Topline
The man who founded DeepSeek, the artificial intelligence company that rattled the U.S. stock market, is 40-year-old Liang Wenfeng, a former hedge fund manager who said he shifted into tech to close the gap between China and the U.S. in the AI industry.
The 40-year-old entrepreneur launched the Chinese AI firm years before it rattled global tech … [+]
Key Facts
Liang told Chinese outlet Waves he grew up in Guangdong, China, in the 1980s—reportedly the child of teachers in the area, which is now known for its tech industry—and he later received a bachelor’s and master’s degree in information and communication engineering from Zhejiang University, according to Reuters.
In 2015, Liang co-founded High-Flyer, a quantitative hedge fund that relies on “mathematics and AI” to create investment strategies.
High-Flyer created its first AI model in October 2016, allowing the company to take its first stock position based on AI projections after previously relying on “traditional machine-learning algorithms.”
High-Flyer started using AI models to develop nearly all of its stock positions by 2017, after which the hedge fund hired a research team dedicated to “AI algorithms and [their] basic applications.”
Liang started accumulating thousands of Nvidia graphics processors for a then-unnamed AI project in 2021, just before the Biden administration restricted trade of those chips to China, according to the Financial Times.
Liang founded DeepSeek in 2023, and the company relied on maximizing the output of processors available in China while being unable to acquire Nvidia’s chips, the Financial Times reported, citing an AI researcher close to the company.
Get Forbes Breaking News Text Alerts: We’re launching text message alerts so you’ll always know the biggest stories shaping the day’s headlines. Text “Alerts” to (201) 335-0739 or sign up here.
What Has Liang Wenfeng Said About Ai?
Liang told the Chinese tech outlet 36Kr that China must “gradually transition” to being a contributor in the AI industry, “rather than continuing to ride on the coattails of others,” according to a translation by the Post. China will “always be a follower” in the industry unless the country closes a “one- or two-year gap” with AI in the U.S., Liang said.
What Are Liang Wenfeng’s Ties To The Chinese Government?
Liang spoke at a closed-door event for China’s Premier Li Qiang with experts from other industries on Jan. 20, according to a summary of the event translated by the Washington Post. Li said during the event the Chinese government should “focus on breakthroughs in key core technologies and cutting-edge technologies.” Some local outlets cite statements from provincial and city governments in China, the Post reported, in addition to an AI expert at Beijing University who said DeepSeek could “create miracles.”
What Is Deepseek?
DeepSeek, reportedly founded by Liang with $1.4 million in capital, released its open-source models for download in the U.S. earlier this month. DeepSeek’s AI model quickly jumped to the top of iPhone download charts, including ahead of OpenAI’s ChatGPT. DeepSeek says its product outpaces rival models in mathematical tasks, general knowledge and question-and-answer performance benchmarks. DeepSeek’s advanced reasoning model R1 has been compared to similar products made by OpenAI and Meta, though they appear to be more efficient with lower costs to train and develop models. The company says it developed its model for a tiny fraction of the cost of most U.S. competitors, rattling American chipmakers and AI-focused companies—though some skeptics have questioned whether the true cost is higher than DeepSeek claims. The model also refuses to answer questions on some controversial topics, however, including queries about China’s treatment of Uyghurs, Taiwan’s status with China and what happened at Tiananmen Square in 1989. Chatbot Arena, a performance platform hosted by the University of California, Berkeley, ranks DeepSeek’s R1 and V3 models among its top 10 products, including ahead of some models developed by OpenAI, Google and Elon Musk’s xAI. Liang told 36Kr the company’s advancements in AI are limited because DeepSeek is unable to access chips developed by Nvidia.
What Does Deepseek Say About Liang Wenfeng?
When DeepSeek is asked about Liang’s background, the AI chatbot told Forbes there is no information publicly available about Liang, including his “place of origin or educational history.” DeepSeek also notes details about Liang’s personal life and academic background are not “extensively publicized,” as he has maintained a “relatively low profile compared to some other tech entrepreneurs.” The Associated Press was the first to ask about Liang.
Key Background
The public release of DeepSeek’s generative AI model ignited a selloff across tech stocks in the U.S., led by Nvidia’s loss of nearly $600 billion. Several billionaires had their fortunes cut by the market reaction, including losses of $27.6 billion for Oracle chairman Larry Ellison and $20.8 billion for Nvidia CEO Jensen Huang. The company’s model appeared to rival OpenAI’s viral chatbot ChatGPT, with speculation about DeepSeek’s product running at a fraction of the cost of U.S.-based rivals. President Donald Trump, who recently announced a $500 billion AI infrastructure project, said the release of DeepSeek’s AI model “should be a wake-up call” for U.S. industries while applauding the company as a “positive development” because of its low costs.
Further Reading
Read the full article here