Moonshot’s Reported $3 Billion Hong Kong IPO Could Test China’s AI Boom

Moonshot’s reported confidential Hong Kong IPO filing is more than a fundraising event. It is an early public-market test of whether a Chinese foundation-model developer can convert model momentum, substantial private financing, and demand for open-weight AI into a durable commercial business.

Reuters reported that the Beijing-based developer of the Kimi large-language-model family has filed confidentially for a Hong Kong listing and is aiming to raise about $3 billion. The company is reportedly valued at around $50 billion in an ongoing funding round, but the proposed offering’s financial details and timing are still subject to regulatory approval and market conditions. 1

The reported deal: $3 billion at a $50 billion private valuation

At roughly $50 billion, Moonshot’s reported valuation would place it among China’s most valuable AI startups. Reuters reported that it is below the approximately $74 billion valuation attributed to DeepSeek, as well as the roughly $66 billion market capitalization of Z.AI.

That comparison provides context, not a verdict on relative quality. A private funding-round valuation and a public-company market capitalization are determined in different settings, and Moonshot’s eventual IPO price, if an offering proceeds, may differ from its current private-market valuation.

Moonshot was founded in 2023 by Yang Zhilin, an AI researcher who pursued doctoral studies at Carnegie Mellon University. Reuters reported that the company has raised more than $5.5 billion in total, including more than $2 billion in May, with Alibaba, Tencent, HongShan and Meituan among its reported backers. Goldman Sachs, China International Capital Corp. and Deutsche Bank are reported to be advising on the proposed offering. 1

Why Kimi K3 is central to the IPO story

The company’s investment case is closely tied to Kimi K3, its latest model. Moonshot said the 2.8-trillion-parameter system is the world’s largest open-weight AI model and said its performance approaches Anthropic’s frontier Fable model.

Its size and reception illustrate both sides of the AI business. Strong interest in Kimi K3 reportedly strained Moonshot’s compute capacity enough that the company temporarily paused new subscriptions. That is evidence of demand, but it also highlights the infrastructure challenge: serving a very large model requires access to substantial, expensive computing resources.

For potential public investors, the key question is therefore not only whether Kimi K3 attracts attention. It is whether Moonshot can reliably serve users and translate usage into recurring revenue while keeping infrastructure costs under control.

U.S. cloud talks could expand distribution—but no deal is announced

Moonshot has reportedly been discussing revenue-sharing arrangements with Microsoft, Amazon and Google that would allow Azure, Amazon Web Services and Google Cloud to host Kimi K3. Reuters reported that Moonshot was seeking up to a 30% share of revenue from K3-related services on those platforms. 2

If completed, an arrangement of that kind could give Moonshot broader distribution and give cloud customers another high-profile open-weight model to deploy. But the talks were reported as negotiations, not signed partnerships. Their commercial value, structure and regulatory viability remain uncertain.

Regulatory structure and geopolitical risk

The confidential filing is also shaped by China’s listing rules. Reuters reported that Moonshot had to unwind its offshore, or “red-chip,” incorporation structure and move to an onshore Chinese domicile before filing for the Hong Kong IPO. 1 That development aligns with wider scrutiny of some mainland companies using offshore structures for Hong Kong listings.

Moonshot also faces geopolitical and compliance questions. A senior U.S. technology official alleged that the company acquired advanced Nvidia chips and distilled capabilities from Anthropic’s Fable model in developing K3. Moonshot has disputed the distillation claim.

These are allegations rather than established findings. Still, they could matter to prospective investors and to any overseas cloud arrangements, where technology-export rules, partner diligence and regulatory reviews may affect what is practical.

Why Moonshot matters for Hong Kong’s IPO market

Moonshot’s proposed listing arrives as Hong Kong’s technology IPO market has revived. Hong Kong raised about $37.22 billion from 115 listings in 2025, according to LSEG data cited by Reuters, and several newly listed Chinese technology companies rose on their January 2026 trading debuts.

A successful Moonshot offering would show that investors are prepared to fund a Chinese AI developer at frontier-model scale in the public market. A weaker outcome would be equally informative: it would indicate that model visibility and private valuations are not, by themselves, enough to overcome questions about monetization, compute supply, competition and cross-border risk.

What to watch next

Moonshot’s IPO remains a reported, confidential filing rather than a completed transaction. The most consequential developments will be whether it obtains approval and publishes final offering terms; whether Kimi K3 demand can be served without recurring capacity constraints; whether the reported cloud talks produce agreements; and how allegations and technology restrictions affect its operations.

The core tension is straightforward: Moonshot appears to have a prominent model, deep financing and significant market interest, but a public listing would require it to demonstrate that those advantages can support scalable, compliant and economically sustainable growth. 12