Shield Xiao Dun on the bug China spent a century patching | Wealth Office
财富办公室

Insights / Asia and the new geography of wealth

Shield Xiao Dun on the bug China spent a century patching

Shield Xiao Dun traces China's technology strength to a culture of work and learning that, in his account, once undervalued science and has been correcting it.

A culture built for work

Explanations of China's rise in technology usually begin with capital, scale or state planning. Shield Xiao Dun, who founded 17 EdTech in 2011 and listed it on Nasdaq in 2020, starts somewhere else. Speaking with Michael Macfarlane in Hong Kong, he traces the country's technological strength to education and to the family, and he offers three threads to explain it.

The first is cultural. Xiao Dun is careful to say he is not an expert in Chinese history, but he observes that China has fought more internal wars than wars of conquest, and that most emperors were more interested in a harmonious society than in invading other territories. Harmony, in the philosophy he describes, comes from everyone performing their role. The father has a duty in society and so does the mother. The child's role is to grow into an adult who can one day fulfil a duty of their own, and learning is how that happens. The value passes from one generation to the next. In his words, valuing work and valuing learning is deeply rooted.

He ties this directly to the demand his first company met. Parents in even the most remote village, he says, regard education as the way to change a child's destiny and the family's destiny with it. He contrasts it with winning a lottery. Money that arrives without a change of mindset tends not to last, he suggests, while education changes what he calls the operating system within a person.

The bug and the patches

The second thread is historical. Before 1900, in Xiao Dun's account, China was a proud and confident country. Around 1900 it was invaded by eight countries, including the UK, and came to realise the power of technology. He lists what followed: two world wars in which China suffered the greatest losses of life, a great famine, the Cultural Revolution, and then Deng Xiaoping's policy of reform and opening up. He reads that policy as a decision to value the power of technology and to embrace capitalism, at least as a means to a more harmonious society. The past 40-odd years, he notes, have brought much faster development than the 40 before them.

Macfarlane frames the present acceleration as a return to a long trend line after a twentieth century of setbacks. Xiao Dun broadly agrees, and summarises the story as an engineer would.

I think the traditional culture of China is a culture that enables people to work well. But there was a bug of not valuing science and technology enough. We have been trying to put patches on this.

— Shield Xiao Dun, Wealth Office

The patches are visible in the classroom. When he was growing up, he recalls, people said that anyone who learned maths, physics and chemistry well would have a good future. He puts the number of STEM graduates and postgraduates China produces each year at around three million, and says they go into every industry and make engineering innovations. He sees a similar competitiveness in India and South Korea. The leadership, in his view, now recognises the importance of science and technology.

Where the engineers end up

The third thread is what those engineers build. Xiao Dun sees it first as a consumer. Electric cars, in his view, have become largely a contest between Tesla and Chinese manufacturers, with established German, Japanese and American brands behind. Manufacturing that once meant clothing and small household items now means digital components. To supply Apple or Tesla, he says, a company is building chips or sophisticated switches and transistors, at low cost and high quality at once. Most of that supply chain, for now, is still in China, and consumer brands are beginning to emerge from it. He illustrates the point with a new pair of glasses he gave William Louey, who told him that another brand he had been trying was better. Xiao Dun tried it and found it better in many ways, and not only cheaper.

On artificial intelligence his assessment is measured. He sees China and the United States as the two clear leaders, with the US ahead in compute and the state-of-the-art models still made by American companies. When ChatGPT appeared, he estimates, Chinese models were perhaps five years behind. Now he believes they are only months behind. He notes that DeepSeek is open source and suggests that developers tend to open-source their second-best models, while allowing that an open model could one day become the Android of the field, a different architecture and not necessarily a worse one.

The price of the patches

Xiao Dun does not present the system as finished. In the same conversation he credits East Asian countries with better long-term, systematic execution, and credits Palo Alto, where he now works, with the self-driven exploration that breaks a system and builds something new. He suggests China can learn from that.

He is also clear that capital markets still price China's political risk. The most successful Chinese technology companies, he observes, have listed in the United States. He says he is not an expert in capital markets, but puts institutional ownership of China's A-share market below 30 per cent, compared with the reverse in the US, and describes a Chinese market that is more volatile and more influenced by policy. Geopolitical tension adds to the pressure. Chinese companies listed in the US, in his account, face much more scrutiny, and he points to ByteDance, whose TikTok business he says was broken into two pieces. Macfarlane suggests that Chinese technology trades at a discount precisely because of that exposure, and Xiao Dun agrees.

For families and investors who follow Chinese technology, his account offers a way of reading it. The strength he describes is cultural and educational in origin, built on families that treat learning as a duty and on a state that, in his view, has spent decades correcting an old blind spot. The risks he describes are political and geopolitical, and he does not wish them away. Macfarlane prefers a world in which nations race over AI models to one in which they race over arms. Xiao Dun's own framing is quieter. A culture that already valued work has been fixing a bug, and in his telling the patches are still being applied.