Summary of Key Points
Today, Zhipu's stock price in the Hong Kong market first soared and then declined (with a maximum increase of 21.75%, closing at a gain of 3.45%). The company also announced plans to raise approximately HK$31.4 billion through a new share placement in the Hong Kong market and an additional RMB 15 billion through an IPO on China's STAR Market (Science and Technology Innovation Board). In total, Zhipu aims to secure around RMB 44.8 billion to fund its large model research and development efforts. The funds will be primarily used for research and development, infrastructure construction, and other related activities, as the competition in the large model industry is becoming increasingly competitive, with leading companies continuously investing heavily in technological advancements.
Detailed Analysis
1. Stock Price Volatility: Market Sentiments Driven by Placement News
Zhipu's stock opened slightly higher today, and bullish investors quickly entered the market, pushing the price up to HK$2,222 (a 21.75% increase). However, it then fluctuated and fell back to HK$1,888 (a 3.45% increase). The reason is straightforward:
- Initial Rise: Investors initially viewed the financing as a positive sign for research and development efforts, leading to buying enthusiasm.
- Subsequent Drop: When they realized that the placement price was 12.99% lower than the market price, they became concerned about the dilution of existing shareholders' equity (as new shares would reduce their ownership percentage), prompting them to sell, causing the stock price to drop.
2. Details of the Hong Kong Market Placement
Zhipu is selling 19.78 million new shares to specific investors, accounting for 4.44% of its current total shares, which will increase to 4.25% after the placement. The price per share is HK$1,588, which is 12.99% lower than yesterday's closing price of HK$1,825 and 6.72% lower than the average price over the past five days.
Why a discount? The discount is designed to make the offer more attractive to investors, motivating them to invest quickly, given that the amount of funds involved (HK$31.4 billion) is substantial. This is a common financing strategy to improve the success rate of the placement.
3. Where Will the Funds Be Used?
The funds will be spent by the end of 2027 on three main areas:
- Research and Development: Hiring model experts, purchasing or leasing computing power (such as cloud servers), taking out loans for computing resources, and optimizing computing resource allocation to support the training and improvement of large models.
- Business Expansion: Engaging in strategic investments or acquisitions (e.g., acquiring technology teams).
- Operational Expenses: Optimizing the capital structure and replenishing working capital.
Combined with the remaining HK$308 million from its previous IPO and the planned RMB 150 billion from the A-share IPO, Zhipu will have a total of around RMB 44.8 billion available for large model development.
4. Why the Urgent Need for Financing?
The large model industry is characterized by high costs. Zhipu explains that the approach to AI computing power has changed this year, with continuous upgrades in large model capabilities and increasing demands from customers (e.g., companies using large models for customer service and analysis). The scope of applications is expanding, requiring significant investments in computing infrastructure and global market expansion.
In short, developing large models is like nurturing a costly endeavor: purchasing computing power and top-tier talent is expensive, and training a model can cost tens of millions or even hundreds of millions. Without financing, Zhipu would not be able to sustain its operations.
5. A+H Listing and Industry Competition
Zhipu also plans to list on China's STAR Market, establishing an "A+H" dual-listing structure that allows it to access funds from both domestic and international markets, providing more stable long-term funding.
The competition in the industry is fierce: OpenAI is working on GPT5.6 and GPT6 with over 4 trillion parameters, while domestic companies like DeepSeek and MiniMax are also developing models with over 2 trillion parameters. In the second half of 2026, there will be a rapid update of models, and competition will focus on optimizing computing power, innovative architectures, and practical application outcomes (e.g., whether the models can generate revenue for companies). Zhipu's financing is essential to stay ahead in this competitive landscape.
In One Sentence
Zhipu is leveraging both its Hong Kong market placement and A-share IPO to secure sufficient funds to maintain a competitive edge in the high-cost race of large model development. The industry is still in its early stages, but leading companies are investing heavily to secure their place in the future. Only those with sufficient capital will be able to survive and thrive.