Summary of Key Points
The recording of Deepseek's founder, Leung Man-fung's fundraising event, has accidentally leaked, revealing several crucial pieces of information that have caused a stir in the AI community: the company plans to reduce prices (which was met with cheers from the employees), aims to recoup its costs within 10 months, and does not focus on maximizing profits but instead seeks to benefit users. These approaches are completely contrary to the usual tactics of larger companies like Baidu and OpenAI. In short, this AI company is not aiming to follow the traditional path of "burning money to monopolize the market and then raising prices to profit," but rather aims to establish a foothold through low prices and high volume sales.
Detailed Analysis
1. Why did employees cheer at the price cut? It's not about a pay cut, but seeing a chance for survival
Many might assume that a price cut indicates financial trouble, but the employees' enthusiasm suggests they see it as a positive move. The reason is simple: the AI industry is highly competitive, with large companies spending heavily on subsidies to attract users. If small companies do not lower prices, they will have no customers. A price cut means the company can attract more users, which in turn generates more data and leads to better models, attracting even more users, creating a virtuous cycle. What the employees see is that the company can survive and even grow, rather than worrying about reduced earnings (after all, the price cut is for the benefit of users, not the employees).
For example, if you run a small noodle shop surrounded by large restaurants, reducing the price from 20 yuan to 15 yuan might result in fewer profits per bowl, but more customers, potentially leading to higher overall earnings and job stability for the staff.
2. Recouping costs in 10 months: Is it realistic or just hype?
Recovering costs in 10 months is very fast in the AI industry (typically, it takes 1-2 years or longer). This is based on two factors:
- Low marginal costs: The main cost of AI models is initial training (such as purchasing servers and hiring engineers); once the models are trained, providing services is relatively inexpensive. If a price cut brings in many users, the initial investment can be quickly recouped.
- Effective cost control: Deepseek may have optimized its model development, using more efficient algorithms to reduce server costs or maintaining a smaller team size. Leung Man-fung's confidence in achieving this goal indicates that they have done the math: once they reach a certain number of users, their costs will be covered.
However, there are risks: if user numbers don't meet expectations or larger companies enter the price war, recouping costs could become difficult. But for now, they are optimistic about this.
3. Benefiting users: A strategy that contrasts with larger companies' approaches
Large companies typically offer subsidies (free trials, discounted offers) to make users dependent on their services before gradually raising prices (e.g., OpenAI's GPT-4 and Baidu Wenxin Yiyuan). Deepseek, on the other hand, focuses on providing good value to users without maximizing profits. This is a smart move for small companies, as they cannot afford to burn money like larger ones. By offering lower prices and higher quality, they can compete on cost-effectiveness. For example, if a large brand sells a phone for 5000 yuan while a smaller brand offers the same features for 3000 yuan, consumers might choose the cheaper option. Deepseek aims to be the "cost-effective" player in the AI industry.
4. The AI community's reaction: A transparent strategy challenges industry norms
The leak caused a stir because few companies openly discuss their plans to reduce prices or recoup costs quickly. Large companies usually keep such information confidential. By being so upfront, Deepseek is telling others, "This is how we operate; you can either follow or not."
This could have two outcomes: other small AI companies might adopt this strategy (since it's a viable model for survival), while larger companies might become cautious, fearing that Deepseek could attract users and force them to lower their prices. The AI community is debating whether this approach will succeed and whether it will change the industry landscape.
5. The potential for this strategy
While Deepseek's approach has its opportunities—such as quickly gaining a user base and potentially forcing larger companies to lower prices—it also faces risks:
- Sustainability with low profits: If user numbers don't meet expectations or operating costs rise, low margins could lead to losses.
- Counteraction from large companies: Larger companies might use their resources to compete on price, making it difficult for small companies to withstand.
- User acceptance: Although lower prices are appealing, users are more concerned about the quality of AI models. If Deepseek's models are not as good as those of larger companies, customers may still prefer the established brands.
In summary, Deepseek's strategy is a unique attempt by a small company to survive in a highly competitive market. It might work if they can quickly build a user base and influence industry trends, but it also carries risks. Ultimately, whether this approach succeeds depends on user acceptance and how larger companies respond.