Liang Wenfeng
Founder of DeepSeek
Quant-fund billionaire turned AI lab founder whose cheap, open-source DeepSeek-R1 model shook global markets in January 2025. Pursues AGI openly and insists China must originate ideas, not just follow them.
Why this score
How the formula works →- +4.1proBuilds AI professionally
- +3.3proInvests in AI companies
- +1.4pro3 pro-AI statements (forceful)
Weights, statement counts, the ×1.25 multiplier for having a stake in an AI-impacted field, and the formula’s smoothing are already baked in — the points add up to the score: Strongly Pro-AI · 8.8/10
In his words
Our destination is AGI, which means we need to study new model structures to realize stronger model capability with limited resources.
The real gap is the difference between originality and imitation. If this doesn’t change, China will always be only a follower.
Open source is more of a cultural behavior than a commercial one, and contributing to it earns us respect.
Biography
Liang Wenfeng (born 1985 in Zhanjiang, Guangdong) studied electronic engineering at Zhejiang University and made his first fortune in quantitative trading, co-founding the hedge fund High-Flyer in 2015 and stockpiling thousands of Nvidia GPUs for its trading models. In 2023 he spun that compute into DeepSeek, an AI lab with an unusual charter: pure research toward AGI, staffed largely by young domestic PhDs, with its models released as open source.
The Sputnik moment
In January 2025 DeepSeek released R1, a reasoning model rivaling OpenAI’s best, reportedly trained at a fraction of American costs — and gave it away. The app shot to No. 1 on US charts, Nvidia shed roughly $600 billion in market value in a single day, and Marc Andreessen called it “AI’s Sputnik moment.” Liang, until then nearly unknown, was summoned to meet Premier Li Qiang and became a national symbol of Chinese AI self-reliance.
In rare interviews with the outlet 36Kr, Liang argued that China’s tech industry had coasted on imitation for decades and that hardcore, original innovation — not commercialization — was DeepSeek’s purpose. His stance is not without complications: R1’s low training cost is debated, US officials have scrutinized how DeepSeek obtained restricted chips, and the model censors politically sensitive topics in line with Beijing’s requirements. His open-source strategy has nonetheless reset industry economics, with rivals from Meta to Sam Altman’s OpenAI forced to respond.
Quote sources
- 36Kr interview, via ChinaTalk, 2024(chinatalk.media)
Sources & further reading
Canonical record: https://battlelines.ai/topic/liang-wenfeng
