- China leads in AI research, publishing more than the US, UK, and EU combined.
- It’s also the top global AI collaborator—while relying less on others.
Artificial intelligence has become more than a technological trend—it’s now viewed as a national asset. A new report from research analytics firm Digital Science shows that China is pulling far ahead in AI research, outpacing the US, UK, and EU in publication volume, patent filings, and researcher activity.
The report, DeepSeek and the New Geopolitics of AI, written by Digital Science CEO Dr Daniel Hook, draws from the Dimensions research database. It reviews global AI trends from 2000 to 2024, including research output, international partnerships, talent flows, and innovation outcomes. The conclusion is clear: China is now the most dominant force in AI research, and it’s widening the gap.
China outpaces global peers in research and citations
Back in 2000, fewer than 10,000 AI papers were published globally. In 2024, that number hit 60,000. But not all growth has been equal. China now produces as much AI research as the US, UK, and EU-27 combined. In terms of research attention, China captured over 40% of global citations in 2024—four times higher than the US and EU individually, and 20 times more than the UK.
More importantly, China is building a research ecosystem that doesn’t rely on others. While the US, UK, and EU remain tightly linked in AI collaboration, they are all more dependent on China than China is on them. Only 4% of China’s AI publications in 2024 involved collaborators from these regions. By contrast, 25% of the UK’s AI papers included a co-author from China, making China its top research partner.
Even the US—despite years of political tension and efforts to decouple—maintains its highest AI research ties with China. These relationships have persisted even as legislation like the China Initiative and chip export controls aimed to limit such collaboration.
Dr Hook argues that AI has become a key tool of geopolitics, similar to energy or defence capabilities. “AI is no longer neutral,” he writes. “Governments are using it as a strategic asset.”
DeepSeek shows China’s technical independence
China’s development of DeepSeek, a cost-efficient, open-source chatbot released in early 2025, shows how the country is finding workarounds to chip shortages and positioning itself for leadership. DeepSeek didn’t require expensive GPU training runs and was released under an MIT license—moves that signal both technical skill and confidence.
Beyond the model itself, DeepSeek is symbolic of something larger. China has built a vast, young, AI-trained population. More than 30,000 AI researchers are currently active in China. Its combined PhD and postdoc base alone is double the size of the entire US AI research population. By comparison, the US has about 10,000 researchers, the EU-27 around 20,000, and the UK roughly 3,000.
What stands out isn’t just scale—it’s structure. China’s AI workforce is overwhelmingly young, with relatively few senior researchers. This suggests China is investing in long-term capacity, rather than relying on a few high-profile experts. It’s also drawing in talent from overseas. The report notes that China is now a net gainer of AI researchers from countries like the US and UK, reversing earlier trends.
Research is translating into innovation
While China’s talent base is growing, its research is also translating into patents and products. The report highlights how China now files nearly ten times more AI-related patents than the US. It’s not just publishing papers—it’s protecting ideas and building businesses around them.
Geographically, China’s approach is different too. Its AI research is spread widely, not just clustered in a few cities. In 2024, 156 institutions in China published more than 50 AI papers each. These included universities, companies, hospitals, and research centres located in places like Beijing, Shanghai, Nanjing, and Guangzhou. The US had 37 such institutions, the EU-27 had 54, and the UK had 19.
This nationwide spread shows that China isn’t betting on a few research hubs—it’s building a broad AI infrastructure that reaches across the country. That could make it harder to disrupt or outcompete.
EU lags, UK punches above its weight
Meanwhile, Europe shows signs of falling behind. While EU countries collaborate well internally, they’re less connected to outside regions and struggle to turn research into patents or startups. Notably, despite France’s vocal investment in AI since 2018, no French research institution published more than 50 AI papers in 2024. Even top performers like Université de Toulouse fell just short.
The UK, despite its smaller size, continues to punch above its weight in AI visibility. Its research consistently draws more citations than expected for its volume, showing a high attention-per-output ratio. But it, too, is now leaning heavily on China for collaboration.
China’s companies are gaining ground
In terms of corporate involvement, China is also closing in on the US. While US companies still publish more AI papers overall, Chinese companies are gaining fast. The number of research-active companies in China has nearly caught up to the US, suggesting that China’s private sector is becoming a bigger player in AI R&D.
Dr Hook notes that much of US AI research now happens behind closed doors, in private firms like OpenAI. That makes it harder to track and may distort the picture. Even so, the report’s data shows the US is at risk of losing its narrow lead in research-focused AI startups.
The global shift has already happened
The report calls attention to a broader trend: China is not just a competitor—it’s becoming the preferred global connector in AI research. While Western countries still have strong academic networks and commercial pipelines, China’s scale, growth rate, and independence set it apart.
DeepSeek didn’t appear out of nowhere. It reflects years of investment, education, and infrastructure-building. It also signals what may come next—not just more chatbots, but a wave of AI tools built by a large, skilled workforce, increasingly operating on China’s own terms.
In the next decade, the report suggests, the real advantage may go to the countries that can not only attract talent and fund research, but also build systems that let that work reach society. China appears to be doing all three—and doing it faster than anyone else.