MiniMax Hong Kong IPO Preview
MiniMax is getting ready to complete its IPO in Hong Kong in early January 2026. MiniMax is expected to raise at least US$500 million and as much as US$700 million.
Backed by Alibaba (9988 HK), MiniMax is targeting a valuation of more than US$4 billion.
MiniMax develops multimodal AI models including MiniMax M1, Hailuo-02, Speech-02 and Music-01, which can process text, audio, images, video and music.
MiniMax is getting ready to complete its IPO in Hong Kong in early January 2026. MiniMax is expected to raise at least US$500 million and as much as US$700 million. Backed by Alibaba (9988 HK), MiniMax is targeting a valuation of more than US$4 billion.
Founded in early 2022 by former SenseTime executive Yan Junjie, MiniMax has emerged as one of China’s prominent AI companies. MiniMax develops multimodal AI models including MiniMax M1, Hailuo-02, Speech-02 and Music-01, which can process text, audio, images, video and music.
MiniMax Company Background
MiniMax is a global artificial intelligence foundation model company focused on advancing general-purpose AI systems capable of performing a full spectrum of human intellectual tasks. The company operates at the intersection of foundational AI research and large-scale product commercialization.
MiniMax believes that artificial general intelligence will reshape productivity, creativity, and economic structures globally, and aims to be a central contributor to this transformation by building scalable, cost-efficient, and widely accessible AI technologies. The company believes that foundation models will become a core infrastructure layer for the global digital economy, similar in importance to operating systems, cloud computing, and the internet itself.
MiniMax operates in a market environment characterized by strong demand for increasingly capable, multimodal, and agentic AI systems. By 2030, the global foundation model market is projected to exceed US$300 billion, reflecting widespread integration of AI models into software, devices, content creation, enterprise workflows, and digital services. According to IDC estimates, artificial intelligence is expected to contribute US$19.9 trillion cumulatively to the global economy by 2030, accounting for approximately 3.5% of global GDP in that year alone.
The company emphasizes that real-world human interaction is inherently multimodal, involving text, speech, images, video, emotion, and context. As a result, MiniMax differentiates itself by committing from day one to the development of multimodal foundation models, rather than focusing narrowly on text-only systems. MiniMax also places strong emphasis on cost efficiency. Rather than pursuing scale at any cost, the company seeks to deliver state-of-the-art performance while reducing computational expense, ensuring that advanced AI remains affordable and accessible to a global user base.
MiniMax has built a comprehensive and competitive suite of proprietary foundation models spanning language, video, speech, and audio generation. These models are designed to work independently or in integrated, multimodal systems.
Large Language Models: MiniMax M Series - The MiniMax M Series represents the company’s flagship large language model family.
MiniMax-M1 - Launched in June 2025, MiniMax-M1 is an open-source, large-scale hybrid-attention reasoning model. MiniMax-M1 was positioned as a foundational platform for developers and researchers building next-generation agentic systems.
MiniMax-M2 - MiniMax-M2 is the company’s latest and most advanced large language model. It builds on the M1 foundation while delivering substantial performance improvements. MiniMax-M2 is designed not only as a reasoning engine but as a practical production-grade model optimized for real-world deployment at scale.
Video Generation Model: Hailuo-02 - The Hailuo-02 series is MiniMax’s proprietary video generation model, capable of producing high-quality video content from diverse input formats. Upon release, Hailuo-02 demonstrated competitive performance on global benchmarks and was commercialized rapidly at scale. Its intuitive interface enables content creators, advertisers, and marketers to generate professional-grade video outputs using simple prompts, significantly lowering the barrier to high-quality visual production.
Speech Generation Model: Speech-02 - The Speech-02 model series focuses on text-to-speech and audio synthesis. Released in April 2025, Speech-02 achieved global recognition as a top-performing speech generation model. Speech-02 serves as the backbone for MiniMax’s audio products and supports interactive, real-time applications across consumer and enterprise use cases.
AI-Native Product Ecosystem - MiniMax translates its foundation model capabilities into a diverse portfolio of AI-native products designed for individual users, enterprises, and developers. As of September 30, 2025, MiniMax’s AI-native products had served over 200 million individual users, reached users in more than 200 countries and regions, and supported over 100,000 enterprises and developers. This scale reflects both global demand and the company’s ability to operationalize AI research into commercially viable platforms.
MiniMax: Intelligent Agent Application - MiniMax is the company’s flagship intelligent AI agent application. It enables users to issue natural language instructions to autonomously execute complex tasks within a unified workspace. MiniMax Agent exemplifies the company’s vision of AI as an autonomous collaborator rather than a passive tool.
Hailuo AI: Visual Generation Platform - Hailuo AI is MiniMax’s flagship visual generation platform, fully integrated with the Hailuo-02 model. Hailuo AI has quickly become one of the most popular AI image and video creation platforms worldwide, particularly among creators, marketers, and advertisers.
MiniMax Audio: Audio Generation Tool - MiniMax Audio provides users with high-fidelity audio and speech generation capabilities via a web-based platform. The product is designed for creators, educators, enterprises, and developers requiring professional-grade audio output.
Talkie / Xingye: Multi-Modal Entertainment Platform - Talkie (international markets) and Xingye (China domestic market) represent MiniMax’s AI-native entertainment offerings. Powered by MiniMax’s proprietary models, Talkie/Xingye demonstrates the emotional and experiential potential of AI beyond productivity-focused use cases.
MiniMax Open Platform - The MiniMax Open Platform provides scalable AI services to enterprises and developers via public APIs. As of September 30, 2025, MiniMax Open Platform is available in over 100 countries and regions. It supports text, video, and audio model integration. Key industries served include smart devices, health care, finance, cultural tourism, and Internet services. The platform ranks among the world’s largest open AI platforms by average daily token volume, signaling strong developer adoption and enterprise demand.
MiniMax’s AI-native product ecosystem has experienced rapid growth. Average monthly active users (MAU) of AI-native product expanded more than sixfold from 3.1 million in 2023 to 19.1 million in 2024, and further increased to 27.6 million in the nine months ended September 30, 2025.
The expanding consumer user base provides continuous product feedback, enabling MiniMax to iterate rapidly and enhance user experience. The company has recorded quarter-over-quarter MAU growth over the past three quarters, driven primarily by strong user engagement and favorable ratings of the Talkie/Xingye application, as well as the launch of Hailuo AI in the third quarter of 2024.
Talkie/Xingye, which represents MiniMax’s multi-modal entertainment offering, continues to benefit from integration of new foundation model capabilities that enhance immersion, aesthetics, and emotional interaction quality. Meanwhile, Hailuo AI, powered by the Hailuo-01 and Hailuo-02 model series, has achieved globally competitive performance, allowing users to produce visually differentiated, cinematic-quality video content compared with other platforms.
Monetization across AI-native products has accelerated alongside user growth. The number of paying users increased from approximately 119,700 in 2023 to 650,300 in 2024, and further expanded to approximately 1.77 million in the nine months ended September 30, 2025.
Source: Company data
Source: Company data
Source: Company data
MiniMax’s AI-native products experienced strong expansion in new user acquisition, with new users increasing more than nine-fold from 11.1 million in 2023 to 104.2 million in 2024. New users also grew year-over-year from 65.4 million in the nine months ended September 30, 2024 to 96.9 million in the same period of 2025, reflecting continued global adoption despite a more disciplined growth strategy.
Monetization metrics continued to strengthen. Average spending per paying customer across AI-native products increased from US$6 in 2023 to US$11 in 2024, and further to US$15 in the nine months ended September 30, 2025. Hailuo AI demonstrated particularly strong monetization momentum, with average spending per paying customer rising from US$36 in 2024 to US$56 in the nine months ended September 30, 2025. Average spending per paying customer for Talkie/Xingye declined due to rapid expansion into a broader, lower-spending user base; however, this reflects widening market reach.
Source: Company data
