Worries about the twilight of American hegemony and China’s rise are producing a new consensus in Washington. At a recent congressional hearing on “the Chinese Communist Party’s political warfare,” U.S. House Representative Jasmine Crockett (D-TX), referencing the small country caught between the world’s two superpowers, captured the mood: “I often tell our friends from Taiwan that the only thing that is bipartisan in this 118th Congress is China.”
A case in point is the CHIPS and Science Act of 2022, one of Joe Biden’s signature legislative accomplishments, which commits $52.7 billion to jumpstart U.S. semiconductor manufacturing through a mix of subsidies, tax credits, and R&D and workforce training expenditures. If not for the unifying threat of China’s technological advances, the bill probably would have fallen apart along the usual partisan lines. Instead, now two years into implementation, the law has been hailed as an economic sea change, a decisive shift from neoliberal free trade to state-directed industrial policy.
The United States doesn’t really make chips these days, instead relying on a complex process of design, production, assembly, and testing that spans the globe. The vast majority of fabrication is done in East Asia; Taiwan, in particular, produces 41 percent of all processor chips and more than 90 percent of the most powerful chips, essential to advanced computing and AI. The supply chain’s concentration in an island nation with which China expressly seeks to “reunify” gives the whole matter unusually weighty stakes. At a White House event to get the bill past the finish line in Congress, Deputy Secretary of Defense Kathleen Hicks put it in terms bordering on thermonuclear: “Semiconductors—it’s not an overstatement to say—are the ground zero of our tech competition with China.”
While companies like Intel and Samsung are huge beneficiaries of the CHIPS Act, everyone understands that the major coup is getting the world’s number one chipmaker, Taiwan Semiconductor Manufacturing Company (TSMC), to build its foundries on U.S. soil. Even as competitors have closed ground in recent years, no company can yet match TSMC’s manufacturing precision, high yields, and pace of production, and it’s reportedly well ahead of the pack in progress toward next-generation “2 nanometer” chips. TSMC—which has now broken ground to build foundries in Arizona—is the most valuable company in Taiwan’s history, its global success unfathomable for a country developed through cheap exports.
What explains its success? The most common answers cite the company’s exclusive focus on chip fabrication and the intense work culture set by hard-driving managers. Still, why Taiwan? What made it possible for this small country, whose sovereignty remains intensely disputed by Beijing, to dominate the ultra-valuable production of advanced semiconductors?
The answer requires a longer view. The dead weight of centuries of colonization has shaped and continues to haunt Taiwan’s semiconductor project. Even as the country’s ultra-successful chip industry has secured its place atop global supply chains, neoimperial entanglements remain unsettled. Today we are witnessing the consequences.
Chips are the heart of all computing. These small integrated circuits—thin squares measuring just a few millimeters on a side—contain several layers of electrical components and can hold billions of very tiny transistors, each of which can rapidly turn an electrical current on or off. This is the physical mechanism that allows computers to process digital information, encoded as strings of 1s and 0s; it relies on the semiconducting properties of elements like silicon, and more transistors crammed into a single chip means more powerful and efficient computation. Improvements in design and fabrication at these microscopic scales have made virtually all modern information technologies possible, from personal computers to smartphones, data centers, and more.
Before venture capitalists and corporate executives recognized the potential for profit, chip production was driven by U.S. Cold War strategy. Pentagon leaders understood they could not compete with the sheer numbers of the Soviet army; new technologies, powered by chips, offered a way to build up advantages in surveillance, intelligence, and weaponry.
Everyone understands that the CHIPS Act’s major coup is getting TSMC to build its foundries on U.S. soil.
The U.S. national security establishment quickly embraced the idea. The Air Force’s 1962 Minuteman II intercontinental ballistic missile, guided by a small Texas Instruments chip, was a proof-of-concept smash success, opening the doors to a host of Silicon Valley firms that designed, produced, and delivered chips to the Pentagon. Consumer demand for electronics and personal computers skyrocketed in the 1980s, but that didn’t mean semiconductors lost their national security valence. “The Goliath of totalitarianism will be brought down by the David of the microchip,” Ronald Reagan declared in the last year of his presidency. “More than armies, more than diplomacy, more than the best intentions of democratic nations, the communications revolution will be the greatest force for the advancement of human freedom the world has ever seen.” The war between capitalism and socialism depended as much on commercial forces as overt military strength.
Today, few companies manufacture chips themselves. Big players like Nvidia, AMD, and Apple design state-of-the-art chips—like the graphics processing units needed for the computation-intensive deep learning behind generative AI—but they rely on other firms to mass produce them.
There’s a reason for this state of affairs: it is incredibly challenging to make modern chips at scale. For one thing, startup costs are massive; a single fab facility—imagine a building spanning more than twenty football fields, filled to the brim with precious, highly specialized equipment—can cost upward of $15 billion. The process is also notoriously difficult and labor-intensive.
In simple terms, across the whole supply chain, it goes like this. Silicon, bought from companies that mine silica-rich sand and quartz in regions around the globe, is melted at an extremely high temperature and cooled into a column; a diamond saw then slices it into uniform horizontal wafers, which are sold to fabricators like TSMC. There, various materials, including a light-sensitive coating called photoresist, are applied to the wafer’s surface. A special type of light is shone on the coating to etch out microscopic patterns, and ions are implanted to control the flow of electricity, resulting in a series of interconnected electrical components. This process of pattern masking and etching is repeated again and again, each time adding another nanoscopic layer of computing power. After months of production time, the wafer is cut into individual chips, which are sent for testing and final assembly. The most advanced chips, boasting unfathomably small transistors, are pretty much the most complex good ever produced by global capitalism. Today, only TSMC and a few other advanced foundries can produce such circuits at reliable yield ratios and a rapid pace of delivery.
TSMC was founded in 1987 by Morris Chang, a Chinese-born, MIT-educated engineer who rose up the ranks at Texas Instruments before being recruited by the Taiwanese government to leave the United States and lead Taiwan’s semiconductor sector. Now a multibillionaire, Chang is often held up as another mythical “titan of industry” in the mold of Henry Ford or Steve Jobs. In reality, as is always the case, larger currents shaped the terrain on which he and TSMC found unimaginable success.
Many of these forces stretch back centuries. In 1683 the Manchu-led Qing Empire of China, then the dominant empire in East Asia, conquered Taiwan to expand its influence and eliminate political rivals at its frontiers. European traders, present as early as 1624, were largely marginal next to regional powers. For more than two centuries, generations of Chinese migrants, mainly from the provinces of Fujian and Guangdong, settled in Taiwan, pushing the Austronesian indigenous people into the mountainous interior. When China ceded the island to Japan following the First Sino-Japanese War in 1895, Taiwan became Japan’s first colony. As part of a larger campaign to colonize Asian territories to its south, Japan sought to run the island as a “model colony” to showcase the civilizing ways of Japanese imperialism to the locals (settler migrants from the Qing days) and to rival colonizers from the West.
Imperial Japan invested heavily in the island’s infrastructure so that it could supply the homeland, launching large-scale development projects that transformed the country from one of subsistence agriculture to industrial production. One such project erected a vast transportation network to connect Taiwan’s main ports of Kaohsiung, Keelung, and Hualien to facilitate the export of sugar. Extraordinary sums were spent on new harbor construction, railroad feeder lines, irrigation canals, and the promotion of the sugar industry. Within ten years of occupation, Taiwan was financially self-sufficient and shipping products and natural resources to Japan. The year-by-year increase in exports as a percentage of gross GDP exploded, particularly in the 1920s and 1930s—rates that wouldn’t be seen again until the country’s shift to low-cost contract manufacturing in the 1970s.
The colonial government financed these industrial projects by radically recasting Taiwan’s economic structure—creating a central bank, imposing land registration policies, and levying taxes on the basis of productivity. These changes generated tax revenue while protecting private property, which the government further bolstered by establishing tax policies that rewarded private savings and capital formation. (The Taiwanese government would employ similar tactics later, in the 1960s, to become a tax haven for exporters.)
As Japan set up these infrastructures of primitive accumulation, its day-to-day administration of the Taiwanese people was relatively peaceful and stable. Despite the colonial government’s suppression of movements for self-rule and political and economic rights, some older Taiwanese people today continue to identify along the lines of Japanese nationalism, appreciating the “civilizing” ways of Japan over barbarous China. The colonizers had implemented an assimilation policy, Kominka, to transform the Taiwanese into Japanese imperial subjects through cultural programs and primary language education, but the reasons, in the end, are mostly material: the Japanese were the ones who built roads and hospitals.
Japanese imperialism had a very different reception in China. In the Second Sino-Japanese War, starting in the 1930s and escalating into World War II, Japan’s inward march to colonize China resulted in brutal violence, subjugation, and the death of millions. Taiwan’s comparatively more accommodating and orderly treatment effectively separated its fate from the mainland’s, producing complicated sentiments of national identity. For many people in Taiwan, China and “Chineseness” were artifacts of bygone times under Qing rule.
When Japan invaded, the Chinese state was in the middle of civil war between the Republic of China, governed by Chiang Kai-Shek’s nationalist Kuomintang party, and Mao Zedong’s Chinese Communist Party (CCP). The two sides entered a tentative alliance to expel the Japanese during World War II, and after Japan surrendered, the United States forced the transfer of Japanese-occupied Taiwan to the Republican Chinese as fighting with the communists resumed. When the CCP won control of most of the mainland in 1949, the military and political elite of Chiang’s regime fled to Taiwan. There, with U.S. support, they instituted martial law that would last for almost forty years, crushing worker activism through repressive labor laws and curtailing dissent via campaigns of “white terror.”
Here is the peculiar case of Taiwan. The small island has been part of multiple imperial formations, making it possible to imagine and project many different ideas of “empire” onto it. For the settler colonist Kuomintang party, in exile from defeat in the Chinese Civil War, the island society was the new home of the Republic of China. To them, Taiwan was the rightful heir of the real China, even as many Taiwanese, after decades of Japanese colonization, no longer saw themselves as Chinese, communist or otherwise.
On the other side, the victorious People’s Republic of China (PRC), governed by the CCP, claimed Taiwan as its territory, invoking boundaries set a half century earlier under Qing rule. On this telling of history, Japanese colonialism was a brief interruption in a grand arc of Chinese nation-building. The United States would codify these dueling nationalist claims in an official communiqué in 1982, claiming that each side agrees “there is but one China and Taiwan is part of China.”
These contradictions of Taiwanese national identity were fully entangled in the geopolitics of the Cold War. Observing the PRC’s support for Korean communists in the Korean War and sensing the spread of communism across the region, the United States quickly adopted an interventionist policy of anticommunist influence in East Asia in the 1950s. It saw in the Kuomintang an opportunity to defeat, or at least isolate, communism and install a brand of global capitalism in its own image. Frantz Fanon rightly observed that the emerging cold war saw “the Americans take their role of patron of international capitalism very seriously.”
The United States thus used Taiwan as an outpost of American imperium: a staging area for military forces and surveillance powers in Asia and an ideological foil to its mainland neighbor. It supported the right-wing nationalist regime with economic aid, military support, and diplomatic backing, and Taiwan’s leaders were all too eager to embrace their benefactors. Between 1950 and 1967, U.S. foreign aid flowed into Taiwan to the tune of nearly $4 billion, about 40 percent of which went to non-military economic development. (All this comes to north of about $44 billion in today’s dollars.) In exchange for this influx of capital, the United States pressed the Taiwanese state to enter international markets through export-led economic policies. By 1973, Taiwan was the seventh-largest source of U.S. imports, and the country’s production had shifted from sugar exports to industrial products like textiles and electronics.
Japanese colonizers had laid the blueprint for industrialization, then, but it was Kuomintang authoritarian control, underwritten by U.S. foreign policy, that rapidly modernized Taiwan’s economic development. Multinational companies recognized the advantage of doing business in Taiwan, seeing an opportunity to exploit cheap, quiescent labor in an authoritarian regime. Between 1960 and 1973, U.S. venture capital deployed in Taiwan grew at an average annual rate of 28.2 percent. Sociologist André Gunder Frank coined the term lumpenbourgeoisie to describe Third World elites who spearheaded their own exploitation from above in order to accept the military, political, and economic support allocated to them to remain in power. He was describing the postcolonial situation in Latin America but could just as well have been writing of postwar Taiwan.
It was out of these forces that TSMC emerged. By the 1980s, Taiwan had some experience in state-backed semiconductor exports, buoyed by a successful 1976 technology transfer agreement with the American corporation RCA to build local manufacturing facilities and expertise.
To further advance the sector, the Kuomintang decided to invest in a homegrown manufacturing company, drawing from the government’s sovereign wealth fund to provide TSMC with 48 percent of its startup capital. (The rest came from the Dutch tech company Philips and wealthy Taiwanese who owned firms in plastics, textiles, and chemicals.) The company launched with a new business model of exclusively fabricating chips for other companies, a plan that played to Taiwan’s strengths in attracting Western capital. The Kuomintang would be there to nurture it every step of the way. As Chris Miller writes in Chip War: The Fight for the World’s Most Critical Technology (2022), “From day one, TSMC wasn’t really a private business: it was a project of the Taiwanese state.”
But exactly what was this state? Economic development had begun to drag following the liberalization of China’s economy in 1978, which created new competition for cheap labor. Many Taiwanese business owners would close local factories and move operations to the mainland to exploit even lower wages. In this environment, Taiwan saw large, capital-intensive development as the main way it could hold onto its rung of the ladder in multinational supply chains.
The nation’s internal politics were also changing. The generalissimo Chiang died in 1975, and the nationalist party, once a party of mainland-born military elites, became a home for technocratic entrepreneurs and professionals. Meanwhile, a prodemocracy movement took off in 1979, gaining momentum over the next decade as it pressed for reforms. The death of the ancien régime seemed only a matter of time, yet the Kuomintang and new president Chiang Ching-kuo devised a strategy that would accommodate the ruling class to the changing tides. As Perry Anderson has described:
Once [U.S. President] Carter had recognised the PRC in 1977 . . . Chiang Ching-kuo, seeing that he could be left high and dry by Washington, moved to relegitimise KMT rule by gradually liberalising its system from above, and then picking a local successor—calculating that this would make it very difficult for the U.S. to abandon the island. Democracy, when it came to Taiwan, was thus the combined result of an opposition pushing against dictatorship from below, and a regime in quest of new credentials from above.
Martial law was lifted in 1987, the same year TSMC was founded. The grassroots democracy movement—fueled, in the main, by small producers, upper-class professionals, and students—was not in a position to counter the exploitation of cheap labor. The reform movement had thoroughly internalized anticommunist ideology, and the triumph of industry was a point of national pride. Taiwan’s new democratic institutions, culminating with elections in 1996, thus consolidated an explicitly bourgeois politics tied to elite stewardship of the economy.
Obviously, the bet has paid off. According to U.S. International Trade Commission estimates, Taiwan’s chip exports—some $184 billion in 2022—account for about a third of the country’s merchandise exports, and its semiconductor industry makes up about 15 percent of GDP. With a market cap on the order of $800 billion, TSMC is by far the country’s most valuable company, driving Taiwan’s stock exchange to record highs.
But the nation’s majority have borne the social costs of modernity. Today Taiwanese society is sharply stratified, and class inequality has ballooned. The labor movement has won modest labor law reforms in recent years, but the climate for workers remains grim. The Taiwanese work the sixth-longest hours in the world, even as Taiwan produces more wealth than many advanced economies. About 6 percent of Taiwan’s workers are unionized, worse even than the United States’ dismal rate of 11 percent. Real wages have been largely stagnant since 2002, despite gains in labor productivity. TSMC, in particular, is noted for its “intense, military-style work environment,” and local commentators have jokingly likened the working conditions to slavery. Chang, its iconic founder, has credited Taiwan’s hostile attitude toward labor unions as key to the company’s rise.
At the same time, the cost of living is rapidly increasing. Taipei and other cities are littered with luxury apartments few can afford. Voter turnout and political participation are declining. And the semiconductor industry continues to take an acute ecological toll, from pollutants and chemical exposure to titanic consumption of water and energy.
The chips were supposed to ease geopolitical uncertainties, not produce them. (These days, many Taiwanese proudly call TSMC the country’s “sacred mountain.”) The reality is the opposite. Nativism is resurgent, and a “new cold war” is on. U.S. industrial policy is explicitly framed as a matter of national security. China, responding to a U.S. export ban on advanced chips, is accelerating its domestic semiconductor R&D; it could also invade Taiwan to take the chip foundries directly. The age of High Imperialism is dead, and neoliberalism is fading, but so far as these developments portend, imperial conflict looks to be dressing itself in new clothing.
Independent and nonprofit, Boston Review relies on reader funding. To support work like this, please donate here.
