BENGALURU: The U.S.-China geopolitical fault lines in high technology have become sharper than ever and nowhere has it been more significant than in the domain of semiconductors. Semiconductors are critical to advanced computing and artificial intelligence (AI) that will form the core of the next-generation industrial revolution. They also have significant implications for advanced and precision weaponry in addition to consumer tech.
The deepening of U.S.-China geopolitical rivalry has rendered the widely distributed global semiconductor value chain susceptible to disruption and choking. Consequently, their value as a critical tech has led to an increased securitisation of the domain. The incessant urge to maintain a competitive edge over the adversary has led to efforts towards creating a closed tech ecosystem through attaining ‘self-sufficiency’ and ‘technology access control’, particularly by the U.S. and, to an extent, China.
U.S. Preponderant Position In Chip Value Chain
While the global semiconductor value chain is too complex and too widely distributed for any self-sufficiency efforts by a single country to succeed, the U.S. holds a preponderant position in the chip ecosystem for several reasons. The most significant of all pertains to the U.S. dominance in the chip designing ecosystem. Not only are the majority of chip-designing IP rights held by U.S.-based companies such as AMD, Qualcomm, Broadcom, and NVIDIA, the U.S. also has a near monopoly in the Electronic Design Automation (EDA) market. Firms that engage in EDA, test chip designs for their performance in real-world scenarios through simulation and verification using both software and hardware. Without EDA, the cost of error in chip manufacturing would be catastrophic as chips once fabricated cannot be reworked.
Another factor that bolsters the U.S. predominance in the chip industry is the specialised leadership of its allies in various segments of the value chain. For instance, the Netherlands-based ASML along with Japan’s Tokyo Electron, Nikon, Cannon and Nuflare are the sole manufacturers of the specialised lithography tools that are used in fabricating advanced chips. Of them, ASML enjoys a near monopoly in manufacturing the most sophisticated lithography tools that uses Deep Ultraviolet (DUV) and Extreme Ultraviolet (EUV) technology. Manufacturing of chips also requires a host of specialised gases and materials such as silicon wafers, a segment that Japan primarily dominates.
Lastly, in the fabrication segment, Taiwan, a key U.S. partner, produces more than 60% of all global semiconductors and over 90% of all advanced semiconductors. Taiwan’s TSMC is the market leader and accounts for around 80% of the total Taiwanese chip output. Its competitors include UMC (Taiwan), SK Hynix (South Korea) and GlobalFoundries (U.S.). It shall be noted that each U.S. ally that specialises in any one segment is dependent on U.S. technologies in some way or the other to maintain their leadership. Thus, the U.S. emerges as a seemingly invincible player in the semiconductor industry.
China’s SMIC is a minor player in this segment that specialises in fabricating low-end semiconductors.
U.S. Insecurity Rising
Yet, U.S. insecurity has grown sharply due to China’s recent advances in chip technology and efforts to attain self-sufficiency largely through unfair means, such as IP theft and forced technology transfer backed by heavy investments and government subsidies.
China’s strength in the global semiconductor supply chain is two-fold. One, it is the world’s largest market for chips produced globally (one-third of the total sales) as bulk of electronic goods manufacturing takes place in China. Secondly, almost all contract manufacturing companies have some of their fabrication capacity or foundries located in China. Even in-house semiconductor companies like Intel and Samsung, operate factories in China to fabricate their chips. Close to 15% of the global chip fabrication (China-owned and foreign-owned) takes place in China.
China has managed to leverage these advantages (as mentioned earlier through IP theft, forced tech transfer and reverse engineering) to achieve marginal gains in its effort to become ‘self-reliant’ in designing and producing lithography machines and fabricating high-end semiconductors. Amidst the deepening contestation with China, the U.S. has become increasingly wary of Beijing’s relative gain and fearful of losing the competitive advantage it currently holds in the semiconductor industry. The anxiety has arisen against the backdrop of the falling share of U.S. domestic semiconductor manufacturing capacity, from around 40% in 1990 to around 12% currently. The fall is attributed to the rise of fabless semiconductor companies in the U.S. that decided to concentrate on chip designing and outsource manufacturing to contract manufacturers or foundries. The others who still manufacture their chips, like Intel, have shifted their manufacturing operation to East Asia, where 80% of the global fab capacity is concentrated.
To address this gap, the U.S. unveiled the Chips and Science Act in August 2022 with an aim to invest $280 billion over the next ten years to boost its domestic semiconductor manufacturing and R&D capability. In October 2022, the U.S. also tightened the stranglehold over China by introducing severe export control measures to deny the latter access to key technologies in chip making without government approval. Building on these measures, in December 2022, the U.S. Commerce Department added 36 of China’s semiconductor-related companies including the state-owned YMTC to its ‘entity list’ that makes Beijing’s access to critical technologies extremely difficult. It also applied the foreign product rule to 21 of the 36 entities that prohibits American companies from exporting products that contain a specified proportion of U.S. technology to China. Finally, the U.S. struck a trilateral deal with the Netherlands and Japan in January 2023 to restrict China’s access to advanced lithography tools.
The U.S.-led efforts to exclude China from its innovation ecosystem have left Beijing anxious. In response, China has lodged a complaint against the U.S. in the WTO on grounds of unfair trade practices. Meanwhile, SMIC and YMTC have reported a fall in their revenue and scaled down their orders for semiconductor equipment. The U.S. measures have also delayed SMIC’s expansion plans.
Facing the heat, China is aggressively pushing for self-reliance but its efforts are significantly constrained by the U.S. restrictions. As a consequence, China is pressing for capturing overseas Chinese talent in the high-tech sector by offering lucrative salaries. But even this effort has met resistance because of the exploitative work culture in China. In contrast, the work environment in the West is more appealing and has therefore prevented any material shift in this regard until now.
Two Separate Ecosystems?
Based on the developments so far, one likely outcome that seems apparent is the emergence of two separate semiconductor ecosystems, albeit with some overlaps. Given its ‘chip overlord’ status, the U.S. has the capability to enforce an ecosystem that is exclusive of China. The American companies have so far complied with the state’s directives. Nvidia and Intel have restructured and downgraded their operations in China in order to comply with the U.S. restrictions. Korea’s Samsung and SK Hynix have also downgraded their operations in China. The Chinese response has been to target the American companies operating in China, either by way of stalling their merger operations or subjecting them to increased surveillance under the newly reformed counter-espionage law. If anything, the Chinese response is likely to strengthen an alignment between the American state and the companies. The second category of consequent actors in realising the U.S. vision of an exclusive chip ecosystem is its allies. Here also, there seems a visible convergence as evident in the responses of Japan, the Netherlands, Taiwan and South Korea. While Japan and the Netherlands have already announced intentions to enforce the trilateral agreement reached with the U.S. to restrict the export of advanced lithography tools to China, Taiwan’s TSMC has also promised to comply with the CHIPS Act and announced plans to open a foundry in Arizona to boost U.S. domestic chip manufacturing.
If the U.S.’ plans to enforce an exclusive ecosystem materialises, China will find it extremely difficult to catch up with its rival and the gap will likely increase. The probable option China will definitely look toward is first attracting overseas Chinese talent that have hitherto been a part of the U.S. ecosystem. Efforts are already underway as multiple Chinese companies have floated lucrative offers for overseas Chinese talent in this high tech sector. In the past, there have been incidents where Chinese citizens trained in Europe and the U.S. have returned and helped establish Chinese tech companies, giving rise to challenges relating to IP theft. The most infamous case involved Zhongchang Yu, a former ASML employee accused of stealing its IP rights and later setting up two chip firms in China.
Therefore, IP protection and close guarding of human capital appears the next front for U.S.-China contestation. It will certainly begin with exclusion of Chinese personnel from the R&D field in these categories. Although the U.S. is yet to announce its plan in this regard, Marvell, a U.S. semiconductor firm, laid off its entire China R&D team in March 2023. Besides, R&D cooperation and collaboration between the U.S. and Chinese nationals have seen a steep decline in recent years. The figures for the same between Europeans and Chinese researchers have stagnated lately.
China’s advances in semiconductors in the coming years may be severely limited if its access to equipment and talent is restricted. This would mean that its domestically produced chips will not be able to compete in the market with those produced by the U.S.-led ecosystem. This would impede commercialisation of Chinese chips. And since there exists a direct correlation between commercialisation and innovation, which is seemingly true for at least the chip ecosystem that is governed by Moore’s law, China will find it very difficult to compete with the United States. Several domain experts feel that given these restrictions, China could fall at least two decades behind U.S. technology with no real chance of catching up.
Of course, there would be significant economic costs for the U.S. as well in building a chip ecosystem exclusive of China, but a portion of the associated cost could be offset by ‘friend shoring’ and support from its allies. The fact that India and Vietnam along with other Asian countries are competing to absorb the functions that are currently undertaken in China, for instance in the ATP segment, will aid the U.S. effort. Furthermore, the Indian government has rolled out a Product Linked Incentive (PLI) scheme to boost domestic semiconductor manufacturing. The EU came up with its own CHIPS Act to boost its manufacturing capacity and cut its reliance on Asian players as the U.S.-China conflict threatens to impact supply of chips from the region. These steps are likely to lessen the burden on the U.S. considering there is relatively deeper convergence between the EU and the U.S. on the issue. Anyway, there is enough momentum within the U.S. to overcome the inertia in pursuing an exclusive chip ecosystem given its increased securitisation and the remarkable capability the world is witnessing in the realm of AI lately.
(The author is a Research Analyst, Indo-Pacific Study Programme, at the Takshashila Institution. Views expressed in this article are personal.)
[/vc_column][/vc_row][/tdc_zone]
