroic compounding assumptions extending five to ten years forward.
The logic was defensible at the time. In the twelve months through end-2025, Pop Mart's plush-toy segment generated revenue of RMB 18.708 billion (approximately US$2.60 billion), up 560.6% year-on-year, with overseas revenue expanding 291.9%. Lao Feng Xiang posted first-half 2025 revenue growth of 251% and net profit growth of 285.8%, fuelled by record international gold prices and consumers willing to pay a significant craft premium for its traditional-technique gold jewellery. Mixue Group's H1 2025 revenues rose 39.3%, underpinned by a franchise network that by June 2025 had reached 53,014 stores globally, with more than half concentrated in China's Tier-3 cities and below.
The critical error — made by both buy-side analysts and retail investors — was treating cyclically elevated growth rates as structurally durable. Shen Meng, director at Chanson Capital, argues that capital markets habitually repackage a company's transitional advantages into a long-term compounding thesis. The repackaging works until the first quarter when the numbers refuse to cooperate.
IP Fatigue Undermines Pop Mart's 'Hit Factory' Investment Thesis
The bear case against Pop Mart has never been about whether LABUBU is a compelling character. It has always been about whether Pop Mart can function as a repeatable IP-generation engine rather than a single-franchise phenomenon.
THE MONSTERS series, home to LABUBU, generated RMB 14.16 billion (US$1.97 billion) in 2025 revenue — 365.7% growth, representing nearly 40% of group turnover. The concentration risk was visible in plain sight. By June 2025, official restocking had already caused secondary-market prices for select LABUBU variants to retreat sharply. When LABUBU 4.0's "Retro Barbershop" series launched in June 2026, Deutsche Bank channel checks found first-day e-commerce sales running at only 30%–50% of prior major IP releases in 2026, with some standard variants trading below the official retail price of RMB 159.
New IP is growing — the Xinxingren IP posted 580.6% revenue growth in H1 2026 — but it is not yet large enough to offset the deceleration in the flagship franchise. Pop Mart founder Wang Ning acknowledged at the August 2026 earnings call that the company would likely miss its 20% full-year growth target. That admission, from the person with the most intimate knowledge of the pipeline, is the single most important data point for investors reassessing the IP-factory thesis.
The number of mainland China public mutual funds holding Pop Mart as a top-ten position collapsed from 311 at end-Q2 2025 to just 86 at end-Q2 2026 — a reduction of more than 70%. Aggregate mutual fund share holdings fell approximately 36% in a single quarter, from 33.248 million shares at end-Q1 2026 to 21.147 million at end-Q2 2026, according to Choice Financial Terminal data.
Gold Price Volatility Strips Away Lao Feng Xiang's Earnings Smoothness
Lao Feng Xiang's H1 2026 headline numbers appear robust in isolation: revenue of RMB 19.808 billion (US$2.75 billion), attributable net profit of RMB 4.267 billion (US$593 million), up 88.2% year-on-year. The problem lies beneath the aggregate.
Reverse-engineering the semi-annual results against the Q1 profit guidance range reveals that Q2 2026 attributable net profit likely came in between RMB 467 million and RMB 667 million — compared with RMB 3.6 billion–RMB 3.8 billion in Q1. The first quarter was artificially inflated by a 20%–30% price increase implemented in late February 2026, which triggered a surge of front-loaded purchases before the new price list took effect. When international gold prices corrected sharply in Q2 and consumer demand softened, management acknowledged the slowdown and opted against immediate strategic adjustment, with the board noting monthly average profit remained above RMB 200 million.
The revision cycle from sell-side analysts tells a stark story. In August 2026, Kaiyuan Securities cut its 2026–2028 attributable net profit forecasts for Lao Feng Xiang from RMB 90.12 billion / RMB 123.05 billion / RMB 155.37 billion to RMB 65.35 billion / RMB 67.43 billion / RMB 80.03 billion respectively. The revised 2027 forecast implies year-on-year profit growth of approximately 3% — a dramatic deceleration that is plainly incompatible with a 140x earnings multiple.
Lao Feng Xiang retains meaningful competitive moats: its ancient-craft brand positioning commands genuine consumer loyalty, overseas revenues continue to grow rapidly, and new product categories such as gold tableware are adding incremental revenue. The issue is not whether the business is good; it is whether the business is good enough to justify the valuation at which it was priced.
Mixue's Unit Economics Crack as Store Density Cannibalises Franchisee Returns
Of the three, Mixue Group's challenge is the most structurally uncomfortable, because it strikes at the core of the franchise model itself.
H1 2026 results showed group revenue increasing a modest 2.3% year-on-year while net profit declined 14.7% to RMB 2.319 billion (US$322 million) — the first post-IPO half-year profit contraction. Total global store count reached 63,987 as of end-June 2026, up approximately 20.7% year-on-year, yet revenue barely moved. The implied conclusion: incremental store openings are generating meaningfully less revenue per unit than the existing base.
The granular metrics confirm this. New franchise openings in H1 2026 totalled 5,455, down 2,266 from the same period in 2025. Net new additions of 4,166 compare unfavourably with 6,534 in H1 2025. Management conceded at the earnings call that per-store sales volumes for both the group overall and the flagship Mixue Bingcheng brand fell by double-digit percentages. Cross-referencing product and equipment sales revenue against average franchised store count yields an implied per-store procurement spending decline of approximately 17.9% year-on-year.
This matters because Mixue's business model derives revenue primarily from selling raw materials, packaging, and equipment to franchisees. Store count expansion drove revenue in earlier years; store density is now working against it, as franchisees in increasingly saturated markets buy less per store. Simultaneously, the cost base is expanding faster than revenue: selling and distribution expenses rose 22.9% in H1 2026, and administrative expenses climbed 39.4%, both running at multiples of the 2.3% topline growth rate.
The locus of risk has shifted from "can Mixue keep opening stores?" to "can existing franchisees remain economically viable?" Franchisee profitability is both the foundation of the model and, at present, its most acute vulnerability.
AI Hardware Absorbs the Premium — But Faces the Same Eventual Reckoning
The capital rotation into artificial intelligence hardware has been dramatic. According to Guosen Securities data, technology-sector holdings as a share of mainland public mutual fund top-ten positions rose 23.6 percentage points to 60.8% in Q2 2026, with the electronics subsector alone reaching 42.9%, up 21.2 percentage points. Consumer sector weighting fell 7.6 percentage points to 9.5% over the same period.
The reallocation is not irrational. Zhongji Innolight, the high-speed optical transceiver manufacturer that has emerged as the most visible AI infrastructure beneficiary in Chinese equities, reported H1 2026 revenue of RMB 41.778 billion (US$5.80 billion), up 182.5% year-on-year, with attributable net profit of RMB 13.651 billion (US$1.90 billion), up 241.7%. Hyperscaler demand for 800G and 1.6T transceivers to build out AI data centres is generating the kind of hard, contracted revenue that makes triple-digit growth rates in consumption look ephemeral. Gartner in September 2026 projected global AI spending would reach US$2.67 trillion in 2026, a 49.5% year-on-year increase, with AI infrastructure absorbing the largest share.
But investor Chen Momo draws the critical distinction: AI capital flows amplified the consumer stocks' decline, but did not originate it. "Capital moving away accelerated the fall. It did not cause the fall," she said. The valuation compression in Pop Mart, Lao Feng Xiang, and Mixue was already underway before the AI rotation intensified in H2 2025 — it simply became more acute as incremental liquidity dried up.
The more important observation is forward-looking. AI large-model companies such as Zhipu AI and MiniMax (MiniMax), which listed on the Hong Kong Stock Exchange in January 2026, illustrate that revenue growth alone does not insulate a company from severe de-rating. Zhipu reported H1 2026 revenue of RMB 954 million (US$132.5 million), up 399.7% year-on-year, alongside an adjusted net loss of RMB 1.964 billion — losses more than doubling revenues. MiniMax posted revenue of approximately US$117 million, up 283%, against an adjusted net loss of US$293 million, up 111.2%. Both stocks had fallen more than 80% from their all-time highs by October 20, 2026.
Hua Xia Fund portfolio manager Li Xiangjie has explicitly noted that institutional capital remains concentrated in upstream AI hardware, with downstream application companies still in a "differentiated observation phase." Even within AI hardware, early signs of rotation are visible: Q3 2026 mutual fund data compiled by Huafu Securities showed electronics-sector allocations declining 6.81 percentage points quarter-on-quarter to approximately 36%, with communications falling 1.69 percentage points to roughly 15%. Zhongji Innolight remains the largest mutual fund holding, but both the number of funds holding it and aggregate share count have declined from Q2 2026 peaks.
Contrarian Capital Enters Where Institutions Exit
The divergence between institutional behaviour and smart-money contrarianism is perhaps the most intellectually interesting subplot. While mutual funds were cutting Pop Mart exposure by more than 70%, investor Duan Yongping — the founder of BBK Electronics and a well-known long-term value investor — was accumulating.
Having stated in August 2025 that he "did not understand" Pop Mart, Duan reversed course after the company's full-year 2025 results. Beginning in April 2026, he initiated a position through cash-secured puts, subsequently converting to direct equity through multiple tranches. Following the Q3 2026 revenue disclosure that showed near-50% contraction, Duan publicly stated he would "continue buying," with a stated ceiling of 10% of total share capital — which would make him the second-largest shareholder behind founder Wang Ning.
The apparent contradiction — institutions selling, Duan buying — resolves when viewed through the lens of investment horizon and price-relative-to-intrinsic-value frameworks. Chen Momo notes that the same set of facts — slowing growth, elevated risks — can rationally support opposite positioning depending on whether an investor believes current prices have already priced in the downside scenario. The market's perpetual challenge is that this question has no objectively correct answer until time resolves it.
Impact Assessment: What Investors, Franchisees, and Suppliers Should Monitor
For Pop Mart, the critical variable is the Q4 2026 performance of non-LABUBU IP franchises globally, particularly whether Xinxingren demonstrates pricing power comparable to the MONSTERS series at a comparable stage of development.
For Lao Feng Xiang, the second-half gold price trajectory and the sustainability of the company's overseas retail expansion programme will determine whether the sell-side's subdued 2027 earnings forecast requires further revision downward.
For Mixue, the franchisee viability data — specifically whether per-store procurement spending stabilises or continues to deteriorate — will be the clearest leading indicator of whether the revenue model retains its network effect.
For AI hardware, investors should track whether Q4 2026 and Q1 2027 earnings cycles from key data-centre suppliers begin to show sequential deceleration consistent with hyperscaler capex digestion periods, which would trigger the same re-rating dynamic currently afflicting the consumer trio.
The 'Consumer Trinity' was not destroyed by AI. It was destroyed by the universal principle that any asset priced for perfection has no margin of safety when execution proves merely mortal.
Related Coverage:
Pop Mart Revenue Rises 24% as Overseas Growth Reverses and Inventory Piles Up
POP MART Reports 204% Revenue Surge Driven by LABUBU Mania and Global Expansion