Competition for chip production will affect Japan, Korea and Taiwan, and Nvidia, Samsung or TSMC
No product has played as prominent a role in shaping the world economy and the balance of military power as semiconductors. Yet for years, this $556 billion industry has attracted scant attention from governments in Washington, Tokyo and other capitals of the developed world. Lately, chips have become a battleground in the competition between the US and China. Big companies and other countries will be affected by the fight.
Few other parts of the economy are so dependent on so few companies, says historian Chris Miller in Chip War: The Fight for the World’s Most Critical Technology (The chip war: the fight for the world’s most crucial technology). Taiwan Semiconductor Manufacturing (TSMC) makes almost all of the world’s most advanced microprocessors. ASML in the Netherlands has a de facto monopoly on the UV lithography machines needed to make the most sophisticated circuitry. Two South Korean giants dominate the memory chip market; three US companies control the semiconductor software.
These so-called choke points are a feature of a hyper-efficient industry that can produce more than a trillion units a year. According to Miller, a professor of US and Russian foreign policy, it is no accident that the US and its allies control most of them. After World War II, pioneering companies like Fairchild Semiconductor, Intel, and others from Silicon Valley cemented America’s technological supremacy.
The subsequent drive to outsource US manufacturing abroad coincided with the country’s efforts to deepen trade and investment links with Japan and the rest of Asia. Cheap and abundant labor allowed companies to reduce their costs; Asian leaders touted better-paying jobs and economic growth; and Washington integrated its allies more deeply into the US economy. In the late 1970s, companies like Intel and Texas Instruments employed tens of thousands of workers in South Korea, Taiwan, and Southeast Asia.
Over time, however, some Asian manufacturers amassed enough supply chain experience and scale to challenge U.S. dominance. South Korea; Taiwan today boasts not only the world’s largest contract chipmaker, but also the world’s leading companies that assemble, test, and package chips.
Andy Grove, the former head of Intel, rightly warned that abandoning “commodity” manufacturing could lock manufacturers out of future emerging industries. Now the former US pioneer is struggling to catch up with $356 billion TSMC and Korea’s Samsung in chipmaking. Meanwhile, natural disasters and the pandemic have exposed the fragility of the global supply chain, which Miller describes as “a perfect picture of globalization gone wrong.”
The combination of weakening leadership in the chip industry and heightened awareness of supply chain vulnerability underpins US technology policies. Chips and Science Act It is providing $53 billion to bring semiconductor development and manufacturing back to the US. China, for its part, has long identified US control over supply chains as a national security threat, and is spending hundreds of thousands of million dollars to wean themselves off foreign technology. US sanctions against Huawei have accelerated these efforts, reminding Miller that chokepoints “don’t last indefinitely.” In response, Washington has increased trade and investment restrictions: In early September, it banned AMD and Nvidia from exporting some advanced AI chips to China.
This intensifying rivalry puts Taiwan, South Korea, and Japan in an awkward position. All three depend on China as their largest trading partner. But if the US is successful in relocating advanced chip manufacturing, the market shares of one or all of its allies must decline, Miller argues. Multinational giants with exposure to China will also be caught in the middle. Nvidia estimates that some 400 million in sales are at risk. Apple’s plans to use Chinese-made chips have drawn scrutiny from Washington.
For South Korea’s Samsung and SK Hynix, the dilemma is even more serious. Both hope to expand in the US, but a provision of the Chips Act states that, in order to receive US subsidies, they will be prohibited from expanding or improving chip production capacity in China for 10 years. The two companies currently produce 20% and 40%, respectively, of their memory chips in the People’s Republic, according to Nikkei.
Miller exposes even more aggressive tools that Washington and Beijing have yet to deploy. The former could pressure TSMC to deploy its new technologies simultaneously in the US and Taiwan, or force it to invest more in the country. China could pressure foreign companies to transfer technologies to local peers, or require firms like Apple to buy local components.
The biggest escalation, of course, would be a military conflict between China and Taiwan, whose sovereignty Beijing claims. Disabling the island’s semiconductor production facilities would be “catastrophic” for the global economy, Miller writes, as the world would produce 37% less computing power as a result, and it would take at least half a decade to restore capacity. .
This is an extreme scenario, and investors and tech bosses are betting that the high stakes will deter Washington and Beijing from escalating tensions. That seems like an illusion.