Taiwan makes 90 percent of the world’s AI chips. If it stops for 30 days, the world loses $10.6 trillion.

The world spent the last 50 years making chips smaller. We are about to learn what happens when geopolitics gets measured in nanometers (nm). No one is considering the paradox. Governments worldwide have designated $380 billionto become independent in semiconductors. In doing so, they are making themselves more dependent on a handful of chokepoints, from Dutch machines to Japanese chemicals to American software. 

Everyone thinks the danger is that China will not be able to make advanced chips. The real danger is the opposite: China is learning to live without them while investing to fill the gap. It is mass-producing the mature-node chips that power 80 percent of the world economy, cars, appliances, etc., and by doing so, it is eroding the silicon shield that was supposed to keep Taiwan safe. The next crisis will not be about who has the best 2-nanometer chip. It will be about who controls the 28-nanometer one.

TSMC in Taiwan makes over 90 percent of the chips below 7 nanometers. The tiny ones. The ones that actually run artificial intelligence (AI), that go into iPhones. If the fabs in Hsinchu stopped for 30 days, factories in Detroit and Stuttgart would idle. There is no backup. Washington and Brussels know it, as do Tokyo and Seoul. They have heavily invested in the CHIPS Act, the EU Chips Act, and subsidies in Japan and Korea. The plan was to bring production home. What happened instead is we added more steps, more cost, but not more resilience. We still have one island that makes the most important product on earth. That gives Taiwan something no supplier normally has. It is not leverage from what it owns. It is leverage from what no one else can do. Even with unlimited money, China needs at least half a decade to replicate TSMC’s advanced technology for microchips below 7nm. You require the machines, skilled operators, and nearby chemical suppliers. Taiwan has long surpassed the role of just a supplier; it has evolved into a strong strategic deterrent known as the silicon shield. This fundamentally alters market dynamics, shifting the focus from purely commercial ties to national security considerations. 

Interdependence was once reliable; today it has become so concentrated that it poses a risk. Trade in the 1990s spread the risk across multiple interchangeable suppliers. Nowadays, the supply chain often depends on a single node, posing a powerful threat to global markets. Countries move from dependency to dependency; sovereignty in one supplier is bought with exposure to the next. The race for microchips is not creating a foundation for independence but a fragile trade environment.  

China’s Growing Dominance in Mature-Node Semiconductors

In the 1990s, Europe and the United States had the highest semiconductor manufacturing capacity. Taiwan increased its capacity at the beginning of the century and has maintained the highest position. China has shown steady growth in microchip manufacturing capacity and is expected to be the world’s leading manufacturer of mature nodes like 28nm by2030, surpassing the United States and Taiwan, with over 30 percent of global capacity, but will still be behind TSMC at <7nm without ASML’s extreme ultraviolet (EUV) lithography machines. Beijing doesn’t need to beat TSMC to neutralize the silicon shield; it just needs to make the shield irrelevant for its own economy.

Europe has a trend particularly close to the United States, with both selling their dominant production positions in the supply chain. Extreme capital intensity, low efficiency, expensive labor, and low profits led the Western countries to sell their manufacturing capacity while keeping semiconductor design and innovation. The assumption that having a supplier contract was as good as being one is no longer true. It is no longer a cheap supply chain but a chokepoint, a boundary, and a weakness for Western countries.

The pattern was the opposite for China. While Western countries saw manufacturing as a liability, Beijing saw it as infrastructure to be owned. It was a long-term strategic return, Chinese leverage over global supply chains and insurance against trade restrictions. After the American export controls in 2018, China’s global chip capacity increased exponentially and so did the number of patent applications. Public policies, such as state subsidies and tax incentives for chip manufacturers, were applied to improve research and development (R&D) and expand the domestic market for microchips. 

Research and development is a key driver for long-run growth. Taiwan, China, and South Korea are the manufacturing leaders in semiconductors, and this results from high levels of R&D investment. China had the highest growth in innovation investment, followed by South Korea and Taiwan. 

For the past decade, China has invested aggressively in the technology industry. This is not a coincidence. US curbs on advanced microchips pushed China to focus on its own production and innovation. China’s target is not to beat TSMC, but to no longer need it. With China’s exponential growth, it is expected that by 2030 China will dominate mature-node manufacturing. The hundreds of billions of dollars invested in domestic semiconductor production will build chip self-reliance and strengthen R&D. Closing the knowledge gap and increasing patent applications in semiconductors is how China aims to achieve capability parity. 

The US-China Semiconductor Technology Gap

Real technological leadership depends on the ability to manufacture the newest semiconductors reliably. In advanced microchips, China is generations behind TSMC, encountering higher marginal costs and lower profitability. This stems from Taiwan’s control of the newest processes. The recent increase in Chinese patents shows a narrowing knowledge gap that will affect manufacturing industries. 

Washington wants to keep its lead in innovation, Beijing wants to end its dependence, and the chip supply chain is where that contest is now fought. The chokepoints stem from high-capital industries where fixed costs are so high that only one firm can afford to produce them.  It is no longer just about who makes the fastest chip, but who controls the chokepoints: one company in the Netherlands (ASML) makes 100 percent of EUV machines for $200 million each, and one island for fabrication. A single disruption is not just a tech problem. It hits car prices, phone availability, and inflation for everyone.

The pace of miniaturization represents an exponential increase in capital demand. In 2004, the most advanced microchips in mass production used a 90-nanometer process, and today the research center IMEC is exploring the possibility of nodes below one nanometer, requiring even more capital and specialized labor. As innovation progresses, the supply chain becomes more capital-intensive, creating further chokepoints and instability. The United States remains a pioneer in microchip research and development; however, mass production has migrated to East Asia. Taiwan is increasing its global presence in the supply chain and controls 60 percent of the world’s contract chipmaking capacity.

Semiconductor Supply Chains Cannot Escape Interdependence

Manufacturing countries control microchip supply chains; however, leadership is defined by those able to manufacture the newest semiconductors. The number of patent applications for semiconductors or R&D spending can measure a country’s innovation progress in technology. 

These high levels of investment over the last two decades have reduced the technological gap between East Asia and Western countries. China is a benchmark in semiconductors and high-tech development. However, R&D investment is necessary for sustained growth, but not sufficient. It must be paired with access to machinery and materials, rare metals, and specialized labor. China became an example of a high-investment country that cannot manufacture leading-edge semiconductors because another country controls the machinery needed.  

China’s Semiconductor Rise Threatens Taiwan’s Silicon Shield

Chinese capability parity poses a future threat to Taiwan. The silicon shield will lose its power if China can meet its own demand for mature node chips domestically. The possibility of disruption causes political and financial instability in Taiwan. In the event of a natural disaster, an unexpected geopolitical shock, or even a Chinese military invasion, it would paralyze TSMC’s operations and affect all sectors of international trade. A full war over Taiwan would have a greater impact than the 2008 financial crisis, with global GDP decreasing by approximately 9.6 percent and financial losses of $10.6 trillion in the first year alone. Samsung would win a significant share of the market, since it is the only non-Taiwan company that can mass-produce 5nm microchips. 

What comes next? One built by China that will be good enough. One built by America, Taiwan, Japan, South Korea, and Europe that will be cutting-edge but insanely expensive. Both will need their own factories, machines, and chemicals. That’s why prices will go up, not down. Brussels thinks it needs $120 billion by 2035 just to keep a seat at the table. We call it the silicon shield because we like the idea that dependence will keep Taiwan safe. The logic only holds as long as both sides need the island: China for its chips and Western countries for its factories. Beijing probably won’t make a 2-nanometer chip like TSMC in the short term, but it doesn’t need to. If it can produce enough decent chips, the 28-nanometer kind used in most cars and other consumer goods, the shield stops being a shield. It becomes the thing that traps us.

About the Authors: Bruno S. Sergi and Mariana P. Gameiro

Bruno S. Sergi is an instructor at Harvard University’s Sustainability and Global Development Practice Graduate Programs. He is also affiliated with the Harvard Center for International Development, the Davis Center for Russian and Eurasian Studies, and the Harvard University Asia Center. He has led the launch of several scholarly journals and book series, including the Cambridge Elements series at Cambridge University Press and Entrepreneurship and Global Economic Growth at Emerald Publishing.

Mariana P. Gameiro is a teaching assistant in microeconomics and macroeconomics at Católica Lisbon School of Business and Economics, where she is completing a Master’s in Economics as part of a double degree in Firms Strategy with the Université Catholique de Louvain. Her research focuses on microeconomic theory and labor economics, including hiring models and wage setting under signaling and monopsony.

The post Why Taiwan’s Silicon Shield Is Failing appeared first on The National Interest.