MiniMax Plunges Over 18% on Lock-Up Expiry Day as Price Support Pledges Fail to Offset Selling Pressure
Taylor Wilson
MiniMax saw 153 million shares — 48.9% of total equity — unlocked on July 9, swelling its free float from under 6% to roughly 50%; despite over 80% of shareholders pledging not to sell, the stock fell more than 18% to HK$296 by midday on six times normal volume.
How large is this unlock?
153 million shares unlocked in a single day, representing 48.9% of total equity and roughly HK$59.4 billion at the prior close.
This means → the free float jumped from under 6% to about 50% — a near-tenfold expansion in tradeable supply and MiniMax's largest-ever lock-up expiry.
The stock opened down about 1%, then selling accelerated. By 13:18, MiniMax traded at HK$296, down over 18%, on volume of more than 12 million shares — roughly six times the daily average.
Over 80% of shareholders pledged not to sell — so why did the stock still fall?
The evening before, cornerstone investors including Aspex, Boyu, IDG and Janchor, state-backed entities such as China Life Investment, and early backers Alibaba, miHoYo, Yunqi and Mingshi all pledged not to reduce holdings — a breadth of commitment rarely seen in a Hong Kong tech lock-up.
Alibaba, the largest single external shareholder at roughly 13%, also subscribed for US$30 million in new shares. miHoYo voluntarily extended its lock-up by 12 months; the founding team did the same — well beyond the industry-standard six months.
In plain terms = everyone who pledged is a strategic holder. But more than a third of legacy shareholders are financial investors — Hillhouse, Sequoia China affiliates, Future Capital — who entered early, at low cost, with paper returns of several to more than ten times. They face fund-life deadlines and DPI (distributed-to-paid-in, the ratio of cash returned to limited partners) pressure. Even a small fraction choosing to lock in profits is enough to create heavy selling in the secondary market.
What is the "overlapping shareholder" dilemma?
Many of the shareholders who pledged to hold MiniMax also own stakes in Zhipu AI: Alibaba (including Ant Group) holds roughly 6% of Zhipu; Tencent invested RMB 200 million; Hillhouse and Sequoia China are early investors in both companies.
This means → these institutions are not picking sides — they are betting on the entire AI sector. But when "long Zhipu, short MiniMax" becomes a popular pair trade in Hong Kong, holders of both face pressure from both directions.
On July 9 afternoon, MiniMax fell more than 18% while Zhipu AI rose more than 5% over the same period — the divergence suggests some overlapping shareholders may have prioritised trimming their MiniMax positions.
How do the two AI companies compare on fundamentals?
Zhipu posted 2025 revenue of RMB 724 million with a 41% gross margin. Its MaaS platform — model-as-a-service, where enterprises call AI capabilities on demand — generates RMB 1.7 billion in annualised recurring revenue, mainly from state-owned clients, with a 95% renewal rate. Market cap exceeds HK$800 billion.
MiniMax reported revenue of roughly US$79 million (about RMB 570 million) over the same period, with a 25.4% gross margin and an adjusted net loss of US$251 million. Consumer-facing revenue accounts for 67% of total sales, but consumer gross margin is just 4.7%; its flagship model M3 was forced to cut prices by 50% within a week of launch, and paid conversion sits below 1%.
In plain terms = Zhipu earns predictable revenue from large enterprise clients; MiniMax relies on consumers but has not yet found a path to profitability. Capital flowing from the latter to the former is not surprising.
What to watch next?
MiniMax has shed more than 77% from its peak market cap of HK$410 billion three months ago, leaving it at roughly HK$100 billion and a price-to-sales ratio of about 290x.
This reflects a sharp market repricing of consumer-AI monetisation — not a rejection of the technology, but growing impatience with the question of whether it can make money.
With financial investors still under pressure to exit, whether the market can absorb this unlock wave smoothly remains an open question.
Content is for reference only, not financial advice.