Liang Wenfeng, who has up-ended the artificial intelligence industry (AI), is a man of few words.
The 40-year-old grew up in China's southern province of Guangdong in an era when the province led the rest of China in adopting market capitalism in the 80s and 90s.
Mr Wenfeng is now one of the players leading China in its AI revolution as the superpower attempts to keep pace with the dominant industry in the United States.
DeepSeek shook US markets on Tuesday after unveiling its new AI chatbot, which it said was cheaper and less resource-intensive than AI chatbots made in the US.
It saw companies such as US chipmaker Nvidia lose almost a trillion dollars of its value.
He had kept a low profile but in January 20 of this year, he was one of nine individuals asked to give a speech at a closed-door symposium hosted by China's Premier Li Qiang.
Before that, he gave two media interviews to Chinese media outlet Waves last year, and in 2023.
But apart from that he has stayed mostly out of the public eye
Despite his limited media appearances and public statements over the years, Mr Wenfeng hasn't been shy about expressing his views on China's role in the AI arms race.
"China's AI can't be in the position of following forever. We often say that there is a gap of one or two years between China's AI and the United States," Mr Wenfeng said in an interview with Waves in July last year.
"But the real gap is the difference between originality and imitation … If this doesn't change, China will always be only a follower — so some exploration is inescapable."
China no longer a follower in tech
Mr Wenfeng studied at the elite Zhejiang University at age 17, majoring in electronics and communication engineering before pursuing a master's degree in information and communication engineering, which he completed in 2010.
He then co-founded a quantitative hedge fund High-Flyer in 2015, which uses complex mathematical algorithms for trading as opposed to human analysis.
The fund's portfolio totalled over 100 billion yuan ($22 billion) by the end of 2021.
But in April 2023, it announced on its WeChat account that it would expand its remit beyond the investment industry and concentrate resources to "explore the essence of AGI".
That's artificial generative intelligence, and a month later DeepSeek was created.
The company released its first AI large language model later that year.
Under Mr Wenfeng's leadership, DeepSeek deliberately avoided app-building.
Instead, it concentrated research talent and resources on creating a model that could match, or better OpenAI.
His approach stood out in a Chinese tech industry that was used to taking innovations from abroad, from smartphone apps to electric vehicles, and quickly scaling them up, often much faster than the countries where the inventions were first made.
This time it was China leading rather than following, a shift Mr Wenfeng had wanted to see.
Making DeepSeek open-source allows for soft power, the founder says
The DeepSeek chatbot became more widely accessible when it appeared on Apple and Google app stores this year.
But it was a follow-up research paper published last week — on the same day as President Donald Trump's inauguration — that set in motion the panic that followed.
That paper was about DeepSeek AI model called R1 that showed advanced "reasoning" skills — such as the ability to rethink its approach to a maths problem — and was significantly cheaper than a similar model sold by OpenAI called o1.
It triggered a global sell-off in tech stocks.
Nvidia, the biggest casualty so far of DeepSeek's dramatic interruption of the AI market, has welcomed the Chinese AI firm's sudden appearance.
Nvidia dropped 17 per cent, or $US590 billion ($940 billion), on Monday night after DeepSeek released its chatbot.
Despite this, Nvidia described DeepSeek's innovation as an "excellent AI advancement".
In a statement, Nvidia said more of its chips would be needed in the future to meet demand for DeepSeek's services.
"DeepSeek's work illustrates how new models can be created using that technique, leveraging widely available models and compute that is fully export control compliant," Nvidia said.
DeepSeek has decided to make all its models open-source, unlike its US rival OpenAI.
In open-source models, the base code is publicly available for any developer to use and modify at will.
"Even if OpenAI is closed-source, it cannot stop others from catching up," Mr Wenfeng said.
"Open-source is like a cultural practice, rather than a business practice … a company that does this will have soft power."
Android era of AI has begun, expert says
Mr Wenfeng's rare public comments reveal a belief that China's tech industry had reached a crossroads.
It lacked the confidence but not the capital needed to engage in fundamental research and development breakthroughs.
"In the past 30 years, (China's tech industry) has only emphasised making money, and ignored innovation," he said in July.
"Innovation is not solely driven by business, it also needs curiosity and a desire to create."
University of Sydney's Chang Xu, an associate professor who specialises in machine learning and computer vision, said due to DeepSeek parent company High Flyer's success, it was not driven by profit.
"The impressive thing is its cost-efficient … I think it is a success of the extreme engineering and algorithm innovation," Dr Xu said.
He said the the limitations on US chips available in China meant companies such as DeepSeek were pushed into the corner leading to innovating both from an engineering and algorithm perspective.
But DeepSeek's decision to go open-source was what Dr Xu believes will lead to the biggest shift within the AI industry.
He drew a comparison with the smartphone market, where competition exists between the closed system and the open-source platforms such as android.
"So I think this actually is a kind of start of the android era for the larger AI models that has been marked by DeepSeek," he said.
"So its open-source framework is quite a success to learn differences from the closed processor such as openAI and Google's large AI models.
"So this open-source framework can be widely accessed by the public, and it's actually quite friendly to the research committee, so we can have the freedom to access the research building model."
What does DeepSeek's AI chatbot say about its founder?
When DeepSeek was asked, "Who is Liang Wenfeng?" its first answer was to name a different Chinese entrepreneur with the same name, at least as spelled in English letters.
When asked: "Where is Liang Wenfeng from and where did he go to university?" it said that as of October 2023 — the most recent knowledge cut-off for DeepSeek's R1 AI model — "there is no publicly available information about Liang Wenfeng's background, including his place of origin or educational history."
"If you are referring to the founder of DeepSeek, details about his personal life or academic background have not been disclosed publicly. For more information about DeepSeek, you can visit its official website," it said.
ABC/Wires
