Kimi Spurgeon re-targeting research board: behind the double listing of the dark side of the moon, China's AI competition enters the capital extension phase

China’s big model competition is shifting from “which model is smarter” to “who has capital to live longer.” Reuters quoted the South China Morning News on September 10 as saying that the dark side of Beijing’s artificial intelligence company is exploring the possibility of a double listing between Hong Kong and Shanghai: the company has discussed with investors a Hong Kong IPO and is considering subsequently landing on the Shanghai research board. Reuters has previously that the dark side of the month has been secretly filed for listing in Hong Kong, with a target fundraising of about $3 billion. For an AI start-up that is still in the phase of model rapid iteration and commercialization exploration, the double listing is not just capital market news, it shows that China’s big model industry has entered a costly “race.”

AI entrepreneurship is returning from technology romanticism to the reality of cash flow. From 2023 to 2024, China’s big model entrepreneurship trend is rapidly heating up with venture capital, Internet giants and founders’ visions; by 2026, computing, talent, reasoning costs, market placement and model training will require continued huge investments.

Double listing, is a financing channel and also a valuation firewall

Kimi, who developed the Dark Side of the Moon, has quickly gained market awareness with its long-text capabilities, and the latest Kimi K3 has also gained strong attention after its release in July this year. Reuters said the company is also discussing revenue sharing arrangements with Microsoft, Amazon and Google to let the U.S. cloud platform host its models. If the relevant partnership landed, the dark side of the moon will no longer be just a domestic AI application company in China, but will seek a position in the global model distribution market.

For AI companies, the listing has long been more than just the story of the founders.Who can get lower cost, longer capital, is more likely to withstand the next round of model iteration and price warfare.

After choosing Hong Kong and considering the research board, it also reflects the real calculation of Chinese AI companies. Hong Kong provides international funds, dollar capital and more open investor structures, but this year several AI concept shares are listed, making the fund competition more intense; the research board has stronger domestic science and technology policy attributes and the RMB fund pool. If both countries can be opened, the enterprise is equivalent to obtaining an international capital window and a domestic policy capital window at the same time.

This arrangement is in line with China's current science and technology strategy.The United States continues to restrict the flow of advanced AI chips and some key technologies to China, while Beijing sees artificial intelligence as a strategic industry, encouraging the development of domestic models, computing and application ecology.When external technology restrictions and domestic industry support are strengthened, capital markets naturally become part of national science and technology competition.

It’s really expensive not to “make a chat robot”, but to maintain cutting-edge modeling capabilities.

There is a reality in the big model industry that is easily overlooked by ordinary investors: user growth is not equal to profit growth. The stronger the model, the training and reasoning often require more computing; when the industry competes with users through free quotas, low-cost APIs and corporate discounts, commercial income growth may not catch the cost. The Chinese market also has a lot of competitors such as bean bags, DeepSeek, Intelligence, MiniMax, AliExpress, price and talent competition are extremely fierce.

Thus, the capital pressure for the future of the dark side of the moon comes from at least four aspects:

• Training costs: Advanced model training requires continuous procurement of computing power, data and infrastructure;
• Reasoning costs: The larger the user size, the bigger the online service itself;
• Talent costs: Top AI researchers are competing for high pay globally;
• Commercialization time gaps: Technical influence and real profitability tend to be many years apart.

That’s why “marketing” is more like a supply station in this round of AI industries than an endpoint. Capital markets will give enterprises more cash, but will also require revenue, profit rate and growth story to be checked quarterly. Private companies can interpret long-term investments with “AGI Vision” and listed companies will eventually have to translate the vision into financial statements.

Beijing's AI industry is forming a mixed model of "national strategy + market burning money"

China's science and technology policy has a distinct feature: the state is responsible for determining the strategic direction, providing the industrial environment and policy capital, and enterprises are responsible for competing at high speed in the market. This model can quickly form industrial clusters, but it is also easy to make repeat investment and valuation expansion. In the past, new energy, photovoltaic has emerged similar pathways, and today AI may become the next capital-intensive course.

The Dark Side of the Moon’s listing plans are therefore worth observing in a larger context. Intelligence, MiniMax and other companies have entered the Hong Kong market, and DeepSeek is alsoly preparing for domestic capital operations. More and more AI companies are moving from the venture capital phase to the open market, meaning that China’s AI competition is gradually transferring risks from a handful of VCs and Internet giants to secondary market investors.

This does not mean that the dark side of the moon lacks technical value. Yang Chinin’s team has a higher industry awareness in terms of context, model architecture and open source direction, and Kimi has also built a strong brand awareness. But the stronger the technology, the greater the demand for capital. Enterprises must find a balance between “making the most advanced model”, “expanding the size of users” and “building sustainable revenue”, which is exactly what global AI companies face.

Focus China believes that the dark side of the moon's double listing is most worth observing is not how high the final valuation will be, but whether it can demonstrate whether Chinese AI companies can build a business model that does not rely on unlimited funding.If capital markets are just replacing risk investments to continue to compete for high cost, then IPO can only extend the race and can not solve the basic economics of the industry.

The next phase of AI will not be determined by the ranking alone.Model capabilities determine who can enter the field, cash flow and funding capabilities determine who can stay on the field.The dark side of the moon simultaneously knocks on the door to the Hong Kong and Shanghai capital markets, which is indicating that China’s big model war has gone from the technology conference to the balance sheet.