Moonshot Is Chasing a $50B Pre-IPO Round and a Hong Kong Listing

Fresh off Kimi K3, China's Moonshot AI reportedly plans an August pre-IPO round at up to $50 billion, well above its current $31.5 billion mark, with a Hong Kong listing possible within months. Goldman and CICC are in talks to underwrite.

Moonshot Is Chasing a $50B Pre-IPO Round and a Hong Kong Listing

Moonshot AI, the Beijing lab behind the Kimi models, reportedly plans to open talks in August for a final pre-IPO round targeting a valuation of up to $50 billion, after a current round that values it at about $31.5 billion, per TechNode. A Hong Kong listing is possible within months, with Goldman Sachs and CICC in talks to underwrite. The sprint follows Kimi K3, the open-weight model we covered that benchmarked near the top US systems.

The read is a near-doubling of the mark, from $31.5 billion to $50 billion, across one model release and one funding round. That is momentum pricing, and it rests on an open-weight strategy that gives the flagship model's weights away. So $50 billion prices something other than API margin: distribution, the enterprise and cloud pull that open weights create, and Moonshot's status as China's designated frontier champion, a status that brings state-linked investors to the table.

The Hong Kong venue matters. A Chinese frontier lab listing in Hong Kong rather than New York reflects both the geopolitics of US-China tech and access to mainland and regional capital. It also sidesteps the US disclosure and access regime, including the kind of government sign-off Washington is now building. For the global picture, Moonshot chasing a listing within months while Anthropic eyes October puts the AI-IPO wave on two continents and two regulatory systems at once. The question the listing has to answer is how an open-weight model compounds into the recurring revenue public markets demand.

A jump toward $50 billion prices Moonshot as China's frontier champion, not as an API business. Watch whether an open-weight model can carry a public listing, because no one has proven that it can yet.