Two founders took the stage together at the University of Science and Technology of China in Hefei on September 6th. One, He Yu, is 34. The other, Liu Sizhe, is 18. Both came through the university’s ‘youth class’, a fast track for gifted children at a school South Korea’s Chosun Ilbo, which reported the scene, calls China’s answer to KAIST.
He Yu’s story begins with a room. As an undergraduate his supervisor gave him a 14-square-metre laboratory, room 5005, and left him to work on quantum measurement. In 2016 he founded Guoyi Quantum, a quantum-instruments company; last month it listed on the Shanghai Stock Exchange. The office Wi-Fi password, the newspaper notes, is still ‘5005’.
Liu Sizhe is a second-year student who studies geophysics. He has founded an artificial-intelligence company that reads the early signs of Alzheimer’s disease in the voice — a field with nothing obvious to do with his degree.
Seen from Seoul, both stories look strange: an undergraduate holding one laboratory for six years, an 18-year-old building a firm in a subject he never studied. But trace the pedigrees of China’s advanced-industry companies and the pattern repeats. The founders of chip designer Cambricon, facial-recognition firm Megvii, self-driving company Pony.ai, humanoid-robot maker Agibot and AI drug-discovery company Xingyao were all groomed inside gifted tracks such as USTC’s youth class and Tsinghua University’s Yao class.
The machinery is deliberate. Xi’an Jiaotong University selects gifted children in middle school, spares them the three entrance examinations of high school, university and graduate school, and attaches a research supervisor to each. Tsinghua’s X-lab work space has logged 2,900 projects over 13 years and produced 950 companies. At USTC this year, five undergraduates hold national research funding to lead work in AI, quantum computing, integrated circuits and robotics.
This is not the Bill Gates model. The Chinese prodigy does not flee the university; he uses it — develops the technology, gathers the colleagues, builds the prototype, and only then hangs up the company sign. Yao-class graduates often found firms with their classmates. Liang Wenfeng, of Zhejiang University, carried what he learned at a quantitative-trading firm built with university friends straight into DeepSeek. The product, by the age of 23 or 24, is what the newspaper calls the ‘nerd CEO’: technique, colleagues and failures already banked.
The attitude to big companies fits the same frame. When an engineer leaves a giant to found a firm, the common verdict is not ‘talent drain’ but ‘a virtuous cycle of the ecosystem’. Peng Zhihui, recruited by Huawei under its ‘genius youth’ scheme, built a self-riding bicycle in his spare time, quit in 2022 and founded the humanoid-robot company Agibot the following year. Much of his core team came out of Huawei’s own laboratories.
The correspondent’s discomfort is for his own country, and it is plainly stated. Korea’s gifted track is narrower, the race through science high schools and university admissions longer, the chance to lead research smaller. The children best at finding the right answer are given the least room to chase a wrong one. His closing question deserves to be repeated, because it applies far beyond Korea: is the frontier of technology crossed by the person who has never been wrong, or by the one who has failed fifty times and kept going? China is betting its best classrooms on the second.