An irony of DeepSeek releasing its cutting-edge open-source artificial intelligence model that wiped around a trillion dollars off the share market value of United States tech giants is that it is wholly owned by a for-profit Chinese hedge fund. In contrast, OpenAI, despite its name, offers a closed source model with limited details of its AI coding. OpenAI, an American company, started as a non-profit but turned for-profit when it became apparent there was serious money to be made. It’s easier to stay profitable with less competition so it is not so much ironic as deeply cynical that Sam Altman, OpenAI’s CEO, was an active proponent of then president Joe Biden’s executive order on AI that raised barriers for newcomers to enter the field. As technology writer Ben Thompson wrote in his blog Stratechery: “it sure is notable that the early AI winners seem the most invested in generating alarm in Washington, DC about AI. This despite the fact that their concern is apparently not sufficiently high to, you know, stop their work. No, they are the responsible ones, the ones who care enough to call for regulation; all the better if concerns about imagined harms kneecap inevitable competitors”. The step by DeepSeek towards AI being a readily available commodity has rendered ineffectual the ostensible guardrails in Biden’s executive order. It wouldn’t have mattered if President Donald Trump hadn’t rescinded the order. The interesting question is why have the roles reversed? Why has a Chinese for-profit company behaved in a way commonly associated with an open society while OpenAI has belied its name? One potential explanation is that it is more profitable for DeepSeek to make high quality AI widely available because it will drive demand for something else it is selling. It is like a seller of printing ink subsidising sales of printers to increase consumer demand for ink. The strategy is called “commoditising your complement”. This explanation will appeal to those who characterise mainland Chinese as cunning deep thinkers with long time horizons but lacking the ability to engage in fundamental innovation. An interview in November 2024 with hedge fund manager Liang Wenfeng, DeepSeek’s founding CEO, who was educated in China and whose father was a primary school teacher in Guangdong province, points to another explanation. The interview provides three key insights. One is that China has reached the point where economic progress requires fundamental innovation rather than commercial application of existing technology. Second, it has developed the ability to do so. Wenfeng was trained entirely in China and so were all the people DeepSeek has hired. Third, Wenfeng recognises a key element in attracting the talent to make breakthroughs is a loose organisational structure that tolerates time spent on projects that turn out to be dead-ends and provides personal recognition of breakthroughs by individuals. He said: “for technical talent, having others follow your innovation gives a great sense of accomplishment. In fact, open source is more of a cultural behaviour than a commercial one and contributing to it earns us respect. There is also a cultural attraction for a company to do this”. It’s a philosophy that probably would work well in a university too! The world is full of surprises. I didn’t expect a mainland Chinese hedge fund manager to be the person making the most effective and impactful case for open knowledge. Well might we say, happy Chinese New Year! Winthrop Professor Raymond Da Silva Rosa is an expert in finance from The University of Western Australia’s Business School