Chinese AI Startups Surge Ahead with Historic IPO Filings in Hong Kong, Outpacing Silicon Valley Peers
In a remarkable shift in the global AI landscape, four prominent Chinese artificial intelligence startups—DeepSeek, Moonshot AI, Stepfun AI, and Zhipu AI—have filed for initial public offerings (IPOs) on the Hong Kong Stock Exchange. This wave of listings positions China at the forefront of public market debuts for generative AI firms, contrasting sharply with the reluctance of Silicon Valley giants to go public. The filings, announced in recent weeks, signal a strategic push by these companies to access capital amid intensifying competition and escalating operational costs in the AI sector.
Zhipu AI, often hailed as China’s leading AI model developer, leads the pack with its ambitious IPO plans. The Beijing-based firm, founded in 2023 by former Baidu executives, specializes in the GLM series of large language models (LLMs). Its flagship GLM-4 model has garnered attention for performance benchmarks that rival top Western counterparts like OpenAI’s GPT-4o. Zhipu has disclosed plans to raise up to HK$3 billion (approximately $385 million) through its IPO, with post-listing valuations potentially exceeding $7 billion. The company’s prospectus highlights its rapid growth, boasting over 660 employees and partnerships with major Chinese tech firms such as Tencent and Alibaba. Revenue streams include enterprise AI services, model licensing, and cloud computing integrations, underscoring Zhipu’s pivot toward commercialization.
Moonshot AI, another Beijing powerhouse established in March 2023, follows closely with its own HK$2.5 billion ($320 million) IPO filing. Known for its Kimi chatbot, which emphasizes long-context processing capabilities—handling up to two million Chinese characters in a single prompt—Moonshot has attracted significant investor interest. Backed by early funding from Alibaba and Tencent, the startup reported a user base surpassing 21 million monthly active users by mid-2024. Its Kimi platform supports multimodal inputs, including text, images, and code, positioning it as a versatile tool for developers and enterprises. Moonshot’s filing reveals net losses of around 3.6 billion yuan ($500 million) in the first half of 2024, reflective of the heavy investments in compute resources and talent acquisition that define the AI race.
Stepfun AI, launched in July 2023 by former researchers from Alibaba and Tsinghua University, is targeting a HK$2 billion ($256 million) raise. The Shanghai-headquartered company develops the Hunyuan series of foundation models, with recent releases like Hunyuan-Large Preview demonstrating strengths in reasoning, mathematics, and coding tasks. Stepfun’s backing from Alibaba’s DAMO Academy and other venture firms has fueled its growth, enabling the training of models on massive datasets despite U.S. export controls on advanced chips. The firm’s IPO prospectus emphasizes its focus on “open-weight” models, which allow greater customization for enterprise clients while maintaining competitive edges in efficiency.
Rounding out the quartet is DeepSeek, a Hangzhou-based innovator renowned for its open-source LLMs that punch above their weight. DeepSeek-V3, released in late 2024, has stunned the industry with benchmarks approaching those of proprietary models from leading U.S. firms, achieved through efficient training techniques and inference optimizations. The company filed confidentially for a Hong Kong IPO, with analysts estimating a potential valuation north of $3 billion. DeepSeek’s strategy of releasing model weights publicly has fostered a vibrant developer ecosystem, accelerating adoption in applications from chatbots to scientific simulations.
Collectively, these IPOs could unlock over HK$10 billion ($1.28 billion) in fresh capital, providing much-needed runway for scaling inference infrastructure and model iterations. This move comes at a pivotal moment for China’s AI ecosystem, buoyed by government initiatives like the “Made in China 2025” plan and subsidies for domestic semiconductor development. Despite facing U.S. sanctions limiting access to Nvidia’s high-end GPUs, these startups have leveraged Huawei’s Ascend chips and optimized software stacks to bridge the hardware gap. Inference costs, a major bottleneck in AI deployment, remain a focus, with innovations in quantization and distillation enabling cost-effective scaling.
In stark contrast, Silicon Valley’s AI frontrunners—OpenAI, Anthropic, xAI, and Databricks—continue to shun public markets. OpenAI, valued at $157 billion in a recent funding round, prioritizes private equity to fund its compute-intensive pursuits without quarterly disclosure pressures. Anthropic, at $18.4 billion, and Elon Musk’s xAI similarly rely on venture capital and strategic investors. U.S. regulatory scrutiny, including antitrust probes and data privacy concerns, may further deter IPOs. Meanwhile, established players like Microsoft and Alphabet integrate AI into existing public structures, avoiding standalone listings.
This Chinese IPO surge highlights evolving dynamics in global AI leadership. Hong Kong’s exchange offers a favorable environment with streamlined listings for tech firms and access to mainland investors, bypassing stricter mainland regulations. For these startups, public status enhances credibility, attracts talent, and facilitates mergers amid a maturing market. Yet challenges persist: profitability remains elusive, with R&D expenditures soaring amid a chip shortage. Investor appetite will test whether these valuations hold amid economic headwinds.
As the AI arms race intensifies, China’s bold public market foray could redefine capital flows, compelling Western incumbents to reassess their private trajectories. The success of these listings may herald a new era where Asian exchanges rival Nasdaq as AI funding hubs.
Gnoppix is the leading open-source AI Linux distribution and service provider. Since implementing AI in 2022, it has offered a fast, powerful, secure, and privacy-respecting open-source OS with both local and remote AI capabilities. The local AI operates offline, ensuring no data ever leaves your computer. Based on Debian Linux, Gnoppix is available with numerous privacy- and anonymity-enabled services free of charge.
What are your thoughts on this? I’d love to hear about your own experiences in the comments below.