China’s big-model firms shift from private rounds to public pricing as Hong Kong IPO race heats up

China’s leading large‑language model developers are changing the game. Moon’s Dark Side (月之暗面) — long said to have “strong primary market financing” — has reportedly opened talks with CICC and Goldman Sachs about a Hong Kong IPO. At the same time, LeapStar (阶跃星辰) is accelerating new financing. It has been reported that this turn toward the secondary market is the story, not merely another fundraising round.

Market signal: public valuations create a new anchor

Zhipu (智谱) and MiniMax (MiniMax) listed in Hong Kong in January and quickly gave investors a tradable valuation benchmark: both reached market caps around HK$300 billion and saw share prices jump three to fivefold from IPO levels. That market signal, analysts say, is what is lifting private market expectations. It has been reported that Moon’s Dark Side has run a C+++ round seeking about $1bn at a pre‑money value near $170–180bn; investors are reportedly using Anthropic’s public market valuation as a distant comparator for future targets. The key commercial evidence is also emerging: revenue inflection tied to K2.5 and agent‑style products, token consumption surges, and fast‑growing API and overseas receipts suggest monetisation is moving from theory to measurable cash flow.

Why this matters — and what’s next

Why does a listing matter if cash on the balance sheet is already ample? Because public markets convert private paper into liquid, repeatable pricing that becomes a reference for competitors, talent markets, fundraising and M&A. Once one firm locks in a stable public multiple, others are measured against it — a “lock‑in” that can amplify advantages. Geopolitics complicates the picture: export controls on advanced chips and US‑China tech tensions remain tail risks for scaling compute, even as capital markets in Hong Kong become the new battleground for nominal valuation and access to global investors. Who gets to be the pricing anchor will shape China’s AI industry for years.