The fight for dominance in semiconductors, the “chips” powering everything from smartphones and cloud servers to military systems, has become the centrepiece of global economic and geopolitical competition. In what’s now called the chip war, the United States and China face off in a high-stakes rivalry, with Taiwan’s unlikely dominance making it the third pillar of this new era. Semiconductors, a half-trillion-dollar industry expected to double by 2030, are the linchpin of AI innovation, military power, and global economic clout.
International Finance unpacks how Taiwan emerged as a chipmaking powerhouse, the US bid to reclaim manufacturing, and how Washington’s export bans and China’s countermoves are reshaping the global economy. Along the way, we bring insights from industry leaders and policy experts on the economic fallout and the future of the silicon struggle.
Global semiconductor supply
Just 100 miles from China’s coast, Taiwan’s TSMC (Taiwan Semiconductor Manufacturing Co.) produces roughly 90% of the world’s most advanced semiconductors. These chips power everything from Apple iPhones and Nvidia AI accelerators to critical infrastructure and defence systems.
Taiwan’s dominance, especially at the smallest transistor sizes, makes it the linchpin of the global tech supply chain: a phenomenon sometimes called the “silicon shield.” The logic is simple. Taiwan’s role in chip supply makes military conflict an economic catastrophe for everyone involved, acting as a deterrent to aggression.
TSMC’s ascent was decades in the making. Founded in 1987 with state support, TSMC pioneered the “pure-play” foundry model, producing chips designed by others and steadily outpacing global rivals.
Today, its technical know-how lets it pack billions of transistors onto fingernail-sized chips, years ahead of competitors. Until recently, nearly all these leading-edge chips were made in Taiwan, a concentration that inspires both awe and anxiety.
On one hand, TSMC is an economic and strategic bulwark for Taiwan, seen as a “sacred mountain protecting the country.” On the other hand, it creates a single point of failure: a natural disaster or geopolitical event could disrupt the world’s chip supply, with devastating consequences.
Policymakers worry about what will happen to TSMC’s “fabs” if China ever attacks or blocks Taiwan. The stakes are enormous: advanced chips are critical for civilian technology and national defence.
Even within Taiwan, there’s anxiety over how much chip technology should be shared abroad. Morris Chang, TSMC’s 91-year-old founder, has called out the dilemma, “The US Commerce Secretary said repeatedly that Taiwan is a very dangerous place [and] America cannot rely on Taiwan for chips… that, of course, is Taiwan’s dilemma.”
While TSMC is expanding overseas, Chang notes that “in the chip sector, globalisation is dead. Free trade is not quite that dead, but it’s in danger.” His warning is that higher costs and less ubiquity for advanced chips will result if the world splits into competing tech blocs.
Taiwan’s role as a global chip linchpin brings leverage and vulnerability, a reality now pushing others to develop their own advanced chipmaking muscle.
‘Make it in America’ chip push
The United States once led the world in chip design and manufacturing. As production shifted to Asia, America’s share of global chip fabrication capacity fell from 37% in 1990 to just 12% by 2020.
Former Commerce Secretary Gina Raimondo summed up the American predicament by saying that America had “dropped the ball,” allowing Asian rivals to surge ahead. Now, after pandemic-driven supply chain shocks, Washington is determined to “onshore” and “reshore” semiconductor production. But is it working?
A flurry of policies followed, from tariffs and trade pressure to hefty investment incentives. President Donald Trump threatened tariffs to push TSMC and others into building US plants. As a result, TSMC agreed to a $12 billion fab in Arizona, later expanding to a $40 billion project, which marked their first advanced facilities outside Taiwan. These fabs, when fully operational, will produce 4nm and 3nm chips, still trailing Taiwan’s 2nm technology but among the world’s most advanced.
President Joe Biden followed with the CHIPS and Science Act, a $52 billion package of subsidies, grants, and tax credits designed to “supercharge” US semiconductor manufacturing. This resulted in big investments from both American and foreign firms.
TSMC secured $6.6 billion in US grants for its Arizona plants, Samsung got $6 billion for a new Texas plant, and Micron announced a $100 billion New York megafab. The irony? Onshoring incentives are also benefiting foreign giants, whose rise was built on decades of government support in Asia.
Building a robust domestic chip industry is proving complex. Chip fabs are among the world’s most sophisticated factories, requiring immense precision and years to build. TSMC and Samsung have faced delays, cost overruns, and skilled labour shortages in the US.
Arizona’s TSMC site even had to “import” technicians from Taiwan, causing friction with US labour unions. Making advanced chips takes armies of PhD-level engineers, yet US immigration policies restrict high-skilled talent.
Analyst Marc Einstein notes, “You can’t just magic PhDs out of nowhere.”
Many experts argue that expanding high-skilled visa programmes is essential for the US chip renaissance.
Chip manufacturing is a global ecosystem. An advanced chip may be designed in California and fabricated in Taiwan using equipment from the Netherlands and materials from Japan and Germany.
“No single country can do everything,” says TSMC Arizona president Rosemary Castanares.
Even US fabs rely on $150 million ASML lithography machines from Europe. For now, US-based fabs remain smaller and a technological step behind Asia’s mega-fabs.
Historian Chris Miller calls TSMC’s Arizona plants “a generation behind the cutting edge in Taiwan,” and much lower in output.
TSMC itself is clear: its most advanced chips, and bleeding-edge R&D, will stay in Taiwan. Arizona’s fabs get slightly older, though still advanced tech.
The US officials might tout reshoring wins, but the centre of gravity remains in Asia. Restoring US chip leadership is a long-term effort, needing not just money and factories but also investment in education, workforce training, and immigration reform.
US-China crossfire in semiconductors
As Washington tried to onshore chipmaking, it also wielded trade weapons to slow China’s technological rise. The US-China trade war, which started with tariffs in 2018, has increasingly focused on semiconductors as a strategic chokepoint. The Trump and Biden administrations have sought to deny China advanced chips and manufacturing equipment to “protect national security.”
A pivotal moment occurred in October 2022 when Washington enacted stringent export controls. These regulations prevent global companies from selling high-performance chips or chip equipment that utilises US technology to China without obtaining a difficult-to-secure license.
If a chip was made with US software or machinery, as almost all advanced chips are, exporting it to China is restricted. The rules even bar US citizens from working for certain Chinese firms, choking off a “key pipeline of American talent.”
American officials argue this is essential to prevent “sensitive technologies” from fuelling China’s military modernisation, since advanced chips are dual-use, meaning they power both civilian and military AI.
Beijing calls this “technology terrorism” and has filed complaints at the World Trade Organisation, accusing Washington of abusing export controls. Chinese officials warn that these moves destabilise global supply chains. The impact is very much real. Huawei’s handset business collapsed after US sanctions cut it off from advanced chips.
Other Chinese firms, like memory giant YMTC, have been blacklisted. Even the United Kingdom-based ARM won’t license its latest designs to Chinese customers. Washington’s allies in the Netherlands and Japan have joined in, restricting exports of crucial lithography and chip equipment to China.
China’s initial response was cautious, but it has since weaponised its own dominance in key minerals. In 2023, Beijing restricted exports of gallium and germanium, both vital for chipmaking, and later banned exports of more minerals to the US.
These tit-for-tat moves signal China’s willingness to hit back with strategic materials. China also imposed its own limited bans, such as restricting US firm Micron’s chips from critical Chinese infrastructure.
At home, China has doubled down on self-reliance, pouring tens of billions into its chip sector through national funds and the “Made in China 2025” campaign.
President Xi Jinping calls on China to excel in key core technologies to ensure domestic innovation, thus preventing the country from being hindered by foreign sanctions.
The trade war has forced Chinese firms to seek new markets and supply chain arrangements, but often with slimmer profits.
Meanwhile, allied equipment makers like ASML now face the loss of lucrative Chinese customers, which raises concerns about lost innovation and revenue.
Nvidia CEO Jensen Huang recently blasted US export controls as “backfiring.”
He notes that Nvidia’s share of China’s AI chip market fell from 95% to 50%, with Chinese firms ramping up in-house alternatives. The bans, he says, “have pushed Chinese companies toward home-grown alternatives, spurring Chinese investment.”
Bill Gates similarly says US pressure has forced China to “go full speed ahead” on its own chips.
While some US officials argue these bans buy the West a crucial lead time in military AI, critics warn the strategy may accelerate China’s self-sufficiency and ultimately weaken the American industry.
The Chinese playbook
China has responded to the chip war with a multipronged strategy. At the core: building a self-sufficient semiconductor ecosystem. State-backed funds and national strategies aim to reduce dependence on foreign tech, especially in critical areas like manufacturing equipment and chip design.
Chinese firms have aggressively recruited global talent, including Taiwanese and American engineers, and have sometimes resorted to industrial espionage. The urgency to innovate has only intensified after US sanctions nearly crippled companies like ZTE and Huawei.
One breakthrough occurred in 2023 when Chinese chipmaker SMIC produced a 7nm chip, used in Huawei’s Mate 60 Pro, despite lacking access to the world’s most advanced lithography equipment.
US experts suspect SMIC adapted older machines with multiple patterning to achieve the feat. The phone’s teardown revealed memory chips from South Korea’s SK Hynix, showing that China can still source key components through unofficial channels or stockpiles.
While 7nm lags behind Apple’s 3nm chips, the achievement signals that China can adapt around sanctions, even though it comes at a high cost. Chinese companies are also developing workarounds, like using open-source chip architectures (RISC-V) and clustering less advanced chips to achieve AI tasks. Diplomatic efforts target partnerships with countries like Russia and some in Southeast Asia.
In parallel, China is building its own software and tooling ecosystem to reduce reliance on US and allied IP.
Still, China faces several pitfalls, including corruption in its state funds, persistent dependence on imported materials, and the technological gap in ultra-advanced manufacturing.
Globally, the chip war is creating a bifurcated tech order. There is a US-led bloc with strict controls and the most advanced chips, while a China-centric sphere relies on indigenous innovation and sometimes older technology.
Countries such as those in Europe, India, and Japan are now pursuing domestic manufacturing as a strategic objective. The aim is to avoid dependency on a single foreign supplier.
The global economy
The US-China semiconductor standoff has global ramifications. Allies, from Europe to Japan and India, are launching their own chip initiatives to bolster supply chain resilience. The European Union’s Chips Act aims to double Europe’s production share by 2030.
Japan is subsidising TSMC’s new Kumamoto plant; India, pitching its low-cost labour and market scale, is working to attract chipmakers despite obstacles like land acquisition and water supply. All these moves indicate a shift since countries want to reduce overreliance on any one supplier.
The balancing act is delicate for Asia’s chip powerhouses such as Taiwan, South Korea, and Japan. While they are US security partners, they rely heavily on China for a significant portion of their chip exports. With extensive operations in China, Korean giants Samsung and SK Hynix have even required waivers from US regulations.
TSMC, while benefiting from the US “friendshoring,” is careful not to sever links with Chinese customers. Diversification, by building plants in the United States and Japan, hedges bets against both geopolitics and American pressure.
A major unintended consequence is tech ecosystem fragmentation. If the world splits into separate tech stacks, innovation could slow (due to duplication and lost scale), but could also spark alternative breakthroughs. If denied access to leading-edge chips, the Chinese firms might focus on alternative architectures or software innovations.
The US and its allies, wary of supply chain risk, are building redundancy at a higher cost. A Boston Consulting Group study estimates that a full US-China semiconductor split could cost US companies $80 billion in lost revenues and $20 billion less R&D annually. Despite these costs, there is hope that competition will spur next-generation innovation, including quantum chips, new materials, and more resilient supply chains.
Governments everywhere are pouring money into chip R&D, education, and mature-node production for critical industries like autos and defence. Recent chip shortages made clear that even older chips are vital.
AI is front and centre in this fight. US restrictions aim to hold back China’s AI progress by limiting access to the most powerful GPUs. In the short term, it’s working, as Chinese companies are scrambling to adapt.
But software-side innovation and hardware workarounds are likely. If anything, scarcity may force efficiency and new approaches to AI development. In the long run, stifling hardware access may backfire by spurring domestic breakthroughs in China and elsewhere.
The chip war also raises fundamental questions about economic sovereignty. Governments now ask whether they can count on secure chip supplies in a crisis.
For many, the answer is “not yet,” driving a rush to build national capacity and regional redundancy, even if it means higher costs. Taiwan’s “silicon shield” still matters, but if TSMC globalises, no one country will wield absolute leverage for long.
Looking ahead, an “armed détente” is possible: both superpowers invest in reducing their vulnerabilities, and a new equilibrium emerges. North America might reach 20% of global chip output by 2030; China could attain partial self-sufficiency in 7nm or 5nm nodes. The world could then operate dual tech systems, trade some chips while restricting others for security reasons.
A total rupture, such as war over Taiwan, remains a nightmare. Disruption to Taiwan’s fabs would cripple the global electronics industry.
So far, fear of mutual destruction has preserved the status quo. As the US and others reduce reliance on Taiwan, that deterrence may weaken over time.
The world is realising, sometimes painfully, how critical and fragile the semiconductor supply chain is. Chips have become the new oil, with control over them reshaping the global balance of power in the 21st century.
Silicon geopolitics is here to stay, and every country will feel its impact.