Governments have a choice in creating national AI strategies, build talent or buy it. China, Daniel W Hook argues, has already done both.
They provide the PRC with a world-transforming advantage ; “AI has become a geopolitical lever, not just a research field.”
“We have never encountered an exponential revolution before, and its societal impacts will provide both tremendous opportunities and profound disruptions. As we have seen in prior revolutions, productivity gains do not come without a price,” he writes.
Mr Hook sets out the state of the talent race in a paper published by his company, data analytics provider, Digital Science. He argues that for now, the United States leads in AI-to-innovation translation, due to university-industry links. But China dominates research demonstrated by the immensely powerful, low development cost LLM from the DeepSeek start-up.
China leads in the talent race both by making the most of its home-grown base and attracting international talent.
It has 30,000 AI researchers, compared to the EU with 20,000 and the US with 10,000. The head-count advantage counts, he says, creating “a generation of innovators and leaders who deeply understand this technology, potentially fuelling business and innovation at a level that no other research economy can match.” Plus, “the US is overwhelmingly losing talent to China.”
Patent output demonstrates what is happening, China is producing ten times more.
“China’s rapid expansion in research output, its evolving patterns of international collaboration, and its ascendant influence in global networks result in a significant share of global voice in AI. When taken with the volume of talent being produced in the country, the direction of travel appears inexorable,” Hook argues.
The consequence is a supercharged version of US-USSR 60’s space-race.
“Whomever controls the best AI will hold a competitive advantage in a variety of fields. One critical advantage will be an acceleration in research capacity—not just AI research itself but a general acceleration across research more generally.
