China’s semiconductor gold rush: A reality check
- Even with enough capital investment, it might take the better part of a decade for China to catch up with TSMC
- Just a handful of Japanese companies dominate the global market in silicon wafers, photoresists, and essential packaging chemicals
The semiconductor industry has become a major front in the US-China tech war. Given that this is a weak link in China’s otherwise impressive technology stack, the US has imposed export controls restricting Chinese semiconductor companies from accessing key equipment, software, and intellectual property.
China has been aware of this weak link. Since 2014, the Chinese state set up two “Big Funds” to raise money primarily from government bodies – the finance ministry, state-owned enterprises, local governments – and invest it in upcoming semiconductor companies.
The government-led investment has led to a “crowding-in” of private enterprise. More than 13,000 Chinese enterprises registered as semiconductor companies in the first nine months of this year, according to Qichacha, a database of information on Chinese companies. That is a significant jump from the nearly 9,000 companies that registered last year. Such is the lure that several new entrants with no prior experience in semiconductors have thrown their hat into the ring.
Does this silicon rush mean that China will become self-sufficient in semiconductors soon? Not quite. China’s state-backed funds may well spur private investment, even producing a few champions, but such moves are unlikely to result in a self-sufficient Chinese semiconductor industry any time soon.
That’s because the semiconductor supply chain is complex. It can be roughly split into three main stages: integrated circuit (IC) design, semiconductor manufacturing, and assembly and testing. The returns of investment for China in each of these stages will be different due to their unique limiting constraints.
Why semiconductors are important in the US-China tech war
IC Design: The first stage is skill-heavy and asset-light. Coming up with newer chip architectures and integrating them requires a large number of high-skilled engineers familiar with specialised software. There were nearly 138,000 design start-ups in China as of September 2020, most of them being registered after the first government-backed fund was announced in 2014. Though there are fears of a talent shortage, government-backed massive investment means that such a shortage could be tackled by poaching talent from Taiwan at higher wages.
The fortune of design shops is on the rise, yet they are not destined for success. One critical bottleneck remains in US control: software. IC design uses Electronic Design Automation (EDA) software that requires massive R&D investment and in-depth knowledge of chip fabrication. As a result, this market has rapidly consolidated over the years.
Currently, there are three dominant global players, two of which are US-owned (Cadence and Synopsys) while the third (Mentor Graphics) is based in the US. Exploiting this asymmetry, the US banned the sale of US-origin EDA software to HiSilicon, Huawei’s IC design unit. With China having no EDA firm at the cutting edge and the US taking a hardline approach on IP piracy, developing self-sufficiency in EDA tools will require considerable time.
The second bottleneck is processors for high-end mobile phones. ARM Holdings is the dominant firm here with over 90 per cent market share. If Nvidia’s acquisition of ARM is approved by regulators, the technology could potentially be subject to American export restriction laws that are being used to target Huawei. But this will depend on an analysis of each product and whether it is deemed to have US-origin tech by the relevant US authorities.
To overcome this dependence, Chinese companies such as Alibaba (owner of the Post) have thrown their weight behind a rival open source architecture known as RISC-V. However, RISC-V still has many years of catching up before it can displace ARM’s core general purpose processors in mobile phones and tablets.
US restrictions could hamper China’s plans for a semiconductor sector
Chip fabrication: Physically translating IC design onto silicon is a capital-intensive process requiring regular and massive capital infusion. Although the Big Funds will help in the significant investment needed for building new fabrication plants, two challenges will constrain China’s hand in the catch-up game.
First, China’s starting point remains a handicap. Even Semiconductor Manufacturing International Corporation (SMIC), China’s national champion in this space, can commercially produce 14-nanometer chips as of now. This is a couple of generations behind the industry leader TSMC’s 5-nanometer capability. This number is indicative of the transistor size; the smaller this number, the more transistors can be packed in the same area leading to higher device performance.
Even with enough capital investment, it might take the better part of a decade for China to catch up with TSMC. And by the time that happens, the cutting edge would likely have moved on to a more advanced process. Until then, companies such as Huawei have no local option for manufacturing their high-end chips.
Second, a fast follower approach is challenging to execute with ongoing US sanctions. For instance, the US has been applying pressure on ASML, the world‘s only producer of the lithography technology required to manufacture high-end chips. Applied Materials and Lam Research, two US companies which supply manufacturing equipment to SMIC, also face US export curbs. Without access to this equipment, SMIC will not be able to produce advanced chips.
Assembly & testing: Finally, the so-called “back-end”, commonly referred to as Outsourced Semiconductor Assembly & Test (OSAT), is a labour-intensive sector with lower profit margins.
OSAT has already moved to countries that have a comparative advantage in the availability and cost of labour. The capital investment is still significant, but not as high as the manufacturing process. These characteristics make OSAT a low hanging fruit for China’s Big Fund investments. Although Taiwan remains the leader in the space by a significant margin, China’s market share has been growing steadily.
China faces an additional geopolitical challenge in chip fabrication and assembly. Just a handful of Japanese companies dominate the global market in silicon wafers, photoresists, and essential packaging chemicals. These companies are well-regarded for their high-quality production capabilities and their products are not easily replaceable even by a manufacturing heavyweight such as China. In a changing world where strategic concerns are guiding technology flows, China’s chip ambitions can be foiled not just by the US but also by Japan and Taiwan.
Judging the Big Funds’ success by the number of registering companies does not provide an accurate picture. Throwing money at the problem is not a solution that can work for all sectors. Crucially, China’s “wolf warrior” approach to international diplomacy has resulted in significant geopolitical barriers for its fledgling semiconductor industry.
To adapt William Gibson’s oft-repeated quote, the future of semiconductors in China is here – it’s just not very evenly distributed.
Pranay Kotasthane and Rohan Seth work on hi-tech geopolitics at the Takshashila Institution, an independent centre for research and education in public policy.
