DeepSeek's Strategic Move and Vision for AGI

DeepSeek, a Chinese artificial intelligence start-up, has made a significant move in the AI landscape by closing its first-ever external fundraising round. This high-stakes deal not only reshapes the country's AI scene but also offers insight into how one of China's most valuable AI start-ups will grow in the future.

After raising approximately 50 billion yuan in a Series A round, Hangzhou-based DeepSeek now has a post-investment valuation of about 400 billion yuan (US$59.2 billion). This makes DeepSeek one of China's most valuable AI start-ups, surpassing the US$30 billion valuation of Moonshot AI and the US$17.7 billion market capitalisation of Hong Kong-listed MiniMax AI as of Wednesday's close. However, it still trails behind Zhipu AI's US$95 billion market value.

Founder's Leadership and Commitment to AGI

Contrary to earlier speculation that the China Integrated Circuit Industry Investment Fund would lead the round, Liang Wenfeng, DeepSeek's low-profile founder and CEO, emerged as the largest backer with around 20 billion yuan from his own pocket, making up nearly half of the total. This unusual arrangement ensures that Liang maintains control over the company's direction as he remains committed to pursuing artificial general intelligence (AGI), which aims to match human capabilities.

Liang reinforced this vision during an online meeting with investors last month, emphasizing that "DeepSeek wants to do anything relevant to enhance the intelligence" of AI models. He expressed no interest in other ventures, according to a source.

Major Investors and Unique Structure

Other participants in the round included Tencent Holdings and Contemporary Amperex Technology Limited (CATL), committing 10 billion yuan and 5 billion yuan respectively, making them the largest outside backers. NetEase and JD.com each contributed around 3 billion yuan. Additionally, China's National Artificial Intelligence Investment Fund, a joint venture between state-backed Guozhi Investment (Shanghai) Private Equity Fund Management and the Big Fund Phase III, chipped in nearly 1 billion yuan.

Liang had "substantial bargaining power" over investors, including the national AI fund, and none of the investors were getting preferential terms or control over DeepSeek. This arrangement discouraged the national fund from leading the round, according to a Beijing-based investor who requested anonymity. The structure was designed to ensure Liang had the final say in DeepSeek's direction and to deter investors seeking short-term gains, the investor said.

Talent Retention and Competitive Landscape

Liang, the founder of High-Flyer Quant, one of the largest and most profitable hedge funds in China, was financially well-positioned to be selective when it came to investors and equity distribution. His decision to lead the investment was driven by a desire to maintain control over DeepSeek.

DeepSeek's emergence as a national AI champion in early 2025 was marked by the development of V3 and R1, open-source AI models that matched certain capabilities of US frontier AI labs like OpenAI and Anthropic at a fraction of the cost. Despite his long resistance to raising outside capital, Liang came around earlier this year due to DeepSeek's loss of talent to local rivals.

The funding round established a valuation benchmark, giving DeepSeek employees a better sense of their equity value and helping stem poaching amid intense competition for talent, sources told the SCMP earlier.

Challenges and Future Outlook

Liang also requested that investors backing its first round should not hire from DeepSeek or encourage its employees to take their equity money and start their own businesses, according to a report published on Wednesday by Elsewhere Biechufasheng, an independent media outlet focusing on Chinese venture capital, citing anonymous sources.

DeepSeek's funding comes as China's AI rivalry becomes increasingly competitive. Economist Pan Helin noted that AI companies are facing mounting pressure to build application ecosystems around their technology, requiring substantial financial backing to sustain.

"[Chinese] AI models have become increasingly homogeneous, with no company holding a clear technological edge over another," Pan said.

This article originally appeared on the South China Morning Post (www.scmp.com), the leading news media reporting on China and Asia.