Generation Y and Z dominate, with over $1 billion on the books, MiniMax knocks on Hong Kong's stock exchange door
TechFlow Selected TechFlow Selected
Generation Y and Z dominate, with over $1 billion on the books, MiniMax knocks on Hong Kong's stock exchange door
MiniMax is poised to set a new record as the AI company with the shortest time from founding to IPO.
Navigating Web3 tides with focused insightsSource: Jiqi Zhi Xin
Unexpectedly, the pace at which large model startups are going public has accelerated so rapidly.
Sunday evening news this week revealed that MiniMax (Xiyu Technology), a well-known domestic AI startup, has published its post-hearing information package (PHIP) version of the prospectus on the Hong Kong Stock Exchange, marking a critical sprint phase in its IPO process.
Founded in December 2021 and headquartered in Shanghai, MiniMax was established by Yan Junjie and other former executives from SenseTime. The company focuses on developing general artificial intelligence (AGI) technologies integrating text, speech, and vision multimodalities. It is widely recognized for its foundational large models MiniMax M1 and M2, as well as AI-native products such as MiniMax Voice, Hailuo AI, and Xingye.
The company is poised to set a new record, potentially becoming the AI firm with the shortest timeline from founding to IPO.
As a globally leading full-modal large model company, MiniMax has systematically disclosed its business layout, user scale, financial data, and future strategy for the first time in its prospectus, showcasing strong momentum and a unique positioning within the global AI landscape.
Over 200 Million Users, Overseas Revenue Exceeds 70%
MiniMax operates under a dual-engine model centered on "large models + AI-native applications." The company states that its open platform has become one of the world's largest enterprise and developer platforms, enabling rapid deployment across industries including smart devices, healthcare, tourism, culture, and finance.
The prospectus reveals that as of September 30, 2025, MiniMax serves over 212 million individual users across more than 200 countries and regions, along with over 100,000 enterprises and developers from over 100 countries. The average monthly active users of its AI-native products surged from 3.144 million in 2023 to 27.622 million during the first nine months of 2025, demonstrating strong user stickiness and growth momentum.
This vast user base spans the globe: MiniMax discloses in its prospectus that overseas revenue accounts for over 70% of total revenue, highlighting significant achievements in international expansion and cross-market commercialization.
Financially, MiniMax generated $53.437 million in revenue during the first nine months of 2025, representing an approximately 174.7% year-on-year increase, reflecting high-speed growth. Revenue primarily comes from two segments: subscriptions and in-app purchases for AI-native products, and services provided through its open platform and enterprise solutions.
Adopting a "model-as-product" philosophy, efficient operations have led to a healthy financial cycle: its accounts receivable turnover days stand at just 38 days, far below the 60–90 day average in the AI or SaaS industry, demonstrating exceptional cash recovery capability and operational efficiency.
On the consumer (To C) side, MiniMax directly serves global consumers via native products like Hailuo AI, Xingye / Talkie (AI social), and MiniMax Voice. As of September 30, To C revenue grew 181% year-on-year, while the number of paying users increased fifteenfold in less than two years.
On the enterprise (To B) side, MiniMax empowers multiple industries through its open platform, offering advanced model capabilities via APIs to businesses and developers. Its open platform processes over one trillion tokens per day on average, achieving a 160% revenue growth in the same period, with a gross margin as high as 69.4%, indicating robust profitability.
MiniMax has also built a cooperative ecosystem with leading tech companies both domestically and internationally. Abroad, its models are available on the global AI platforms of Amazon, Google, and Microsoft’s three major cloud providers; it supplies video and voice technologies to LinkedIn and Monks (a digital creative agency). In China, MiniMax’s technology powers core products at Alibaba, Tencent, ByteDance, Xiaomi, Kingsoft Office, and provides human-like interaction capabilities for humanoid robots such as Zhixiang Robotics. This broad and deep industrial integration proves the practicality and reliability of its technology.
According to the prospectus, MiniMax’s adjusted net loss in 2025 remained nearly flat compared to the same period last year, achieving effective narrowing of losses amid rapid growth. This is attributed to a diversified revenue model and efficient cost management—by September 2025, R&D expenses increased by only 30% year-on-year, while sales and marketing costs dropped by 26%.
Full-Modal AI Technology Strategy
MiniMax’s success rests on a clear and efficient path: building solid foundational competitiveness through a forward-looking “full-modal” technical matrix.
MiniMax is among the few large model companies that have focused exclusively on full-modal model development since inception. Since its founding, the company has maintained intensive technological iteration, achieving breakthroughs in speech, video, and text models, frequently drawing attention from the tech industry.
-
Starting in 2023, MiniMax launched Speech 01, the first domestic speech large model based on the Transformer architecture. It released the upgraded version Speech 02 in 2024, ranking first in overall performance. To date, MiniMax’s speech model has helped users generate over 220 million hours of audio content.
-
In August 2024, MiniMax launched its video generation model Video 01 and the video product Hailuo AI. In June 2025, it introduced the upgraded Hailuo 02, ranking second in the AA Video Arena. To date, MiniMax’s video model has enabled creators to produce over 590 million videos.
-
In October 2025, MiniMax released and open-sourced its next-generation text large model MiniMax M2, ranking among the top five globally and first among open-source models in Artificial Analysis—the first time a Chinese open-source large model entered the global top five on this leaderboard.
Clearly, its research progress is not limited to isolated breakthroughs but represents systematic advancement across four key modalities—text, video, speech, and music—forming a complete and mutually reinforcing technical matrix.
Notably, the launch and open-sourcing of the next-generation text large model MiniMax-M2 on October 27 this year attracted particular attention.
This large model offers nearly twice the reasoning speed of Claude Sonnet 4.5, with comparable performance but priced at only 8% of the latter. At AWS re:Invent 2025 held in December, Amazon announced the inclusion of MiniMax-M2 into Amazon Bedrock, its fully managed generative AI cloud service, as a representative of Chinese-developed models.
MiniMax-M2 excels not only in planning and reliably executing long-chain tool-calling tasks but also coordinates calls to Shell, browsers, Python code interpreters, and various MCP tools.
The key technology behind MiniMax M2, "Interleaved Thinking," has been widely discussed in overseas AI research communities. This approach enables the model to continuously accumulate contextual understanding within a "think-act-reflect" loop and dynamically adjust strategies based on feedback. Mimicking the real-world workflow of engineers, this significantly enhances the agent execution capability of MiniMax M2, delivering stronger planning, higher execution stability, and more reliable self-correction in complex tasks—forming its most distinctive core advantage.
During the release period of M2, it quickly rose to become the top domestic model by token usage on OpenRouter, a global model aggregation platform, and ranked third globally in token usage for programming scenarios.
Massive R&D Investment, Pursuit of Extreme Efficiency
Like many AI large model companies, MiniMax remains in a phase of sustained high R&D investment, rapidly iterating its large models without yet achieving profitability.
In the first nine months of 2025, the company reported an adjusted net loss of $186 million (non-IFRS metric). Despite revenue growing over 170% year-on-year, the adjusted net loss increased only slightly by 8.6%. The primary reasons for the loss include R&D expenditures, fair value changes in financial liabilities, and listing-related expenses.
R&D spending reached $180 million in the first nine months of 2025, equivalent to 337.4% of total revenue. The R&D expense ratio has decreased from over 2000% in 2023 to 337.4%. This shift reflects both operating leverage from rapidly scaling revenues and continued strategic investment in model iteration and infrastructure.
Notably, despite massive R&D investment, MiniMax demonstrates exceptionally high organizational and commercial efficiency.
The company employs approximately 385 people, with around 300 in R&D—nearly 80% of the workforce. Core team members come from top-tier global tech firms such as Microsoft, Google, Meta, and Alibaba.
Its R&D and product teams are highly youthful, predominantly composed of post-1995 generations, with a significant number of post-2000 engineers. This structure aligns closely with contemporary AI development paradigms—natural fluency with new tools, automated workflows, and agent-based, AI-native working styles—significantly amplifying per-capita output.
In terms of R&D organization, administrative hierarchy under the CEO consists of no more than three layers, with project-driven, minimalist management drastically shortening implementation paths. This allows the company to maintain an extremely high cost-effectiveness ratio even in the capital- and compute-intensive race.
From inception to September 2025, MiniMax cumulatively spent $500 million, achieving:
-
A leapfrog evolution from leadership in single modality to full-modal leadership.
-
Top-tier international standing across three modalities, each with breakout products.
-
Breakthroughs every year, consistently ascending to new levels.
The loss in the first nine months of this year ($186 million) was less than the quarterly marketing spend of some internet giants on their large models.
This efficiency edge is also evident in recent financial trends.
In the first nine months of 2025, R&D expenses grew only about 30% year-on-year, while revenue surged 174.7%. Meanwhile, sales and marketing expenses declined by 26%—an uncommon pattern for an AI company still pursuing aggressive growth.
This "scissor gap" indicates that growth is not driven simply by "burning money for scale," but rather by improvements in model capabilities, organic product reputation, and unleashed organizational efficiency—not reliant on heavy advertising or market subsidies.
Steered by the Post-1995 Generation
Judging from information disclosed in the prospectus, MiniMax’s board composition strongly reflects its corporate characteristics of being "young and deeply technology-driven."
The average age of its four executive directors is just 32 (all post-1995), an extreme rarity in the history of Hong Kong-listed companies.
They are deeply involved in frontline R&D and business execution, serving directly as responsible leaders, including:
Founder & CEO Yan Junjie (36 years old), COO Yun Yeyi (31), Head of Large Language Model Research and Engineering Zhao Pengyu (29), and Head of Vision Model Research and Engineering Zhou Yucong (32).
Non-executive and independent non-executive directors play more governance, oversight, and institutional check-and-balance roles.
This board structure is highly aligned with a young AI company still in a phase of rapid technological evolution, led by founders and engineering teams, and consistent with its overall strategy emphasizing long-term technical investment and high organizational efficiency.
MiniMax stated in its prospectus that proceeds from this listing will be used primarily for the following purposes:
-
Approximately 70% for R&D over the next five years, including large model upgrades and AI-native product development;
-
Approximately 30% for working capital and general corporate purposes.
The company said it will continue advancing its vision of "continuously enhancing intelligence and making it accessible to everyone," improving societal productivity and personal quality of life through technological inclusivity.
The Story of AGI Should Be Young
MiniMax’s push for a Hong Kong listing signifies more than just another AI unicorn going public. It is telling a younger story about AGI.
Within this young team, a group of people are using AI-native thinking to build entirely new kinds of ventures—making their story bold enough to firmly bet on full-modality and globalization. This all-or-nothing sprint inevitably brings both challenges and opportunities, but may well represent the right path toward AGI.
As reflected in MiniMax’s prospectus performance, with continuous advancements in technology and business, this company will keep breaking new ground, fulfilling its long-term vision: Intelligence with Everyone.
Join TechFlow official community to stay tuned
Telegram:https://t.me/TechFlowDaily
X (Twitter):https://x.com/TechFlowPost
X (Twitter) EN:https://x.com/BlockFlow_News
@Synced_Global
