Despite U.S. efforts to suppress China’s microchip industry through export controls, China is making significant advances in fields like photophysics.
A recent analysis found that the majority of foundational research supporting future computing hardware now comes from China. If this research translates into commercial applications, the U.S. may find that maintaining its dominance in high-performance microchip design and manufacturing through export restrictions alone is unrealistic.
Published on March 3, the study found that between 2018 and 2023, Chinese institutional authors appeared more than twice as often as U.S. authors in chip design and manufacturing research. And it’s not just about quantity—50% of the top 10% most-cited papers in their publication year had Chinese institutional co-authors, compared to 22% for U.S. institutions and 17% for European ones.
This research spans various disciplines, from traditional compute chips and AI-optimized graphics processors to entirely new architectures. To compile and categorize the data, ETO analysts trained a machine learning algorithm, focusing on emerging chip technologies rather than commercial advances, which tend to be incremental and proprietary, Arnold explained.
The study included only papers with English abstracts, which are more likely to target an international audience when authored by Chinese researchers. China has pushed aggressively for scientific output across many fields.
These findings align with what Yunji Chen observes in China. Chen, director of the National Key Laboratory of Processors in Beijing and co-founder of AI chip design company Cambricon, pointed out that while China’s chip design is advancing, its manufacturing capacity lags—partly due to U.S. export controls.
Since October 2022, the U.S. Department of Commerce has restricted the sale of advanced chips and manufacturing equipment to China, citing concerns that China uses AI to surveil citizens and modernize its military.
However, Chinese research is making a major academic impact. Chen noted that his group’s work on deep-learning architectures has amassed over 10,000 citations on Google Scholar. While the U.S. still holds the largest share of publications in this field at 41%, China is rapidly expanding its influence.
The ETO team also identified key growth areas in chip research, particularly neuromorphic computing—architectures inspired by neuron biophysics—and optical computing, which transmits information via light instead of electrons. China is leading in both subfields in terms of publication volume.
These speculative technologies hold promise for AI if they can be commercialized, China is heavily experimenting with next-generation technologies, and once they reach commercial viability, suppressing them will be difficult for the U.S. If China succeeds in commercialization, it’s not just about catching up—it’s about potentially leaping ahead.
Source: Nature, VnExpress International
