MiniMax Files for Hong Kong IPO as AI Unicorn Sees Revenue Surge
MiniMax, a leading Chinese artificial intelligence startup, has filed for an initial public offering in Hong Kong, marking a significant milestone as the company seeks to capitalize on the global boom in generative AI. The Shanghai-based firm, formally known as MiniMax, submitted its post-hearing information pack to the Hong Kong Stock Exchange on Sunday, positioning itself to potentially set a record for the fastest listing by an AI company from inception.
The filing reveals aggressive commercial momentum, with MiniMax reporting revenue of US$53.4 million for the first nine months of 2025, a year-on-year increase of approximately 175%. This growth is underpinned by a massive user base that has expanded to over 212 million individual users across more than 200 countries and regions. Notably, over 70% of the company's revenue represents contributions from overseas markets, highlighting significant traction in its global expansion strategy.
Despite the capital-intensive nature of training large models, MiniMax demonstrated improved operating leverage. The company reported an adjusted net loss of US$186 million for the first three quarters of 2025, which remained largely flat compared to the previous year despite the exponential revenue growth. The listing is set to test investor appetite for China's burgeoning AI sector, with the company backed by prominent investors including Tencent, Alibaba, and gaming giant miHoYo.
Founded in late 2021, the company has rapidly scaled its operations, distinguishing itself through a high-efficiency organizational structure and a "full-modal" technical approach. As one of the highest-valued AI startups in China, the IPO proceeds are intended to fund the sustained development of foundation models and the expansion of its AI-native product ecosystem.
Narrowing Losses Amid Global Expansion
MiniMax has managed to decouple its expense growth from revenue expansion, a trend described in its prospectus as a healthy "scissors gap." While revenue surged nearly 175% in the first nine months of 2025, research and development (R&D) expenses grew by only 30%, and sales and marketing expenses actually utilized a decrease of 26%. This divergence indicates that growth is being driven by product capability and user retention rather than aggressive subsidies or marketing burn.
The company's financial health is further evidenced by its cash conversion cycle. The accounts receivable turnover capability stands at 38 days, significantly lower than the industry average of 60 to 90 days for AI and SaaS companies. As of September 30, 2025, MiniMax held cash reserves exceeding US$1.1 billion, providing a runway for over 50 months of operation.
Dual-Engine Growth Strategy
The company operates on a "Model as Product" philosophy, driving revenue through two primary segments: AI-native consumer applications and enterprise services.
On the consumer side (To C), products such as the AI companion app Talkie (known as Xingye in China) and the productivity tool Hailuo AI have seen paid users increase 15-fold in under two years. Consumer segment revenue grew 181% year-on-year. On the enterprise side (To B), MiniMax provides API access via its open platform, which handles over a trillion token requests daily. The enterprise segment recorded a gross margin of 69.4%, demonstrating strong profitability potential in its developer services.
Strategic partnerships have bolstered this ecosystem. MiniMax models are available on global cloud platforms including Amazon Web Services, Google Cloud, and Microsoft Azure. Domestically, its technology supports core products for major tech firms such as Xiaomi Corp and Kingsoft Office.
Full-Modal Technical Ambitions
MiniMax distinguishes itself as one of the few startups focusing on "full-modal" R&D from day one, covering text, speech, and video simultaneously. The company’s technical matrix includes the MiniMax M2 text model, the Speech-02 audio model, and the Hailuo-02 video generation model.
In October 2025, the company released its open-source MiniMax M2 model, which utilizes "Interleaved Thinking" technology to improve complex task planning and coding capabilities. The model ranked among the top five globally on the Artificial Analysis leaderboard. In the video domain, the Hailuo-02 model recently ranked second globally in video competitive arenas.
The prospectus highlights the efficiency of MiniMax's R&D spend. Cumulative spending from inception to September 2025 totaled roughly US$500 million (RMB 3.5 billion). The company asserts this represents less than 1% of the cumulative expenditure of global leader OpenAI, despite achieving tier-one status across all modalities.
Gen Z Leadership and Governance
MiniMax is characterized by an exceptionally young leadership team and workforce. The company has approximately 385 employees, with an average age of 29. Nearly 74% of the staff are dedicated to R&D, including alumni from Google, Meta, and Microsoft.
The board of directors reflects this demographic, with four executive directors averaging just 32 years of age. Founder and CEO Yan Junjie, 36, is a former vice president at SenseTime Group, while other key technical leaders are part of the "Gen Z" cohort. This structure is designed to foster an "AI-native" organizational culture that prioritizes flattened hierarchies and rapid iteration.
According to the filing, approximately 70% of the net proceeds from the IPO will be allocated to R&D over the next five years to upgrade large models, with the remaining 30% designated for working capital and general corporate purposes.