🧠 Food for thought

Implications, context, and why it matters.

MiniMax’s top-range IPO price leaves gaps in judging if HK$165 makes sense

• MiniMax plans to price at HK$165, the top of its range 1. Investors still lack basic metrics to judge it. It made US$53.44 million in revenue in the first nine months of 2025, which was up 175% year over year, but model training costs (teaching AI systems using large datasets) squeeze margins 2.
• The expected deal size of HK$4 to 5 billion in primary proceeds implies a valuation between US$3 billion and US$4 billion, above its March 2025 Series A valuation of US$2.5 billion 32. Details on gross margins and compute costs are scarce 2. Customer acquisition economics and total shares outstanding are also unclear 2. The tag looks rich given legal risks from Disney, NBCUniversal, and Warner Bros. Discovery lawsuits filed in September 2025 2.
• China’s required approval for outbound share issuance adds regulatory risk 2. Geopolitical export controls that restrict graphics processing unit (GPU) access could raise inference costs (running models to generate outputs) and slow scaling 2.

Hong Kong’s January 2026 AI IPO rush opens near-term sales for GPU, cloud, and MLOps vendors

• Eleven AI-related firms plan Hong Kong listings this month and seek up to US$4.1 billion in proceeds 3. That cash makes them buyers of AI infrastructure. MiniMax and Zhipu AI, a Chinese AI model developer, cleared listing hearings in mid-December 2025 3. Zhipu AI’s developer-oriented application programming interface (API) business topped CNY (Chinese yuan) 100 million in annual recurring revenue (ARR) with 2.7 million paying users 3.
• GPU suppliers and cloud infrastructure providers can sell into this wave as these firms scale training and inference workloads. Zhipu AI expects to double operating revenue in 2025 and lift its API business share to 50%, which would expand standardized services that need heavy compute 3.
• Vendors in MLOps (machine learning operations), model evaluation, and AI safety tooling should move fast on business development with this cohort of newly public AI companies. They face rising scrutiny on model governance and often speed up purchases of compliance and monitoring solutions post-IPO.