The U.S. can no longer claim the top spot in published and accepted research papers on chip design and leading-edge technologies. Following the latest entries for the International Solid State Circuits Conference (ISSCC 2023), one of the world's most renowned events on semiconductor circuits, it has become clear that particular accolade now belongs to China - despite a prolonged climate of heavy technological sanctions being levied against the country.
With 59 accepted papers from China against 42 out of North America, the U.S. can no longer claim supremacy in advanced chip research. Part of that reason stems from the fact that the U.S. submitted 30 fewer research papers for ISSCC 2023 than it did for the preceding edition, while China increased its number of submissions. In fact, at 129 accepted papers out of the 198 total, the entire Far East region, which comprises East and South East Asia, now has almost double the submitted (and accepted) papers compared to the EU and the U.S. combined (27 and 42, respectively).
At the same time, ISSCC's panel says that China's research contributions have been increasing in quality at an impressive clip, with increasingly large numbers of its submitted papers surviving the cutting room floor, especially since 2017. So the numbers are more significant, but so is the quality. These aren't papers being hacked together by something like Meta's Galactica AI.
Interestingly, ISSCC also announced that the sources submitting papers within each region have shifted in recent years, with increasingly large numbers of submitted and approved papers originating from universities. Interestingly, in 2011, there was a 50-50 split between corporation and academia submissions. However, as of the 2023 edition, around 75% of the submissions originated in academia, showcasing that universities have become increasingly important in developing cutting-edge tech.
China's Investment in Higher Education Pays Dividends
The number of cutting-edge research papers originating from Chinese universities is nothing to scoff at. Out of the 49 accepted papers out of China, 15 hailed from the University of Macau, 13 from Tsinghua University, and six from Peking University. Tsinghua University’s research antics, for one, have been covered at Tom’s Hardware; one relates to precisely the technologies that ISSCC is looking for, such as the world’s smallest transistor - the Sidewall design. So it’s not by chance that it’s ranked the third most prestigious university in Asia.
You may also remember that Tsinghua Unigroup, one of the many arms of Tsinghua University, owns NAND-maker YMTC - one of China’s crown jewels and a heavy target of U.S. sanctions.
As quoted by Business Korea, one of the attendants said, “China increased its selected research paper count in every category, and the Chinese government played an important role in this.” The funding history for Chinese universities tells an interestingly similar story; over the last decade, the Chinese government’s funding for higher education has more than doubled. It’s currently estimated at $179 billion and has been seeing increases yearly, with several universities receiving over $5 billion in funding.
To put that number in perspective, Cerebras, one of the world’s leading-edge chip designers, has received around $720 million in funding from its various capital raises. However, the U.S. still spends more on higher education funding as a percentage of GDP (Gross Domestic Product, or the total amount of value in goods and services produced by a country’s economy). For example, according to data from the World Bank, the U.S.’s expenses in 2018 totaled 4.9% of GDP, while China’s number for the same year was 3.8% (in 2012, China’s highest value only reached 4.1, in contrast to the U.S.’s 5.1% in 2017).
Reducing the Impact of Sanctions
It also bears mentioning that China's investments in university research (and research paper hegemony) are geared toward the tomorrow of technological development. China isn't researching the present in its universities; it's researching its future. And its end, if the current geopolitical situation is anything to go by, passes by reducing the impact of the US's sanctions as much as possible. But, of course, that can only be done if China isn't technologically dependent on the US, to begin with, and that's precisely the problem future chip fabrication will solve. The US can sanction current technology imports, but it can't forbid China from developing anything. The key here is the speed in unlocking technologies such as Sidewall transistors, quantum computing, and AI. And the higher the quality of research, the higher the rate.
Somewhere, in a now speculative future, there's a chance China could even close - or begin to close - some doors on US technological development itself. Research frequently leads to patents. And patents can sometimes occupy a lot of efficient design space for a very long time. In truth, China merely wants to hit that future faster. What happens after that is anyone's guess.
