DeepSeek launches first external financing: valuation exceeds US$10 billion, plans to raise US$300 million

One sentence conclusion

DeepSeek has launched its first external financing with a valuation of over US$10 billion and plans to raise US$300 million, marking the Chinese AI star company's shift from technology research and development to market expansion. For AI developers, this means that the commercialization of open source models will accelerate, and the enterprise-level AI solution market will usher in new players, but the fundamentals of the open source model will not change in the short term.

What happened

According to foreign media The Information, China’s large model unicorn DeepSeek is launching its first external equity financing since its establishment. According to multiple people familiar with the matter, DeepSeek’s financing target valuation is no less than US$10 billion (approximately 68 billion yuan), and plans to raise no less than US$300 million (approximately 2.05 billion yuan).

This news caused shock in the investment circle. A person from a large state-owned equity institution said: "There is feedback from channels that the news that DeepSeek is launching its first external financing is very likely to be true, and it is completely impossible to invest at present." As of now, DeepSeek has not officially responded, but many people close to the transaction revealed that it is "extremely possible."

DeepSeek had previously relied entirely on the exclusive support of its parent company Huanfang Quantitative. Founder Liang Wenfeng had repeatedly stated that he "refuses financing" and insisted on a self-supply model. This financing marks the technology-driven company’s official entry into the global AI capital race.

Direct impact on practitioners

Cost level

For developers using the DeepSeek open source model, there will be no direct cost changes in the short term. DeepSeek-V3 series models will continue to remain open source and free. But in the long term, commercialization pressure may prompt DeepSeek to launch more paid enterprise-level services.

Efficiency/capability level

Financing will accelerate the research and development progress of DeepSeek-V4. According to the previously announced technology roadmap, DeepSeek-V4 will completely run on Huawei's Ascend 950PR chip, and the technical architecture will be fully shifted from CUDA to CANN. This means:

• Domestic computing power ecology will be substantially promoted
• Enterprise-level deployment solutions will be more complete
• API service stability and response speed may be improved

Compliance/Risk Level

After introducing external capital, DeepSeek may face more commercialization pressure and need to balance the interests of the open source community and commercial returns. However, considering its technical genes and founder’s background, the possibility of significantly changing its open source strategy in the short term is low.

Core data comparison

Our judgment

**This thing is underrated. **

The market generally focuses on a valuation of US$10 billion and a financing amount of US$300 million, but ignores three deeper impacts:

1. A turning point in the domestic AI ecosystem: DeepSeek chose Huawei Ascend as the computing power base of V4. Together with this financing, it will substantially promote the ecological construction of domestic AI chips and frameworks. For developers who rely on the NVIDIA ecosystem, now is the time to focus on domestic alternatives.

2. A new paradigm for open source commercialization: DeepSeek proves that open source models can also achieve high valuations. This provides a financing template for other open source AI projects - technology leadership + ecological construction + domestic alternative narrative.

3. Reshuffling of the enterprise market: Currently, China’s enterprise AI market is dominated by a few manufacturers. With its technical advantages and open source ecosystem, DeepSeek has the opportunity to take a share of the enterprise market. For business users, this means more choices and better bargaining power.

Which type of people have the greatest impact? Enterprise-level AI solution provider and Domestic computing power ecosystem participant. Which type of people are least affected? Individual developers and small startups as the core open source model is expected to remain free.

3 things you can do now

1. Assess the cost of migrating domestic computing power: If your project relies heavily on the NVIDIA ecosystem, you can now start to evaluate the cost and feasibility of migrating to Huawei Ascend or other domestic chips. DeepSeek-V4’s technology roadmap provides practical case references.

2. Follow DeepSeek enterprise-level products: Register a DeepSeek developer account and follow its official updates. Once an enterprise-level API or deployment solution is launched, test and evaluate it as soon as possible to seize market opportunities.

3. Re-evaluate the commercialization path of open source models: If you are operating a service based on open source AI models, study DeepSeek’s financing cases and commercialization strategies. This provides new ideas for the sustainable development of open source projects.

Related tools (jump within the site)

This article involves tools: DeepSeek, Huawei Ascend, and open source large models