By Jiemian News
DeepSeek is making headlines again with its latest funding news.
On May 6, media reports revealed that the National Integrated Circuit Industry Investment Fund is in talks to lead DeepSeek’s first external funding round, with the company’s post-investment valuation expected to reach around $45 billion.
According to sources cited by the Shanghai Securities News, the talks are indeed underway, but the final valuation has yet to be determined. Multiple internet giants and state-backed funds are participating in the funding negotiations, with all parties maintaining a cautious and low-key approach. The final investment list and transaction terms are still being discussed.
As early as April 18, rumors about DeepSeek’s first external funding round had been circulating in the industry. At the time, a report by the China Securities Journal mentioned that a person from a large state-owned equity institution confirmed the news, saying, “It’s currently impossible to invest, and we can’t disclose more information.”
By April 22, another report emerged, stating that Tencent and Alibaba were in talks to invest in DeepSeek, with the company seeking to raise funds at a valuation of over $20 billion.
DeepSeek’s appeal to investors lies in its unique growth path and technical capabilities. Public information shows that the company was founded in July 2023 by Fantasia Quantization, with Zhejiang University alumnus Liang Wenfeng as its founder.
In its early days, DeepSeek relied entirely on internal support from Fantasia Quantization for research and development funding, without any external equity financing. Despite its short history, the company has made a significant impact in the domestic and international AI fields with its outstanding performance in large models like DeepSeek-V3 and DeepSeek-R1. DeepSeek has become the fastest application to break through 30 million daily active users and has topped the app store rankings in 140 countries.
Last February, there were rumors about DeepSeek considering its first external funding round, but a person close to the company denied the news. At the time, it was reported that Alibaba and state-owned funds were interested in investing in DeepSeek, with Alibaba planning to invest $1 billion at a valuation of $10 billion, accounting for 10% of the company. However, Alibaba’s vice president, Yan Qiao, later denied the rumors.
Notably, before and after the latest funding news broke, DeepSeek underwent significant equity adjustments. According to business registration information, on April 27, 2026, the company’s registered capital increased from 10 million yuan to 15 million yuan, with Liang Wenfeng increasing his stake from 1% to 34% through direct investment. The original major shareholder, Ningbo Chengen Enterprise Management Consulting Partnership, saw its stake diluted to 66%. After the adjustment, Liang Wenfeng controls approximately 84.29% of the company’s shares through direct and indirect holdings, further consolidating his absolute control over the company. Sources also revealed that Liang Wenfeng plans to participate in the funding round in his personal capacity, demonstrating his confidence in the company’s long-term development.
The core driver behind DeepSeek’s soaring valuation is its recent technological breakthroughs.
On April 24, the preview version of DeepSeek-V4 was officially released and open-sourced. This large model, equipped with a new attention mechanism, has achieved a significant milestone in the development of domestic AI, enabling soft and hard synergy and autonomous development.
In the Agentic Coding evaluation, DeepSeek-V4-Pro has reached the best level among open-source models, and it has also performed exceptionally well in other Agent-related evaluations. In the world knowledge assessment, DeepSeek-V4-Pro has significantly outperformed other open-source models, slightly trailing behind the top closed-source model Gemini-Pro-3.1. In mathematics, STEM, and competitive coding evaluations, DeepSeek-V4-Pro has surpassed all publicly evaluated open-source models, rivaling the world’s top closed-source models.
Currently, the domestic large model competition is intensifying, with internet giants like ByteDance, Baidu, Alibaba, and Tencent leveraging their financial, computational, and ecological advantages to accelerate their self-research efforts. Top AI unicorns like Zhishu AI and MiniMax are rapidly emerging, with technological iteration and commercialization processes speeding up. Globally, companies like OpenAI, Anthropic, and Google have formed a triple barrier of technology, capital, and talent, with talent competition and computational resource investment costs continuing to rise.
In this context, DeepSeek may need more funding to expand its computational resources and support the continuous iteration of its models, while also increasing salaries to retain top research and development talent and ensuring the large-scale commercial deployment of its models.
Last year, at least five core research and development members of DeepSeek confirmed their departure, including the core author of the first-generation large language model, Wang Bing, the core contributor to V3, Luo Fuli, the core researcher of R1, Guo Daya, and the core authors of the OCR series, Wei Haoran, and the multi-modal achievement core contributor, Ruan Chong. They mostly chose to join internet giants like ByteDance, Tencent, and Xiaomi.
As the largest national-level industry investment fund in China, the National Integrated Circuit Industry Investment Fund’s potential participation in DeepSeek’s funding has significant implications. The fund was established in 2014 based on the State Council’s “National Integrated Circuit Industry Development Promotion Outline,” adopting a market-oriented fundraising and professional management model. The third phase of the fund was formally established in 2024, with a registered capital of 34.4 billion yuan, focusing on investing in key areas of the integrated circuit industry. Industry insiders believe that the combination of state capital and market-oriented capital can help DeepSeek achieve balanced development in technological innovation and industrialization.
As the transaction progresses, this domestic large model company, which has relied on technology to establish itself, will enter a new stage of development, from technological deepening to large-scale expansion. The progress of its funding and subsequent development will continue to influence the competitive landscape of China’s and the world’s AI industries.
