Liang Wenfeng, founder of the Chinese artificial intelligence start-up DeepSeek and quantitative hedge fund High-Flyer Quant, has outlined a distinctive strategic vision that prioritizes long-term technological breakthroughs over immediate commercial gains. A recently leaked transcript from a closed-door meeting with potential investors in May offers an in-depth look at Liang’s approach, which emphasizes corporate restraint, low-margin pricing, and an open-source model.
During the nearly four-hour discussion, Liang repeatedly stressed the principle of “restraint,” explaining that DeepSeek deliberately sets prices to earn reasonable profits rather than maximizing short-term income. According to the transcript, this strategy is intended to make AI technology more accessible, thereby increasing the likelihood of achieving artificial general intelligence (AGI)—a level of AI capability comparable to or exceeding human cognition.
Liang acknowledged DeepSeek’s current status as far from fully commercialized but positioned the firm’s low-margin approach as a tactical sacrifice aimed at enhancing the probability of reaching AGI. “I have no doubt that AGI will have tremendous commercial value,” he stated, “but my priority is not to gain more market share now, but to increase the probability of achieving [AGI].”
DeepSeek, which was spun off from High-Flyer Quant in 2023 to pursue focused AGI research, attracted intense investor interest during its first external funding round in May. The company’s valuation surged from an initial US$10 billion to approximately US$59 billion by June. Liang personally invested nearly US$3 billion in the round, maintaining majority control and strategic direction. However, the company has reportedly paused plans for a second funding round aimed at raising at least 10 billion yuan (HK$11.57 billion) following the leak.
Liang’s relatively low public profile contrasts with his estimated net worth of close to US$3 billion, making him one of the wealthiest founders among pure-play AI model developers worldwide. His reclusive demeanor and measured communication stand in contrast to the high-profile personas common in the global tech sector.
Industry experts note that DeepSeek’s corporate culture and strategy reflect a broader cultural divergence between Chinese and Western AI ventures. Guo Tao, a member of the Chinese Association for Artificial Intelligence, highlighted that while American AI labs such as Anthropic and OpenAI focus heavily on commercial rollout and ecosystem building, DeepSeek is uniquely centered on achieving advanced technological breakthroughs without seeking rapid market dominance. Similarly, Qian Yenying, dean of the Giant Institute of AI and Business Innovation in Shanghai, observed that DeepSeek eschews emphases placed by Western competitors on large-scale deployment, institutional governance, and technical efficiency.
Addressing the geopolitical context, Liang acknowledged the technological competition between China and the United States, pinpointing access to computing resources—particularly advanced AI chips from Nvidia—rather than talent as the principal challenge for Chinese firms. He indicated the necessity for DeepSeek to develop its own computing clusters but said the company had yet to determine whether it would pursue proprietary chip design.
Closing the meeting, Liang described DeepSeek’s team as “a group of very ordinary people,” challenging Silicon Valley’s narrative of exceptional genius. “I prefer the narrative that ‘a group of ordinary people did extraordinary things’ rather than ‘geniuses did extraordinary things,’” he said, framing the company’s achievements within a philosophy of humility and collective effort.