Summary of Key Points
DeepSeek, which has always been low-key, has recently made a dramatic expansion: on one hand, it has launched a large-scale recruitment campaign covering 7 major categories and 33 positions (ranging from algorithms to functional departments, with opportunities in Beijing, Hangzhou, and for interns), aiming to double the size of all its departments; on the other hand, there are reports that it has just secured a $51 billion Series A financing round (with a post-investment valuation of $400 billion), led by investors such as Tencent and CATL. Behind this move is DeepSeek's strategic shift from focusing solely on developing fundamental AI models to expanding into areas such as intelligent agent applications, specialized industries (e.g., small languages/medicine/law), and infrastructure (data centers). At the same time, the company is also facing the pressure of transitioning from a laboratory-based approach to a commercial one.
Detailed Analysis
1. Recruitment: Moving from Selective Hiring to Systematic Expansion
DeepSeek used to be very selective in its recruitment, only hiring core research personnel on an individual basis. This time, however, it aims to double the size of all departments by opening positions across seven major categories (algorithms, R&D, operations, product development, data engineers, and functional roles), with opportunities available in both Beijing and Hangzhou, including for interns. The company has assured new hires that they will be involved in the most critical tasks. For example, the Agent Harness team, which is working on code-based intelligent agents (similar to Anthropic's Claude Code), is still hiring, with the head of the team stating that interviews are happening daily. This indicates that DeepSeek is no longer a small-scale research group; it is preparing to build a comprehensive organizational structure for significant growth.
2. Strategic Change: Focusing on Application Implementation Rather than Model Capability
While other AI companies have been competing on the quality of their underlying models, DeepSeek has shifted its focus to practical applications:
- Intelligent Agent Products: The company is actively recruiting for its code-based intelligent agent team, with the goal of creating its own "DeepSeek Code Harness."
- Specialized Fields: It is hiring data product managers for small languages and medicine/law to target specific industries.
- User Experience: It is looking for AI product operations specialists to improve the user experience of its products. The reason for this shift is that the capabilities of AI models are likely to become increasingly similar in the future, so differentiation will come from the implementation and specialization of applications. For instance, the needs of doctors using AI differ greatly from those of programmers, and companies that can excel in these areas will have a competitive advantage.
3. Building an "AI Fuel Station": Data Centers as a Critical Foundation for Expansion
DeepSeek is also recruiting for its IDC (Internet Data Center) team, with locations including Beijing, Hangzhou, and Ulanqab in Inner Mongolia. The choice of Ulanqab is strategic: the region offers affordable electricity and a cold climate, which reduces energy costs for data centers. DeepSeek emphasizes that this is not just a standard data center but a vital infrastructure for its AI intelligent agents. Just as cars rely on gas stations to operate, AI agents need data centers to provide the necessary computing power for their functions.
4. Financing: More Funds, But Greater Pressure to Generate Revenue
Before this recruitment round, DeepSeek was reported to have raised $51 billion in Series A financing, with a post-investment valuation of $400 billion. Investors include Tencent ($10 billion), CATL ($50 billion), NetEase, JD.com, and others. The benefits of this funding are clear:
- It allows the company to hire top talent at high costs in the competitive AI industry.
- Access to resources from industrial investors like Tencent can help accelerate the implementation of its technology (e.g., Tencent's ecosystem can support product development).
However, there are also challenges: as a former laboratory-based company, DeepSeek no longer has the luxury of focusing solely on research. With this capital, it must demonstrate its ability to generate revenue through expansion, data center construction, and product development.
5. Recruiting Innovative Talents for AGI
A special position is designed to recruit "cross-disciplinary AI talents" with extraordinary abilities and a passion for participating in the development of AGI (Artificial General Intelligence). The desired qualifications include a non-traditional background, expertise in a specific field, or entrepreneurial experience. DeepSeek aims to break conventions by bringing innovative ideas from diverse backgrounds to its AI research, potentially hiring artists or engineers with unique perspectives.
Conclusion
DeepSeek's expansion is a result of the funding round, as it focuses on building teams and infrastructure while shifting from a model-focused approach to one that emphasizes application development. Whether it can succeed in transforming its technology into tangible commercial value remains to be seen.