After years of concern in the United States that China might surpass its ambitions in artificial intelligence (AI), the biggest threat to Silicon Valley’s dominance did not come from one of China’s four major tech companies but from an unknown startup.
Without prior warning, 39-year-old Chinese entrepreneur Liang Wenfeng, the founder of the Chinese AI company DeepSeek, emerged and, within a few weeks, became the face of China’s tech industry and its hope to overcome the strict export controls imposed by the United States.
According to Reuters, Liang maintained an extremely low profile until just days ago, having only given two rare media interviews with the Chinese media outlet Waves in 2023 and 2024. Other than that, he largely remained out of the public eye.
On January 20, he was one of nine individuals invited to give a speech at a closed seminar hosted by Chinese Premier Li Qiang. His youthful appearance contrasted with the gray-haired academics, seasoned Chinese officials, and state-owned conglomerate executives seated around him, as shown in photos and videos published by China Central Television (CCTV).
Reuters highlighted the fact that Liang was invited to share his views on government policy, emphasizing Beijing's recognition of DeepSeek’s role in shifting the global AI landscape in China’s favor.
Talent Focus
Last year, Robin Li, CEO of Baidu, spoke at a similar seminar chaired by the Chinese Premier.
In an interview that same year, Li—who announced China’s first ChatGPT rival in March 2023—said that China would never be able to replicate OpenAI's success and that Chinese companies should focus on applying existing AI models for commercial purposes.
Last week, DeepSeek launched a free AI assistant, claiming it uses less data at a fraction of the cost of existing services, which triggered a global sell-off in tech stocks.
According to Reuters, under Liang's leadership, DeepSeek deliberately avoided building applications. Instead, it concentrated its research talent and resources on developing a model that could rival or surpass OpenAI. The company hopes to continue focusing on advanced models that other businesses can use to build AI products for consumers and enterprises.
Reuters added that Liang’s approach stands out in China’s tech industry, which has traditionally focused on rapidly scaling innovations from abroad, whether in smartphone apps or electric vehicles, often at a much faster rate than in the countries where the inventions were first developed.
Complex Algorithms
Liang grew up in Guangdong Province, which led China’s market capitalism boom in the 1980s and 1990s. At the time, he was surrounded by people who preferred starting businesses over academic study, but he was more inclined toward academia, as he later stated.
At 17, he enrolled at the elite Zhejiang University, majoring in electronic and communication engineering, and later earned a master’s degree in information and communication engineering, which he completed in 2010.
In 2015, he co-founded a quantitative hedge fund that used complex mathematical algorithms for trading instead of human analysis.
By the end of 2021, the fund’s total portfolio exceeded 100 billion yuan ($13.79 billion). However, in April 2023, he announced via his WeChat account that he was expanding beyond the investment industry and would focus on exploring the essence of artificial general intelligence (AGI)—which OpenAI defines as autonomous systems that surpass humans in most economically valuable tasks.
One month after Liang’s statement, DeepSeek was founded. According to Reuters, most of its employees are graduates and PhD students from top Chinese universities, who prefer working at the company because it tackles some of AI’s biggest challenges.
It is reported that Liang is personally involved in the company’s research and has expressed his preference for hiring local talent for the company’s headquarters in Hangzhou—home to Alibaba—instead of workers who studied in the United States or abroad, according to The Guardian.
In July, he stated:"What attracts the best talent is solving the world’s toughest problems. Our goal remains reaching artificial general intelligence."
Open Source Approach
In his interviews, Liang expressed the belief that China’s tech industry has reached a crossroads, lacking confidence and capital to achieve fundamental breakthroughs in research and development.
He has also embraced the open-source culture, which American tech experts previously claimed gave Silicon Valley an advantage over China.
As a result, DeepSeek decided to make all its models open-source, in contrast to its American competitor OpenAI.
In open-source models, the core code is available to the public, allowing any developer to use and modify it as they see fit.
In a July 2023 interview with Waves, Liang said:"Chinese AI cannot remain a follower forever. We often say there is a one- or two-year gap between Chinese and American AI, but the real gap is between originality and imitation."
He added: "For the past 30 years, China’s tech industry has only focused on making money, ignoring innovation. True innovation is not just about business—it requires curiosity and a desire to create."
He continued: "Even if OpenAI remains closed-source, it cannot stop others from catching up. Open-source is more of a cultural practice than a business strategy, and companies that adopt it will gain soft power."
Navigating U.S. Chip Restrictions
According to The Guardian, before former U.S. President Joe Biden announced sweeping export controls on semiconductors to China in 2022—aimed at preventing the country from acquiring the equipment needed for rapid AI development—Liang had already purchased 10,000 Nvidia chips.
While Nvidia’s powerful H100 chip was banned, Liang’s company developed H800 chips, a less powerful version for the Chinese market, which trained DeepSeek’s AI model—although these were also banned in 2023.
At the time, Liang stated:"Our biggest challenge was never money—it was the ban on high-quality chips."