Summary of Key Points
On July 9th, MiniMax, a Hong Kong-based AI large-model company, experienced the first round of restricted shares being released from lock-up. On that day, its stock price plummeted by more than 20%, and its market value fell below HK$100 billion (a decrease of approximately HK$320 billion from its historical high). In contrast, Zhipu saw a 13% increase on the same day despite the release of restricted shares. The main differences between the two companies stem from MiniMax's higher proportion of released shares (63%) and the larger presence of financial investors, which put significant pressure on the stock price. Additionally, recent feedback on MiniMax's M3 model has been mixed, and there are doubts about its commercialization strategy, along with ongoing high losses, further exacerbating the stock price pressure. However, some institutions still see potential for its long-term commercial success, and the future valuation recovery will depend on model iteration and monetization capabilities.
Direct Causes of the Sharp Drop on the Release Day: High Proportion of Released Shares + Desperate Need to Monetize by Financial Investors
"Release from lock-up" simply means that shares that were previously restricted from trading are now available for free. This time, MiniMax released 63% of its total Hong Kong shares (146 million shares), a particularly high proportion—almost two-thirds of the company's shares could potentially be sold on the market. More importantly, a large portion of these released shares were held by financial investors who are not interested in long-term cooperation but aim to profit from price differences, giving them strong incentives to sell.
In contrast, Zhipu had a lower proportion of released shares, and the majority of the releasing investors were state-owned "cornerstone investors" who typically commit to holding shares for the long term. Nearly 70% of these investors stated they would not reduce their holdings, resulting in almost no pressure on the stock price; instead, market confidence led to a 13% increase.
Although MiniMax's strategic shareholders, including Alibaba and MiHaoYou, as well as its founding team, have promised not to reduce their holdings, the significant selling pressure from financial investors was too much to overcome.
Differences from Zhipu: More Than Just Release of Restricted Shares—Model and Commercialization Issues
Half a year ago, both companies were leaders in the AI large-model field, but now their market values differ by HK$70 billion. The core reasons for this difference lie in their "hard capabilities" and profit-making prospects:
- Model Competitiveness: Zhipu's model may be more stable, with better user feedback. However, MiniMax's M3 model received mixed reviews from the market after its release in June, and its pricing strategy caused dissatisfaction among developers, leading to a 15% drop in the stock price on that day.
- Commercialization Prospects: Investors buy stocks based on their expectations of future profitability. Zhipu appears to have a clearer path to monetization (e.g., multiple corporate clients and stable revenue growth), while MiniMax's monetization strategy is uncertain, with pricing controversies raising doubts about its ability to convert technology into cash.
The M3 Model Fails to Appeal: Poor Market Response, Adding to Stock Price Pressure
MiniMax's M3 model was expected to be a major strength, but it became a weakness:
- After the release, some felt that the model's improvements were minimal, and others criticized its pricing (e.g., high fees for developers), leading to negative market sentiment and a 15% drop in the stock price.
- As a result, Citibank lowered MiniMax's target price from a higher figure to HK$533, citing weak market response, uncertain user retention, and unclear monetization strategies, which will continue to pressure the stock price in the short term.
Mixed Performance: Heavy Losses Despite Some Optimism from Institutions
MiniMax is still in the "money-burning" phase:
- Concerns: Although revenue increased by 159% (US$79.04 million) in 2025, losses also rose by 302% (US$1.87 billion), indicating that more money was spent on research and marketing than was earned.
- Optimism: Institutions like Goldman Sachs are still hopeful about its future. They believe the management can achieve annual recurring revenue (ARR) of US$1 billion by the end of 2026, which represents a sustainable source of income. Additionally, the industry's "price war" may be coming to an end, as competitor DeepSeek announced price increases, suggesting that AI large-model pricing is becoming more rational. This could allow MiniMax to charge fair prices and compete more effectively.
Recovery from Short-Term Pressure: What Determines Valuation?
The release of restricted shares is just a short-term challenge. Whether MiniMax's stock price can recover depends on two key factors:
1. Model Iteration: Can the company develop a better model that gains market recognition for its technical capabilities?
2. Commercialization Success: Can it effectively monetize its technology, such as by acquiring more corporate clients and achieving its ARR targets?
If these goals are met, investor confidence will return, and the market value should recover. Otherwise, the recent drop may just be the beginning of a longer-term downward trend.
In summary, MiniMax faces both short-term pressures and long-term challenges. The release of restricted shares was merely a catalyst; the real test lies in transforming its technical advantages into commercial success. For investors, when evaluating AI companies, it's essential to consider more than just concepts—whether their models are competitive and whether they can generate profits, as these factors determine the long-term direction of stock prices.