They are all under 50, all products of China’s elite university system, and all running companies that now compete directly with OpenAI, Anthropic, and Google at the frontier of artificial intelligence. An article published by South Korea’s Chosun Ilbo has crystallized what the global tech industry has been observing for months: four Chinese founders have emerged as the defining figures of the current AI era, and their companies are no longer playing catch-up.
The four (Liang Wenfeng of DeepSeek, Yang Zhirun of Moonshot AI, Tang Jie of Zhipu AI, and Yan Junjie of MiniMax) represent distinct pathways into AI leadership. But they share a common thread: all were trained in China’s top institutions, all chose to build their companies domestically, and all are now generating more than half their revenue outside mainland China.
The Youngest King: Yang Zhirun and Moonshot AI
At 34, Yang Zhirun is the youngest of the four and arguably the one generating the most anxiety in Silicon Valley right now. Born in 1992 in Shantou, Guangdong, Yang graduated from Tsinghua University with top honors before completing a PhD at Carnegie Mellon University in just four years. He worked at both Google and Meta, reportedly rejected an offer from Apple, and returned to China to found Moonshot AI in 2023.
His company’s flagship model, Kimi K3, scored 57 points on the Artificial Analysis benchmark, ranking third globally behind only Anthropic’s Claude Phaebus 5 (60 points) and OpenAI’s GPT-5.6 Sol (59 points). The gap between China’s best and America’s best has never been this narrow. Moonshot is now planning a pre-IPO funding round in August that would value the company at $50 billion, a figure that would have seemed absurd for a three-year-old Chinese AI startup even 12 months ago.
The White House has accused Moonshot of distilling Anthropic’s Fable model to build Kimi K3, a charge the company denies, and the Commerce Department’s Bureau of Industry and Security is formally investigating the firm over chip usage. None of this has slowed its momentum.
The Quant Turned AI Pioneer: Liang Wenfeng and DeepSeek
Liang Wenfeng, 41, took the most unconventional path to AI leadership. A graduate of Zhejiang University with both a bachelor’s and master’s degree, Liang built a quantitative hedge fund managing over 100 billion yuan before spinning off its AI research division as DeepSeek in 2023. His background in financial modeling gave him an unusual perspective on computational efficiency — one that has defined DeepSeek’s approach.
At a recent investor meeting, Liang offered what may be the most concise summary of the current US-China AI dynamic: “The fundamental US-China AI gap stems from computational resources, not talent. DeepSeek achieves performance lagging 12-18 months behind the US with one-twentieth of computational resources.”
That statement captures both the challenge and the achievement. DeepSeek closed a $7.4 billion funding round at a $50 billion valuation in June, making it China’s most valuable AI startup. Liang has been explicit that AGI is the company’s sole goal and that commercialization can wait — a luxury afforded by the hedge fund’s deep pockets.
The Professor: Tang Jie and Zhipu AI
Tang Jie, 49, is the elder statesman of the group and the only one who remained in academia before founding his company. After completing his PhD at Tsinghua University in 2006, Tang stayed on as a professor and led the “Wuda” project — a research initiative involving over 100 people that laid the groundwork for Zhipu AI, which he founded in 2019.
Zhipu’s trajectory has been remarkable. Its stock has surged over 1,600% since its January listing, and the company’s chairman Liu Debing has seen his personal wealth reach $22.4 billion. Zhipu’s GLM-5.2 model, released through its subsidiary Z.ai, was flagged by NIST as the most capable open-weight model at release, and it was trained entirely on Huawei chips, demonstrating that China’s domestic hardware stack can now support frontier-class models.
The Corporate Climber: Yan Junjie and MiniMax
Yan Junjie, 37, took perhaps the most traditional corporate route. After earning his PhD from the Chinese Academy of Sciences in 2015 and completing a postdoctoral fellowship at Tsinghua, he joined SenseTime and rose to vice president in just six years before founding MiniMax in 2021.
MiniMax has distinguished itself through international reach. In 2025, the company generated $57.7 million in overseas revenue — representing 73% of its total sales, across more than 214,000 enterprise and developer clients in over 100 countries. Its M3 model launched in June broke into the global top three on multiple benchmarks, and the company has joined the Hang Seng Tech Index alongside Zhipu AI, earning blue-chip status barely months after going public.
The Talent Pipeline Behind the Founders
What makes the “Four Kings” narrative significant beyond individual achievement is what it reveals about China’s AI talent ecosystem. According to data from the MacroPolo think tank, the proportion of top-20% global AI researchers who completed their undergraduate education in China rose from 29% in 2019 to 51% in 2025. China’s elite programs, Tsinghua’s Yao Class and Peking University’s Turing Class, have become the world’s most prolific producers of frontier AI researchers.
This shift has been documented extensively in the Stanford AI Index, which found that while the US produced 59 notable model releases in 2025, China produced 35, and no other country managed more than eight. The talent pipeline is no longer a future threat; it is a present reality.
China’s strategy of recruiting AI talent directly from high schools, mandating AI education from primary school, and overhauling PhD programs to align with industrial priorities is designed to ensure the next generation of “kings” emerges from within — rather than returning from abroad.
What Comes Next
All four companies are now generating more than half their revenue outside mainland China, a fact that complicates Washington’s efforts to contain their influence. The US is weighing bans on foreign-made open-source models and investigating several of these firms, but as Nvidia CEO Jensen Huang said this week, American companies should “absolutely” be allowed to use Chinese AI models because “these Chinese models are excellent.”
The four founders are no longer insurgents. They are incumbents reshaping the terms of competition, and the global AI industry is adjusting to that reality whether Washington likes it or not.