DeepSeek Raises Capital as Founder Liang Wenfeng Increases Direct Stake to 34%
Chinese AI startup DeepSeek has completed a capital increase that significantly boosts founder Liang Wenfeng's direct ownership in the company, according to a report by Shanghai Securities News.
According to the Shanghai Securities News report published on April 28, 2026, business registry data from Tianyancha shows that on April 27, Hangzhou DeepSeek Artificial Intelligence Basic Technology Research Co., Ltd. increased its registered capital from RMB 10 million to RMB 15 million (US$2.1 million).
The capital injection primarily came from founder Liang Wenfeng, whose subscribed registered capital increased by RMB 5 million from RMB 100,000 to RMB 5.1 million. This raised his direct shareholding from 1% to 34%. Meanwhile, the stake held by Ningbo Chengen Enterprise Management Consulting Partnership (Limited Partnership) decreased from 99% to 66%.
Through both direct holdings and indirect ownership via Ningbo Chengen Enterprise Management Consulting Partnership, Liang now controls approximately 84.29% of DeepSeek's total equity.
The capital increase comes shortly after DeepSeek launched its latest AI model. On April 24, 2026, the company officially released DeepSeek-V4 preview version, completing open-source development while opening access through its official website, mobile app, and API services. The new model features million-token context capabilities, marking a significant milestone in making ultra-long context processing more accessible to the broader market.
Industry analysts suggest that DeepSeek-V4's release addresses gaps in China's open-source AI landscape, particularly in million-token context handling and advanced agent capabilities. The dual-version deployment strategy is expected to lower barriers to ultra-long context technology adoption and accelerate large language model applications across code development, document processing, and complex reasoning scenarios.
The strategic capital restructuring positions DeepSeek to further capitalize on growing domestic demand for advanced AI technologies while maintaining its commitment to open-source development in China's rapidly evolving large language model industry.