Single-day surge of nearly 43%, Zhipu welcomes the God of Wealth, with a total market value exceeding HKD 320 billion, surpassing JD.com and Kuaishou

On the fifth day of the Year of the Horse in 2026 (February 21), when the朋友圈 was flooded with “Congratulations on your wealth,” an AI company was also completing a “God of Wealth welcoming” event.

Red Star Capital Bureau reported on February 21 that on February 20, the Hong Kong stock market experienced an AI frenzy ignited by Tsinghua alumni. Despite the Hang Seng Tech Index falling nearly 3%, the leading AI large model company Zhipu (02513.HK) defied the trend, closing up 42.72% in a single day to 725 HKD, setting a new high and reaching a total market value of 323.2 billion HKD. In just 43 days since listing, the company’s stock price has increased by 524% from the issuance price of 116.2 HKD.

On February 20, Kuaishou (01024.HK) fell 2.78%, with a total market value of 289.4 billion HKD; JD.com (09618.HK) declined nearly 2%, with a market value of 294.6 billion HKD; Ctrip (09961.HK) dropped nearly 1%, with a market value of 271.8 billion HKD; Baidu Group (09888.HK) decreased 6.25%, with a market value of 354.8 billion HKD. This means Zhipu has now entered the top tier of market capitalization in the Hong Kong TMT sector.

In terms of news, the company officially launched its new flagship model GLM-5 on February 12, which improves average performance in programming and development scenarios by over 20% compared to the previous generation, with a real programming experience approaching Claude Opus 4.5 level. It achieved top performance in three agent evaluations—BrowseComp, MCP-Atlas, and τ2-Bench—among open-source models.

After releasing GLM-5, due to high demand, Zhipu immediately increased the price of its GLM Coding Plan packages, with a 30% increase in China and over 100% overseas, becoming the first domestic AI native company to raise prices for large model commercialization services. The new packages sold out immediately, setting a new industry record for paid Chinese-made AI programming models.

Founded in 2019, Zhipu originated from Tsinghua University’s Knowledge Engineering Laboratory (KEG, one of China’s leading AI labs). Professor Tang Jie of Tsinghua’s Computer Science Department serves as Chief Scientist, Dr. Zhang Peng, a leader in Tsinghua’s Innovation Engineering, is CEO and General Manager, and Tsinghua alumnus Liu Debing is Chairman.

This company is known as “China’s OpenAI” and is a top competitor to OpenAI. By the end of 2020, Zhipu developed the GLM pretraining architecture; in 2021, it trained the 10-billion-parameter GLM-10B model and successfully trained a converged trillion-sparse model using MoE architecture. In 2022, it released and open-sourced the bilingual 130-billion-parameter GLM-130B, making it the only Asian large model included in Stanford’s evaluation in 2022.

In 2023, Zhipu launched the 100-billion-parameter ChatGLM dialogue model and upgraded it twice. The open-source version, ChatGLM-6B, enabled local fine-tuning and deployment for large model developers. Subsequently, Zhipu also open-sourced multimodal models CogAgent-18B and the 2024 GLM-4-9B series. In 2025, Zhipu announced open-sourcing the 32B/9B series models under the MIT license, allowing free commercial use by enterprises.

At the end of 2025, Zhipu released its new flagship model GLM-4.7, topping the Hugging Face global trend list. In major international large model evaluation rankings, GLM-4.7 ranked first overall among both open-source and domestic models.

As of 2025, Zhipu’s open-source models have been downloaded over 60 million times worldwide, with more than 2,000 related open-source projects on GitHub.

In December 2025, Zhipu disclosed its prospectus, announcing its pursuit of becoming the “world’s leading large model company” in Hong Kong. The prospectus shows that from 2022 to the first half of 2025, the company achieved revenues of 57.4 million, 124.5 million, 312.4 million, and 190 million yuan, respectively. The gross profit margins during these periods were 54.6%, 64.6%, 56.3%, and 50%. The losses during these periods were 144 million, 788 million, 2.958 billion, and 2.358 billion yuan.

The main cause of losses is R&D expenses. During the reporting periods, R&D spending was 84.4 million, 528 million, 2.195 billion, and 1.594 billion yuan. Computing power services accounted for a large portion of R&D costs, representing 17.3%, 58.9%, 70.7%, and 71.8% of R&D investment in each period. Zhipu stated in the prospectus: “The company is committed to leveraging the rapid expansion of the AI industry, which requires frequent and large-scale model training and retraining, necessitating more computing power from third-party providers.”

The prospectus shows that before the IPO, Zhipu completed eight rounds of financing, raising over 8.3 billion yuan, with a post-investment valuation of 24.38 billion yuan. Investors include well-known industry capital and investment institutions such as Meituan, Alibaba, Tencent, Xiaomi, Sequoia, Hillhouse, Yunhui, and Today Capital, as well as multiple state-owned funds.

Notably, after successfully listing in Hong Kong, Zhipu’s A-share IPO guidance is accelerating. On February 13, the China Securities Regulatory Commission website updated Zhipu’s IPO guidance progress. Zhipu withdrew its original guidance filing submitted in April 2025 and registered a new guidance record.

The latest guidance report indicates Zhipu plans to pursue the STAR Market (Science and Technology Innovation Board). Its IPO guidance agencies have changed from China International Capital Corporation (CICC) to Guotai Haitong Securities and CICC.

Red Star News Reporter Wu Danru

Editor Sun Zhicheng

Reviewer Guan Lì