Huawei, the Chinese smartphone giant, has created ripples within the strategic and business community with its newly unveiled Mate 60 Pro which houses the Kirin 9000 processor. The chipset reportedly used Semiconductor Manufacturing International Corp (SMIC)’s second-generation 7nm fabrication technique, thereby demonstrating China’s capability to manufacture a 7nm chip.
Challenges before China’s quest
Consequently, observers have claimed that the capability marks a major breakthrough in Beijing’s drive to attain self-sufficiency in manufacturing advanced chips. The fact that China succeeded in achieving this feat despite American sanctions on key semiconductor technologies has led many to even question the efficacy of the sanctions. However, while China’s technology demonstration deserves appreciation, the capability itself might not mean much as several challenges besiege its self-sufficiency drive.
To begin with, the fabrication technique used by Huawei-SMIC to manufacture the Kirin 9000 processor is highly inefficient. The wafer yield (a metric of efficiency) of the deployed technology is way less than 50%. In contrast, Taiwan’s Taiwan Semiconductor Manufacturing Company Limited (TSMC)’s 7nm fabrication technique has a wafer yield in excess of 90%. This makes the SMIC’s process extremely expensive — up to 10 times the costs incurred by other players in the market, and therefore highly uncompetitive.
Second, the 7nm fabrication technique represents the zenith of China’s capabilities with the available Deep Ultraviolet (DUV) lithography tools. The United States’s sanctions that cut off Beijing’s access to the most advanced lithography tool in the market — the Extreme Ultraviolet (EUVs) — meant that China had to rely on DUVs to fabricate the Kirin 9000 chipset. While DUVs can technically be used to make 7nm chips, the process is extremely messy and inefficient, thus lowering its yield. For instance, the SMIC technique used multiple rounds of masking or layering on the wafer to manufacture a 7nm chipset, leading to multiple exposures. On the other hand, the TSMC with the EUVs can perform the same task of high complexity with a single exposure.
Third, it is doubtful that Huawei-SMIC could produce the current chipsets on a large scale. The fact that the U.S. and its allies have restricted China’s access to even DUVs lately means that large-scale production of 7nm chips would be a challenge for Chinese companies.
Thus, low yield rates, inefficient and costly procedures along with difficulty in achieving the scale are likely hurdles in Huawei’s attempt to commercialise its new technology product. This is significant because the failure to achieve commercialisation will impact incremental innovation as they reinforce each other. And products that fail to innovate alongside their competition eventually fade away.
The U.S. and China systems, a comparison
Besides, there are several other challenges that plague China’s chip ecosystem compared to the U.S.-led ecosystem. The fact that America’s encompasses the most advanced economies of the world confers upon it several advantages that China’s isolated ecosystem will find nearly impossible to compete with.
First, the extensive and distributed nature of the U.S.-led tech ecosystem allows individual countries to achieve functional specialisation according to their respective comparative advantages. The existing supply chain — where the U.S. specialises in EDA tools and designing, the Netherlands in producing lithography tools, Japan in manufacturing specialised materials, and Taiwan and South Korea in fabrication — corroborates the claim.
China, on the other hand, not only has the mandate to become self-sufficient in each segment of the value chain but has to also achieve sophistication in each of these to remain competitive. Achieving specialisation in any one segment of the chip value chain itself is highly capital intensive; to achieve so in each of them is impossible. It is also important to remember that advanced chips are only one of the many core technologies that China aims to become self-sufficient in. Given that the Chinese ecosystem is not as elaborate as the U.S.’s and is rather isolated, there are limits to Beijing’s potential with its finite resources.
As for the U.S.-led ecosystem, the costs can be distributed among the participating countries, most of which have much higher per-capita income than China. Therefore, to compete with a larger pool of resources, China will need to strike a higher success rate on every dime it spends on research, which is difficult to achieve given that breakthroughs in basic research are capital intensive and may not yield success as often. If that alone is not enough to burden the scientific community, the pressure to perform and deliver in China’s authoritarian system further compounds the problem. To sum it up, China’s appetite to absorb failures is extremely limited when compared to the U.S. and its allies. This makes for quite an unconducive environment for innovation given that it thrives best in a free environment. Finally, the U.S.-led tech ecosystem allows it to source talent from diverse regions, owing to its open immigration policy and distributed network. China, on the contrary, will have to solely rely on its national or overseas talent pool as the movement of human capital to China becomes more difficult due to deepening rivalry in the high-tech sectors.
Going forward, China’s tech ecosystem faces a daunting challenge to succeed in everything all by itself. It may score a victory in odd areas, but replicating the feat in every single domain is impractical.
Amit Kumar is Research Analyst, Indo-Pacific Studies Programme, the Takshashila Institution, Bengaluru
