Moonshot AI has reportedly filed confidentially for a Hong Kong initial public offering (IPO) that could raise US$3 billion, a striking figure for a company quietly founded three years ago. This matters for Hong Kong’s financial hub in more ways than one.
The Beijing start-up behind the Kimi large language model (LLM) was valued at US$50 billion in an ongoing funding round. Its planned listing follows the January debuts of model developers Zhipu AI and MiniMax. What initially looked like a cluster is beginning to resemble a market.
China’s AI companies already have the models, engineers and a vast home market. What many still need is capital able to keep pace with their ambitions. Hong Kong is increasingly filling that role, bringing in international investors and giving young technology firms a clearer route to the public market.
That capital access matters because frontier AI becomes expensive long before it earns steady revenue. New models require advanced chips, data-center capacity and highly paid engineers. Venture funding can support the first releases. Building a lasting platform eventually calls for deeper and more dependable financing.
Moonshot’s needs are already growing. Its Kimi K3 model has 2.8 trillion parameters and has drawn enough demand to strain the company’s computing capacity. Moonshot has raised more than US$5.5 billion from investors including Alibaba, Tencent, Meituan and China Mobile.
Hong Kong began laying the ground for this wave several years ago. Chapter 18C, introduced in 2023, gave specialist technology companies a way to list before meeting the usual profit or revenue requirements. That fits an industry where a model may win users long before the company behind it earns steady income.
The route is open to AI labs, chip designers, robotics firms and autonomous-driving companies. More recent reforms brought early guidance and confidential filings into the process. Young firms can therefore work with the exchange while keeping sensitive technology out of public view until they are ready.
Six companies raised US$2.5 billion through Chapter 18C in the first quarter of 2026, more than twice what they raised under the chapter in 2024 and 2025 combined. All six finished their first trading day above their offer price.
Zhipu and MiniMax develop LLMs. Biren Technology and Iluvatar CoreX design AI chips. Other issuers work on industrial robots, autonomous driving, digital twins and drug discovery.
Together, companies across the AI value chain raised US$4.9 billion in Hong Kong during December and January. Investors can now buy direct exposure to Chinese models, computing infrastructure and commercial applications rather than relying on shares in larger technology groups as a rough substitute.
A market becomes more useful when companies from different layers list together. Model developers depend on chips and computing capacity, while application firms show where AI is finding paying customers.
Money is now arriving from several directions. Four January listings involving Biren, Zhipu, Iluvatar and MiniMax raised US$2.5 billion, with cornerstone investors committing about US$1.3 billion. Institutions from the mainland, the United States, Singapore, South Korea, Switzerland and the United Arab Emirates participated.
That mix is where Hong Kong earns its capital hub role. Mainland companies can raise international capital under familiar disclosure and trading rules, while eligible shares can later reach mainland investors through Stock Connect.
The city is now setting valuations that travel in both directions. MiniMax had surged about 300% from its January debut by early June. Both MiniMax and Zhipu then began preparing for mainland share sales, reversing the older pattern of listing domestically first.
For firms following this route, Hong Kong sets the first public price. International institutions assess the technology, revenue prospects and management. That gives a company a benchmark before it approaches investors in Shanghai or Shenzhen.
Moonshot would raise the scale again. Its reported US$3 billion target exceeds the combined proceeds of several early AI listings. A successful offering would show that Hong Kong can finance established technology groups and the newer laboratories building China’s foundation models.
The wider market has already recovered much of its momentum. Hong Kong listings had raised US$41.2 billion by mid-August, up 142% from a year earlier, with Chinese technology companies providing much of the activity.
Yet this wave’s value will be measured after the opening bell. AI companies need to turn fresh capital into computing power, stronger models and customers prepared to pay for them. Public markets, in return, demand regular evidence that research spending is creating a business.
The early signs are encouraging. Zhipu’s first-half revenue rose 400% to 953.9 million yuan, while MiniMax recorded a 283% increase to US$116.6 million. These are young businesses, but public reporting makes their progress easier to judge.
Indeed, going public changes the conversation. A company has to show investors where the money goes and whether demand is turning into revenue. Hong Kong can provide that capital while giving China’s AI sector a clearer measure of commercial progress.
The city should keep widening that ecosystem. Research-heavy firms need patient investors, credible disclosure and a liquid secondary market. A planned mainland exchange-traded fund tracking the Tech 100 could give domestic investors another route into Hong Kong-listed technology companies.
Shanghai and Shenzhen will remain central to financing Chinese innovation. But Hong Kong adds a different kind of financial strength. It can give a mainland AI company an internationally recognized price, a diverse shareholder base and a financial language understood across global markets.
Moonshot’s filing therefore carries meaning beyond one large fundraising round. If the offering proceeds, another of China’s leading model builders will have chosen Hong Kong as the place to finance its next stage and prove its commercial case on the global stage.
China’s AI race is being run in laboratories reaching from Beijing to Hangzhou to Shenzhen. Yet, increasingly, its companies are being valued in Hong Kong. The city’s opportunity lies in turning that flow of listings into a lasting capital base for the technology China is building next.
