
Zhipu weighs a multibillion-dollar Hong Kong share sale after a 2,000% IPO surge. The listing could fund AI expansion and raises dilution questions.
Chinese AI firm Zhipu is in early talks to raise billions of dollars through a secondary share sale in Hong Kong, according to people familiar with the matter. The Beijing-based company, whose large language models compete with those from OpenAI and Baidu, has seen its shares surge almost 2,000% since its January initial public offering on Shanghai's STAR Market.
A secondary listing in Hong Kong would give Zhipu access to a broader pool of international investors. It would also provide an exit path for early backers including Alibaba Group Holding Ltd. and Tencent Holdings Ltd. The company's market capitalization has ballooned to over $50 billion following the stock's rally, the people said.
The discussions are preliminary, the people added. The size and timing of the deal could change. Zhipu's board has not yet formally approved the plan.
Hong Kong has attracted a wave of secondary listings from Chinese technology companies in recent years. Firms seek to diversify their investor base and hedge against regulatory risks on the mainland. Zhipu's move would be among the largest such offerings in 2025 if it proceeds.
For investors, the potential share sale raises several considerations. A large secondary offering could dilute the holdings of existing shareholders, including retail investors who rode the stock's meteoric rise. The additional capital could fund Zhipu's heavy spending on AI model training and computing infrastructure. The company faces intense competition from domestic rivals like Baidu's Ernie and SenseTime, as well as global players.
Zhipu declined to comment on the report. The company's IPO in January raised roughly $1.5 billion and was oversubscribed multiple times. A Hong Kong listing would be another milestone for the Chinese AI sector, which has attracted global investor interest despite export controls on advanced chips and ongoing geopolitical tensions.
If the deal proceeds, Zhipu would join a growing list of Chinese tech firms with dual listings in Shanghai and Hong Kong. The Hong Kong Stock Exchange has relaxed some listing rules in recent years to attract such issuers.
The surge in Zhipu's stock is part of a broader rally in Chinese AI shares this year. The CSI Artificial Intelligence Index has gained more than 50% since January, driven by optimism over China's ability to compete in AI despite chip restrictions. Zhipu's rise outpaced many peers. People familiar said the valuation reflects its position as one of the few Chinese startups offering a viable alternative to Western large language models.
A Hong Kong listing could also help Zhipu attract longer-term institutional investors. Many of these are restricted from directly buying Shanghai-listed shares. That investor base could provide more stable support for the stock.
The offering comes with risks. The Hong Kong market has seen volatile trading for Chinese tech stocks. Secondary listings often face pressure from geopolitical risks and local liquidity conditions. Zhipu's high valuation could also be tested if AI adoption in China slows.
Zhipu's 2,000% gain since January ranks among the largest stock surges for any large-cap tech company in China this decade. The rally reflects investor belief that Zhipu's models can capture a share of the rapidly growing Chinese AI market, which McKinsey estimates could be worth $1 trillion by 2030. The secondary listing would also test global investor tolerance for Chinese AI stocks. Western funds have generally been cautious due to export controls and data security rules. A successful Hong Kong listing could encourage other Chinese AI startups to pursue dual listings.
The company's board is expected to discuss the listing plan in coming weeks. No timeline for a filing has been disclosed.
Prepared with AlphaScala research tooling and grounded in primary market data: live prices, fundamentals, SEC filings, hedge-fund holdings, and insider activity. Each story is checked against AlphaScala publishing rules before release. Educational coverage, not personalized advice.