Summary in Plain Language
DeepSeek, which became an overnight sensation in the past six months for its ability to create world-class large models at low costs, has moved at an unprecedented pace: it only raised funds for the first time in April, and in less than half a year, it has decided to go public on the STAR Market, with its valuation soaring to 500 billion yuan. Many people assumed its eagerness to go public was due to a lack of funds in the AI industry. However, the truth is quite the opposite. Currently, there is a long queue of investors eager to invest in DeepSeek, ranging from internet giants and local state-owned enterprises to wealthy family offices, competing for shares. The timing of its public offering is due to three unavoidable new challenges:
First, the "geek model" it relied on—using a small team and algorithm optimization to cut costs—can no longer sustain the long-term capital-intensive nature of the AI industry.
Second, the capital in the primary market has created a complex web of nested equity structures, so complex that even the founders find it difficult to understand, posing risks to compliance and control.
Third, DeepSeek's once unique "cost-effectiveness" advantage is being eroded by a general price-cutting trend in the industry. Now is the best time to capitalize on its valuation of 50 billion yuan; going public will help to clarify the technical advancements made behind closed doors and establish a foundation for long-term growth by applying public regulations.
---
Detailed Explanation
1. The AI Miracle Created by Cost Savings Can No Longer Sustain a Long-Term Battle
DeepSeek was once seen as an outlier in the AI community: while others spent billions on computing power, it managed to develop a model on par with the world's top performers with just a 30-person team and through algorithmic optimizations, without any significant external funding. This seemed to represent a low-cost path for China to catch up in AI. However, this cost-saving strategy is unsustainable. Initially, the spending was one-time; once the model was trained, it was done. But with the increasing number of users, every interaction involving AI (chats, code writing, etc.) consumes computing resources, leading to continuous financial outflows. In the first seven months of this year, DeepSeek spent 11 billion yuan on servers and chips, while its total revenue was only 475 million yuan, resulting in a loss of over 700 million yuan. It also plans to hire 150 more core technical staff, doubling its team size and expanding into new businesses like AI services, which will further increase its expenses. The parent company, Huanfang, could previously support these costs, but with such ongoing investments, no shareholder can sustain them in the long term, necessitating access to massive, long-term funds from the public market.
2. The Complex Web of Equity Created by Competitive Capital
To prevent external interference in its technology and control, DeepSeek designed a rule during its initial fundraising: all external investors could not directly hold shares but had to invest through a partnership controlled by founder Liang Wenfeng, with their shares locked for five years. This seemed like a win-win solution. However, with so many investors competing for a limited number of shares, some intermediaries started creating multiple layers of investment structures. These intermediaries charged fees and took a portion of the profits. The resulting complexity makes it impossible for even the founders to track who is actually investing and whether any illegal funds are involved. If one of these intermediaries causes issues, it could affect the company's control and compliance. Clarifying this situation through public regulations after going public would be the most cost-effective solution.
3. The 500-Billion Yuan Valuation Is Not a Bubble but Represents the Best Current Opportunity
Some argue that DeepSeek's valuation of 50 billion yuan is inflated, given its annual revenue of just a few hundred million yuan. However, this valuation reflects its scarcity: it is the only Chinese company that has reached the global top tier of large models without spending hundreds of billions. This advantage is unique in the current AI market. However, it is weakening as DeepSeek's subsequent models have failed to replicate the initial success, and the entire industry is cutting prices. OpenAI has reduced the price of its API by 80%, and other cloud providers have also lowered their model prices. DeepSeek's former advantage of "near-top performance at a fraction of the cost" is no longer unique. If it waits until this advantage disappears before going public, it may not be able to fetch such a high valuation. Going public now, when the market still recognizes its technological leadership, is the most beneficial for both the company and its investors.
4. Rules Will Change Completely After Going Public
Going public is not just a means of raising funds for DeepSeek; it marks a transition from a tech-driven startup to a public company. It will have to abandon its previous approach and follow public regulations. No longer can it focus solely on developing groundbreaking models without considering profitability or explaining its spending. It will need to disclose all its financial information regularly, including how much computing power it uses, how many employees it plans to hire, its losses, and its revenue growth. Its valuation will no longer depend on its next model's success but on its ability to generate revenue from selling its models to businesses and individuals, gradually reducing its losses.
5. For Ordinary Investors, This Is a Chance to Invest in Core AI Technologies—but Be Cautious
Previously, only top-tier investors could invest in DeepSeek. Now, ordinary investors have a chance to participate in the growth of this leading AI company. However, the 50-billion yuan valuation already includes potential future profits, so if its model iterations fail to meet expectations, the valuation could fall. Moreover, since all AI companies are still in a competitive phase, high growth comes with high risks. Investors should be cautious and not overreact to the hype.