DeepSeek has raised approximately RMB 100 billion (around USD 15 billion) in its latest financing round, marking a milestone in China's AI large model market. While the pre-investment valuation remains around RMB 500 billion, the structure of the investors indicates a diverse funding source, with existing shareholders continuing to participate and new capital from industries such as semiconductors and automotive, highlighting the industry's long-term confidence in the next wave of model training and infrastructure development.
The financing amount has been raised from the previously reported target of RMB 50 billion, doubling in less than half a month, reflecting optimism from various parties regarding DeepSeek's growth rate and global competitiveness. The investment amounts from both new and existing shareholders are roughly equivalent, indicating that market expectations for the next few years regarding GW-level computing power and self-built data center capabilities are on the rise; at the same time, the company's current valuation has not been inflated, suggesting that investors prioritize the stability of long-term platformization and commercialization.
DeepSeek is advancing its GW-level computing power park, which requires substantial capital for self-built data centers and long-term model iteration investments. These investments are crucial for constructing a globally competitive cloud inference platform. According to Bloomberg, this funding aims to enable DeepSeek to compete with international giants such as Anthropic and OpenAI in the global market. This is also one of the breakthrough performances in China's AI large model sector following this round of financing, setting a new record for single-round financing in the industry.
Investor Structure and the Long-Term Impact of Industrial Capital on DeepSeek
Among the list of investors, in addition to continuing participants such as Tencent, CATL, IDG Capital, LSI Capital, and Shixiang Technology, new industry investors including Stone Creek Capital, SMIC Capital, CPE Yuanfeng, Junlian Capital, and Geely have also been introduced. CATL alone plans to add approximately RMB 10 billion to its previous investment amount, indicating a growing long-term demand from the battery and new energy supply chain for automation and high-computing platforms; furthermore, the involvement of the semiconductor and automotive manufacturing sectors suggests that the entire ecosystem will form closer synergies in core platform technologies, computing power supply, and commercial implementation. This capital structure will help DeepSeek achieve financial stability in the global market and balance local deployment with global layout in the future.
According to industry analysis, DeepSeek is planning a GW-level computing power park, which requires large-scale self-built data centers and long-term investments in model training iterations. The investors' acceptance of long-term technology stacking and operational costs often determines the company's pace of global market deployment and data center establishment. This round of funding is also seen as a strategic move to bring it closer to competitors like OpenAI and Anthropic, further enhancing China's influence in the construction of advanced AI infrastructure.
In the domestic market, this single round of financing has set a new record, demonstrating the capital market's recognition of the long-term growth of large AI models and related services. For DeepSeek, the stability of funding and the entry of industrial capital can provide substantial support for its data center design, energy-saving and cooling technologies, as well as cross-regional data governance, thereby enhancing the competitiveness and visibility of the entire ecosystem. Both officials and media view this financing as an important indicator of China's AI industry in the global competitive landscape.
Based on background information, DeepSeek's next steps will focus on scaling and optimizing the energy efficiency of self-built data centers while advancing the training and commercialization of multilingual and cross-domain models. If these efforts proceed smoothly, the company is expected to establish core computing power nodes in the global market within the next few years and collaborate with several large technology companies and verticals in automotive and industrial sectors, driving the rapid development of new infrastructure in the digital economy. The diversification of investors and long-term capital investment will provide DeepSeek with significant growth momentum in terms of technological capabilities and capital structure. It is important to note that the global competitive landscape still has variables, and market demands for model safety, data governance, and compliance will become significant issues in the future.