SourceSCMP Technology•Stalecollected in 45m
MiniMax prepares for mainland China IPO
💡A major Chinese AI model firm is moving toward a mainland IPO, signaling significant capital shifts in the AI sector.
⚡ 30-Second TL;DR
What Changed
MiniMax hired Citic Securities to manage the mainland China share sale process.
Why It Matters
This listing signals a maturing AI ecosystem in China, potentially unlocking significant capital for large-scale model training. It also highlights the growing trend of AI companies seeking dual-listing status to bridge international and domestic capital markets.
What To Do Next
Monitor MiniMax's public filings for insights into their compute infrastructure and model scaling strategy as they prepare for public disclosure.
Who should care:Founders & Product Leaders
Key Points
- •MiniMax hired Citic Securities to manage the mainland China share sale process.
- •The company aims to offer yuan-denominated shares to onshore investors.
- •This move complements their existing presence and growth in the Hong Kong market.
- •MiniMax is positioning itself as a key player in the Chinese AI model landscape.
🧠 Deep Insight
Background and context from public sources — not the original article. 25 sources cited.
🔑 Enhanced Key Takeaways
- •MiniMax already completed an IPO on the Hong Kong Stock Exchange on January 9, 2026, with its shares surging by approximately 400% since its debut, valuing the company at HK$264 billion (US$33.7 billion) as of May 2026.
- •The planned mainland China IPO is widely anticipated to occur on the STAR Market, a technology-focused board under the Shanghai Stock Exchange, which would introduce core large language model (LLM) assets to a market previously dominated by AI infrastructure and computing-power firms.
- •MiniMax's annual recurring revenue (ARR) surpassed US$300 million as of May 2026, demonstrating over 100% growth in the preceding two months, and the company serves approximately 300 million global users and over one million enterprise and developer customers.
- •The company has attracted significant investment, including a $600 million round led by Alibaba Group in March 2024, valuing it at $2.5 billion, and a subsequent $300 million Series B extension in July 2025, which raised its valuation to $4 billion with participation from Shanghai state-owned capital.
- •MiniMax is recognized as one of China's "AI Tigers," with its M2 model achieving top scores on prominent leaderboards, placing it ahead of models like Google DeepMind's Gemini 2.5 Pro in some intelligence indices for open models.
🛠️ Technical Deep Dive
- MiniMax develops multimodal foundation models capable of understanding, generating, and integrating text, audio, image, video, and music.
- Key models include MiniMax M2.7, MiniMax Music 2.6, Hailuo 2.3/2.3 Fast Video, MiniMax Speech 2.8, MiniMax M2-her Text Model, ABAB 6.5 series, MiniMax-01 LLM product line (MiniMax-Text-01 and MiniMax-VL-01), and Speech-02.
- The MiniMax-M1 and M2 models utilize a Mixture-of-Experts (MoE) architecture for computational efficiency.
- MiniMax-M1 features 456 billion total parameters, with approximately 45.9 billion activated per token, and incorporates a Lightning Attention mechanism for efficient scaling and native support for 1 million token context windows, extrapolating to 4 million.
- The training cost for MiniMax-M1 was significantly lower than estimated costs for GPT-4, being nearly 200 times cheaper.
- The MiniMax-M2 model has 230 billion total parameters, but only activates about 10 billion parameters during inference, contributing to its efficiency.
- The Hailuo 02 video model, built on NCR architecture, was ranked #2 globally on Artificial Analysis benchmarks, outperforming Google's Veo 3.
- Hailuo 2.3 offers enhanced character animation and supports various stylized content, while Hailuo 2.3 Fast provides a cost-optimized variant for batch video production, reducing costs by up to 50%.
- Speech-02 is a text-to-speech model that supports over 30 languages.
🔮 Future ImplicationsAI analysis grounded in cited sources
MiniMax's dual-listing strategy will solidify its financial position and market leadership in China's AI sector.
The mainland IPO on the STAR Market will provide access to domestic capital, complementing its Hong Kong listing, and allow it to compete more effectively with other Chinese AI firms for funding and market share.
The success of MiniMax's efficient AI models could influence global AI development towards more cost-effective and resource-efficient architectures.
MiniMax's models, particularly M2, demonstrate comparable performance to cutting-edge US models at a fraction of the cost, driven by efficient architectures like MoE and Lightning Attention, which could become a benchmark for future AI development, especially given chip export restrictions.
MiniMax's expansion into the A-share market will shift the valuation logic in China's AI sector from hardware speculation to core technology pricing.
The successful listing of core LLM assets like MiniMax on the A-share market will fill a gap previously dominated by infrastructure and computing-power firms, highlighting the value of foundational AI technology.
⏳ Timeline
2021-12
MiniMax founded by former SenseTime researchers.
2022-10
Launched its first product, Glow (later Talkie/Xing Ye).
2024-03
Alibaba Group led a $600 million financing round, valuing MiniMax at $2.5 billion.
2025-07
Closed a $300 million Series B extension at a $4 billion valuation, led by Shanghai state-owned capital.
2026-01-09
Listed on the Hong Kong Stock Exchange.
2026-05-29
Submitted listing guidance filing report to Shanghai Securities Regulatory Bureau for A-share IPO.
📎 Sources (25)
Factual claims are grounded in the sources below. Forward-looking analysis is AI-generated interpretation.
📰
Weekly AI Recap
Read this week's curated digest of top AI events →
👉Related Updates
AI-curated news aggregator. All content rights belong to original publishers.
Original source: SCMP Technology ↗
This is a summary, not the original. Read the source, or get the weekly briefing.
The weekly digest
One email a week. Unsubscribe anytime.