Moonshot AI, the Beijing company behind the Kimi models, is preparing for a Hong Kong initial public offering that could happen within six months, according to Bloomberg reporting.
The status matters and is easy to overstate. Moonshot has not filed. What has been reported is preparation: the company has distributed a shareholder resolution seeking investor approval for a potential listing, and has held discussions with Goldman Sachs and China International Capital Corporation about roles in the offering. A stated intention to list within six months is a planning assumption, and such timetables slip routinely.
The number that frames it
Alongside the listing preparation, Moonshot is completing a private funding round that could value it at more than $30 billion. Its annual recurring revenue reached $300 million in June, up from $200 million in April.
Those two figures together are the story. A $30 billion valuation against $300 million of annualized revenue is a multiple of roughly 100 times. For comparison, mature software companies typically trade in the high single digits to low tens of times revenue, and even fast-growing enterprise software rarely sustains much beyond 20 times outside of exceptional periods.
That is not automatically irrational. Revenue growing 50 percent in two months compounds very quickly if it continues, and a buyer at 100 times current revenue is really paying perhaps 10 to 20 times what they expect revenue to be in a couple of years. The entire proposition rests on that continuation, which is precisely what a public market will be asked to price.
Why Hong Kong
The venue choice is the more interesting structural question.
A Chinese technology company listing in New York faces persistent political and regulatory risk, including audit-inspection requirements and the possibility of restrictions changing mid-course. Hong Kong offers access to international capital under Chinese sovereignty, and has actively courted technology listings, including companies that are pre-profit.
There is also a corporate-structure dimension that explains the timing. Moonshot is dismantling its red chip structure, an arrangement in which an offshore holding company owns the mainland operating business. Such structures were built to make foreign investment and overseas listing easier; unwinding one is substantial legal work, and companies do not undertake it casually. That the work is underway is stronger evidence of intent than any stated timetable.
What Moonshot actually sells
For readers who have not encountered it, Kimi is Moonshot's family of AI models and its consumer chatbot, competing in China with offerings from Alibaba, ByteDance, DeepSeek and others.
The most recent model, Kimi K3, has 2.8 trillion parameters, a measure of model size that correlates loosely with capability and directly with the cost of training and serving it. The company has acknowledged that K3 remains behind Anthropic's Claude Fable 5 and OpenAI's GPT-5.6 in overall capability, while the independent evaluator Artificial Analysis ranked it ahead of Anthropic's Opus 4.8 on frontier benchmarks.
That combination, competitive on some benchmarks and behind the leaders overall, is roughly the position of the strongest Chinese labs generally, and it is commercially relevant: a model that is close to the frontier at lower cost is a viable business even if it never leads.
The question a listing forces
Private valuations are negotiated between a small number of well-informed parties and can stay disconnected from operating results for years. A public listing replaces that with a continuous price set by anyone willing to trade, including investors who will mark the company against its reported revenue every quarter.
For an AI model company, that is an uncomfortable transition, because the cost side is brutal. Training frontier models requires enormous capital spending, and serving them costs real money per query, so revenue growth does not automatically convert into profit. Moonshot has not disclosed profitability, and nothing in the reporting suggests it is profitable.
The listing, if it happens, will therefore be a useful test of something broader than one company: whether public investors in Asia will underwrite Chinese AI labs at valuations set in private rounds, or mark them down to what the revenue currently supports.
