DeepSeek’s first round financing structure exposed: Liang Wenfeng 20 billion, Tencent 10 billion, CATL 5 billion
Reuters disclosed the details of DeepSeek's first round of financing: Liang Wenfeng personally invested 20 billion yuan as the largest investor, Tencent 10 billion, CATL 5 billion, and the post-investment valuation is 350-400 billion yuan. CATL uses investment to enter the AI data center power supply market.
According to Reuters, the structure of DeepSeek's first round of financing has surfaced: founder Liang Wenfeng personally promised to invest 20 billion as the largest single investor, Tencent plans to invest 10 billion, and CATL plans to invest 5 billion. The scale of this round of financing is about 50 billion yuan, with a post-investment valuation of 350-400 billion yuan. It plans to introduce no more than 10 investors and is expected to be completed in the next few weeks.
In addition to the above three companies, the National Artificial Intelligence Industry Investment Fund, NetEase, and JD.com are in the final stages of negotiations, and IDG Capital and Monolith Capital are also among potential investors.
The entry of CATL is the most unexpected - this global power battery leader is extending into the field of AI data center power supply. In April, it invested about 4.1 billion yuan in Zhongheng Electric, the leader of HVDC in intelligent computing center (holding 49% of the shares). In May, it invested about 942 million U.S. dollars in 21Vianet (about 38.1% of the shares). In June, it planned to invest 5 billion in DeepSeek - an AI infrastructure investment chain of about 15.5 billion yuan in three months. It has exceeded the scope of financial investment and clearly points to the migration of strategic tracks. DeepSeek is recruiting data center engineers in Inner Mongolia, and its computing power expansion plan coincides with the CATL energy storage order growth curve.
Tencent is trying to catch up with ByteDance’s Doubao and Alibaba’s Tongyi Qianwen by investing in DeepSeek. Although the self-developed hybrid model continues to advance, its C-end user scale and growth rate have been significantly outstripped by its opponents.
Reviews