Zhipu plans $4bn Hong Kong share sale after AI stock surge

Zhipu AI, also known as Knowledge Atlas Technology, is seeking to raise about $4 billion through an accelerated bookbuild in Hong Kong. The company plans to offer 19.78 million new shares at HK$1,588 to HK$1,698 each, a discount of roughly 7% to 13% to the last close of HK$1,825. The placement would increase the company’s total share capital by about 4.2%. China International Capital Corp is the sole overall coordinator for the transaction.

Zhipu intends to use the proceeds for research and development, including hiring, computing capacity and technical services, as well as for business expansion, strategic investments, mergers and acquisitions, working capital and general corporate purposes. The company has said it will also seek to optimize its capital structure.

The stock climbed 13.4% in the most recent trading session and has risen nearly 1,500% since its January listing. Earlier this year Zhipu raised HK$4.35 billion in a Hong Kong initial public offering and it plans to raise about 15 billion yuan through a proposed listing on Shanghai’s STAR Market.

Chinese regulators have met with major technology firms to discuss potential limits on overseas access to advanced domestic AI models, creating regulatory questions for developers. Zhipu recently launched its GLM-5.2 large language model and made the model freely available to developers to encourage wider use.

Other Chinese technology and semiconductor companies have also tapped Hong Kong capital markets. Nexchip Semiconductor priced an IPO expected to raise about HK$6.98 billion. Shanghai Iluvatar CoreX Semiconductor is seeking about $850 million through a Hong Kong share sale and Shanghai Biren Technology raised about $900 million through a placement.

“Raising $4 billion gives Zhipu significant firepower to compete, but increases pressure to show heavy AI investment can deliver sustainable commercial returns,” Glenn Yin, research director at brokerage ACCM, noted.

The company says the placement will fund expansion of computing capacity and hiring to scale its models and commercial offerings.