A new concept is on the lips of everyone in China's semiconductor industry: the "Post-Moore's Law Era." It's a term that conveys not frustration, but rather hope for China's ability to someday surpass the West's indigenous chip capacity.
Since its initial mention by Chinese Vice Premier Liu He at a meeting on innovation in digital electronics in May, it has been quoted with regularity by Chinese academics, municipal governments, industry journals, private companies and other observers concerned with China's lackluster standing in the global chip sector.
The idea is simple: By 2025, scientists believe, advances in chip technology will no longer keep up with Moore's Law. Coined by Intel co-founder Gordon Moore, the rule states that transistors per unit area double every one to two years. While Moore's prediction has so far proven accurate during an era of dramatic improvements in chip performance, in the future engineers expect to hit the physical limits of existing chip materials, such as silicon. That will result in a drop-off in chip technology advancements unless alternatives are devised.
The idea that dominant chip makers — who are overwhelmingly foreign — face a near-certain obstacle has become a beacon of hope for Chinese firms, which analysts say lag at least one to two wafer generations behind their competitors. For example, research conducted by the U.S.-based think tank Information Technology & Innovation Foundation (ITIF) suggests that, in logic chips, Semiconductor Manufacturing International Corporation (SMIC), China's largest chipmaker, manages scale production at the 14nm level, whereas South Korean Samsung and Taiwan-based TSMC are able to accomplish 5nm. On the memory chip side, U.S.-based Micron and Samsung lead with 10nm wafers, while China's best, Changxin Memory Technologies, only manages 19nm.
To pin down the next fundamental semiconductor material, whatever that may be, could give Chinese firms a shortcut to competitiveness or even allow them to leapfrog foreign competitors.
To this end, Beijing has announced it will research alternative substances to help its advanced chip technology drive.
In August, China's Ministry of Industry and Information Technology unveiled plans to include "carbon-based materials" in its plan to achieve "breakthrough" technologies. Namely, the ministry will incorporate carbon fiber, graphene, silicon carbide and other carbon-based composites into the country's 14th Five-Year Plan's raw materials development strategy as well as its 14th FYP's scientific innovation strategy.
It marks China's highest-level backing of this research to date, after various industry observers, including the MIIT's own affiliated technology journal, discussed the possibilities such materials presented in a series of articles exploring the "Post-Moore's Law Era."
"We need more breakthroughs in theoretical research, especially in the fields of compound semiconductor and material science, which are led by the U.S. and Japan," Huawei founder Ren Zhengfei said during a private speech at an August innovation summit.
"If we pursue only what's practical, we may forever lag behind," Ren said.
Not tried and true
Even with support from the central government, there's no guarantee China will pull ahead in semiconductors. For one, pursuing unproven technologies would mark a drastic new business model for China's wafer companies, for whom R&D has not traditionally been a priority.
For one, the country continues to throw money at manufacturing over design. During its first phase spanning 2014 to 2019, the giant, state-led National Integrated Circuit Industry Investment Fund, which raised a total of RMB 138.7 billion (USD $21.7 billion), spent two-thirds of its invested funds on fabrication companies. According to German think tank Stiftung Neue Verantwortung, as of mid-2020, less than 1% of the fund's investments were made in domestic electronic design automation (EDA), software tools used to draw up new chips.
Similarly, according to ITIF, R&D at Chinese semiconductor companies accounted for just over 8% of sales in 2018, in contrast with U.S. firms' 18%, EU companies' 14% and Taiwanese firms' 10%. Likewise, Chinese companies held just over 6% of the semiconductor patents granted by the United States Patent and Trademark Office in 2018, a fraction of those granted to firms based in the U.S., Japan, Taiwan and South Korea, with 29%, 23%, 17% and 14% respectively.
"If you're getting 40% of your revenues, as SMIC is, from the Chinese government, then you have a source of income that isn't as reliant or dependent upon you innovating the next generation of products," ITIF Vice President of Global Innovation Policy Stephen Ezell said during an online seminar in July.
Yangyang Cheng, a postdoctoral fellow at Yale Law School who researches the history of science in China, echoed this sentiment. "This kind of hardware development takes a very long time," she told Protocol.
Making reference to China's history of persecuting scientists and intellectuals, especially under Mao Zedong's rule, Cheng said that the country has hamstrung its own progress relative to those that never stopped chasing tech advancements.
"If there's a few decades being lost because of political campaigns or different kinds of strategies, then this is time that will be paid for later," she said.
What's more, Cheng said, the emphasis on speed disincentivizes unglamorous fundamental research in favor of sensational claims, such as in the case of a distinguished, young Chinese computer scientist who declared he had created one of the country's first microchips in 2003, later discovered to be based on a design by Motorola.
"These seem like isolated incidents, but … [they] speak to some systemic issues and play into the current state where China does lag behind," Cheng said.
To be sure, China continues to court foreign companies despite the decoupling rhetoric, export restrictions and scrutiny in merger and acquisition attempts from a growing number of Western governments, calling into question the vehemence of the new R&D push.
Data from the Chinese Ministry of Commerce shows that, in 2020, China's overseas tech investments jumped. According to its September report on last year's outbound foreign direct investments, spending on information transmission, software and IT service industries grew 67.7% year-over-year to $9.19 billion, while investments in overseas scientific research and technical service industries also grew 8.7% YOY, to $3.73 billion.
At home, meanwhile, the MOC is pressing foreign companies, including Intel, Germany's Infineon, French-Italian STMicroelectronics, U.S. chip industry association SEMI and others on cross-border collaboration. Items on its wishlist include core IP, tech transfers, EDA, packaging, production, training and more.
These kinds of partnerships, along with trends such as talent acquisition from foreign firms, universities and Silicon Valley remain gray areas where governments are slow to respond, according to Dylan Patel, chief analyst at SemiAnalysis, who characterized some of these efforts as "poaching."
However, given the high precision required of semiconductors, making disruption by carbon materials potentially "10 years away forever," the battle for the foreseeable future will remain in the realm of existing chip technologies that are not exclusive to the most advanced companies, he said.
"The majority of chips made are not bleeding edge, so you can eat that value chain and move up," Patel said, citing China's continued appetite for chip manufacturing. "For China it's very important to transition from the current [industrial society] to services, to an industrial powerhouse. This sort of transition needs to happen or the economy will stagnate, so these sorts of investments need to be made."