What Do We Learn about Capitalism from Chip War? – Monthly Review

Integrated circuit on a microchip. Public Domain, Link.
It would not be an exaggeration to say that the chip has become an essential commodity for economic development, just as steel was a century ago. The subtitle of Chris Miller’s Chip War itself refers to “The World’s Most Critical Technology.” Chips are central to the technology making our lives better—personal computers, mobile phones, the Internet. In 2021, the computer chip industry produced a greater number of transistors than all the goods put together in human history.¹
Semiconductor chips truly are a marvel of human endeavor, and show the extent to which we can go in harnessing the physical world. Today, an iPhone12 is powered by an A14 processor, which contains around twelve billion transistors carved into its silicon; a self-driving vehicle is like a smartphone in the sense that specialized chips are key to its functioning. The world’s biggest auto companies can use chips worth more than $1,000 each in a single car.
The modern chip industry seems to represent the triumph of private large corporations over state bodies. We hear constantly about corporate players like Apple, Intel, Samsung, IBM, Facebook, Amazon, and many more. Chips truly represent globalized production. As Miller states in his preface:
A typical chip might be designed with blueprints from the Japanese-owned, UK-based company called Arm, by a team of engineers from California and Israel, using design software from the United States. When a design is complete, it’s sent to a facility in Taiwan, which buys ultra-pure silicon wafers and specialized gases from Japan. The design is carved into silicon using some of the world’s most precise machinery, which can etch, deposit, and measure layers of materials a few atoms thick. These tools are produced primarily by five companies, one Dutch, one Japanese, and three Californian.… Then the chip is packaged and tested, often in Southeast Asia, before being sent to China for assembly into a phone or computer.²
Chip War is informative, especially for those who are not familiar with the specificities of the semiconductor industry, and provides a broad understanding of an industry that is pivotal to the present order. It gives readers a sweeping historical account of the shifting technology, business models, and landscape of the critical semiconductor industry and its ever-changing dominant forces and actors. It presents the history of the industry since its beginnings in the late 1940s, when transistors were invented at Bell Labs by William Shockley and his colleagues, to the present day. In these ways, the book rightly deserves the praise and excellent reviews it has drawn from almost all corners, including leading think-tank authorities, strategists, economists, and the military. But the most instructive aspect of the book, even more than the rich details about the semiconductor industry, is what one learns about capitalism.
The first thing that we learn from Chip War is that we are well and truly in the age of monopoly capital. Perhaps this can be best illustrated through the example of Advanced Semiconductor Materials Lithography (ASML), a key producer of the photolithography machines used to etch transistor structures onto silicon wafers. Dutch ASML was started in 1984, when Philips spun out its lithography division. Once Intel’s Silicon Valley Group was bought by ASML in 2001, it established complete control over the supply of photolithography machines worldwide, without which no advanced chips can be made. Once such monopolies are established, it is almost impossible to dislodge them given the massive scales of capital expenditure required, perhaps like no other industry in the world, and more so as the global economy considerably slows down. At present, the global chip industry spends over $100 billion annually on capital expenditure.
To elaborate on the functioning of monopoly capital in the semiconductor industry, let us take the example of development of ASML’s extreme ultraviolet light (EUV) lithography technology. Miniaturizing chips requires harnessing EUV, which uses a wavelength of 13.5 nanometers (one one-thousandth of the thickness of a sheet of paper). Developing an EUV tool was one of the biggest technological gambles of our times; every step required breakthrough innovations. An EUV machine costs more than $100 million, with each component designed to last at least 30,000 hours—that is, almost four years of functioning. ASML personnel are posted on site for the entire life of EUV machines for maintenance. These machines are, supposedly, the “most expensive mass-produced machine tool in history.”³
Years before ASML had produced a functional EUV tool, its biggest customers—Intel, Samsung, and the Taiwan Semiconductor Manufacturing Company (TSMC)—each heavily invested directly in ASML to ensure that the company had the requisite funding to continue developing EUV tools. Intel alone invested $4 billion, over and above billions of previous grants and investments over the years. Mastering the challenges to produce only the laser of the EUV tool took a decade, as it needed the finest of mirrors, the likes of which had never been made before. Each laser required 457,329 parts, with most of them being sourced from different companies globally. As ASML produced only 15 percent of components for its machines in-house, it bought several of its suppliers and invested in others in order to establish a reliable global supply chain for the regular production of EUV machines. This machine with hundreds of thousands of components, which took billions of dollars and decades to develop, came into being in the mid-2010s.
Monopoly capital does not entail monopolization of the entire value chain by any single actor. Semiconductor production today requires an elaborate division of labor at a global scale, without which not a single chip can be made. Chips from Taiwan provide 37 percent of the world’s computing power, just two South Korean companies produce 44 percent of the world’s memory chips, and Dutch ASML builds all EUV lithography machines. By the 2000s, the semiconductor industry had split into three categories:
Perhaps the best example of this sort of global division of labor is what Miller calls “the fabless revolution.” “Fabless” refers to the outsourcing of chip production by the companies that design them. This describes the reality that today, the most important U.S. chip companies do not make even a single chip by themselves, and have outsourced production mostly to the corporations of the Global South. One instance is the case of the semiconductor company Nvidia, the share price of which recently has become red hot. Nvidia has become the most valued semiconductor company in the world in terms of market capitalization. Founded in 1993, it specializes in 3D graphic interfaces, parallel computing, and many other such high-tech applications. In 2007, it released Compute Unified Device Architecture software, which cost $10 billion to develop. Following the logic of monopoly capital, the platform was released for free, but worked only with Nvidia chips, which are manufactured by TSMC. Another leading U.S. semiconductor company, Qualcomm, designs chips for moving voice calls across frequencies. Qualcomm is built on millions of lines of code, but the chips are fabricated by TSMC and Samsung. The UK-based company Arm was a start-up in 1990, funded by Apple, but now is owned by Softbank Japan. It sells its chip architecture to fabless design firms, which in turn outsource the manufacturing to foundries like TSMC. Arm found its niche in energy efficient portable devices like mobile phones, which today make up the market for one-third of all chip sales.
While Apple designs the main processor for an iPhone in-house, there are a dozen of different chips involved in controlling various processes on the phone: connecting with the cellular network; sensing images and motion; battery management, and so on. According to Miller, no company other than TSMC has the skills and capacity to produce all of these chips for Apple. “Designed by Apple in California. Assembled in China,” etched on a phone, he writes, “is highly misleading. The iPhone’s most irreplaceable components are indeed designed in California and assembled in China. But they can only be made in Taiwan.”⁴
Interestingly, what comes out of Miller’s account is that despite the scale, automation, and capital intensity in the high-tech and innovation-intensive semiconductor industry, it is as much about surplus value extraction from the labor in production as anything else. Cost-cutting in the production process is critical for the competitive advantage of monopoly capital. Perhaps this is best epitomized by the ruthless culture at Intel under Andrew Grove. In order to supply microprocessors for personal computers, Taylorism was systematically employed to improve the efficiency of twenty-first century semiconductor production, very much like the automobile industry a century ago. Pioneers of the semiconductor industry in Silicon Valley like Fairchild Semiconductor employed women and nonunionized workers and shifted production to relatively poor, less unionized places in the United States in search of cheap wages. In the 1960s, the company relocated to places in Southeast Asia, such as Hong Kong; Texas Instruments, Motorola, and others quickly followed. Fairchild then moved to Singapore, where trade unions were practically outlawed, then Malaysia, and so on. In fact, Miller goes on to say that semiconductor/electronics jobs became a major force against radicalization in Hong Kong, Singapore, Malaysia, Taiwan, South Korea, and Philippines in the 1970s, during the heyday of both the Cold War and a very hot war in Southeast Asia. However, this account stereotypes Asian workers, and Miller uncritically quotes a group of U.S. workers, who after a tour of Japanese semiconductor factories in late 1970s, commented that a “foreman [in Japan] put a priority to the company over his family.”⁵ Micron, started in Idaho in the 1980s, scored victories over both Silicon Valley and Japanese rivals by making employees work in “sweatshop” conditions.
One may tend to believe that, with all the automation and technology, human skills are marginal to the industry, but Miller reveals otherwise. Not only are human skills central, but these skills have been cultivated mostly as collective practices, either in public institutions or through publicly funded programs. Early pioneers of Silicon Valley, mostly young men, were educated in elite U.S. universities such as Harvard, MIT, Stanford, and Berkeley. Perhaps this is best epitomized in the magical and long career of Morris Chang. Born in China in the 1930s, Chang grew up in Hong Kong; was educated at Harvard, MIT, and Stanford; got involved in the early building of Texas Instruments, worked with U.S. military, and, finally, built TSMC from scratch. Institutions like the Semiconductor Research Corporation and the Defense Advanced Research Projects Agency funded programs at Carnegie Mellon and University of California-Berkley that generated new industry start-ups and new software-based chip design tools in the 1980s. These tools are now used by the industry worldwide and are critical to the U.S. strategic hold over the global semiconductor supply chain.
However, monopoly capital does not mean a lack of competition. In fact, one finds cutthroat rivalry among the semiconductor players in Miller’s account. Perhaps this is best exemplified by the decline of Intel. Intel has consistently focused on short-term profits and cost cutting since Grove’s early leadership. Despite being involved in design and fabrication, they failed to have an edge in either, and could not move into emerging applications, like mobile phones, artificial intelligence, and data centers, which led to the rise of rivals like Nvidia. By 2020, half of all EUV lithography tools funded and developed by Intel were installed by TSMC, their archrival in fabrication, while Intel had barely begun using these tools. The decline of Intel happened in spite of a research and development budget of more than $10 billion annually throughout the 2010s. Intel could maintain its position only in the personal computing market, as most PC architectures are defined by Intel’s x86 chip—despite more efficient ones being available since the 1990s. Its hold is primarily due to enormous switching costs involved for the industry in shifting from one standard to another.
Intellectual property figures prominently in Chip War, and its role can be best summarized by what K. T. Li, minister of the economy of Taiwan in 1968, told visitors from the U.S. semiconductor industry, including Chang, at the time of a visit with Texas Instruments. Intellectual property, Li said, was something that “imperialists used to bully less-advanced countries.”⁶ The charge of so-called intellectual property theft is a weapon that is repeatedly deployed by Silicon Valley to stave off any possibility of competitors from any part of the world. For example, in the 1980s, when Japan was regarded as “the Saudi Arabia of semiconductors” by the U.S. chip industry, Japanese firms were regularly accused of intellectual property theft, protected markets, and government subsidies—including access to cheap capital.⁷
Perhaps the most striking feature of Chip War is the constant presence of the state and rivalry among nations, notwithstanding all talk of the state receding with the rise of global corporations. Miller begins his argument with the shifting geography of the semiconductor industry, starting with Silicon Valley, the failed attempts of the USSR to catch up, then the rise of Japan, followed by South Korea and Taiwan, and finally, includes a large part on China-U.S. rivalry.
A 1959 CIA report assessed that the USSR was only two to four years behind the United States in semiconductor technology. In 1962, Zelenograd, Russia, received funding from Nikita Khrushchev dedicated to the development of semiconductors, much like Silicon Valley. Miller also reminds us of Soviet scientist Alferov Zhores, who shared the Nobel Prize in Physics for his work on integrated circuits in 2000 with Jack Kilby for fundamental work done in the 1960s. In Miller’s assessment, the USSR could never catch up due to a top-down, militarized strategy of mechanically copying Silicon Valley’s model and products.
Although the United States was loath to share its knowhow with the Soviets, they were very willing to license technology to Japan to entice it away from the communist camp. The postwar Japanese state also played a key role. For instance, government-owned telecom monopoly Nippon Telegraph and Telephone bought only from Japanese firms. By 1990, Japan was making half of the chip-manufacturing investment in the world. It also acquired a leadership position in lithography machines—Nikon, for example, had a 70 percent global share in 1980.
As Japan became a serious threat to U.S. dominance in the 1980s, Silicon Valley began seriously lobbying against its technology industry, seeking Washington’s support. Silicon Valley was closely tied up with the military: no less than 17 percent of U.S. military spending was solely for electronics hardware. All of this lobbying resulted in significant cuts in the capital gains tax, pension funds being allowed to invest in venture capital funds, and the tightening of intellectual property rights, resulting in a rush of funds into Silicon Valley start-ups. In 1986, the United States placed a quota on Japanese chips. Japan’s share of DRAM chips declined from 90 percent to 20 percent within a decade. As Japanese semiconductor competitiveness declined in the 1990s, primarily due to global geopolitics, Miller’s assessment is similar to his previous one of the USSR. According to him, “Japan’s seeming dominance had been built on an unsustainable foundation of government backed overinvestment.”⁸
One of the central moves in Silicon Valley’s strategy to outmaneuver Japan was to cultivate South Korea. Samsung, emerging as a key semiconductor player, made a huge investment in manufacturing and began selling chips that it produced under the Intel brand. By 1998, South Korea had become the largest producer of DRAM chips in the world.
TSMC started in the mid-1980s as a Taiwanese state project, with the government providing 48 percent of the start-up capital and bringing in Chang to lead it. Chang was given a free hand, with the only condition being to leverage his prestige and network in Silicon Valley in order to find a chip firm that would be willing to provide advanced production technology. Dutch multinational Philips put up $58 million and transferred its production technology and intellectual property license in exchange for a 27.5 percent stake in TSMC; the rest of the capital was provided by wealthy Taiwanese individuals who were “asked” by the government. Chang brought in most of the mid-level hires, who had work experience with Silicon Valley companies such as Motorola, Intel, and Texas Instruments; most of them trained at top U.S. universities. According to Miller, to his potential customers, “Chang promised never to design chips, only to build them” for others.⁹
In all the din about outsourcing manufacturing from the United States, what is often missed is U.S. control over key choke points in the semiconductor supply chain without which not a single chip, at least of the advanced variety, can be made. Moreover, the United States maintains a stranglehold over chip design, intellectual property rights, and the capacity to arm-twist other nation-states due to its military and economic power. Within Silicon Valley there is a set of corporate players with unique features, all of which have a monopoly hold over critical aspects of the semiconductor supply chain:
The U.S. share in world chip production has been steadily declining, from 37 percent in 1990, to 19 percent in 2000, and 13 percent in 2010. Even so, taking the value capture across the global semiconductor chain, aggregating chip design, intellectual property, tools, fabs (fabrication facilities), and so on, the United States still captures 39 percent of the value, followed by South Korea (16 percent), Japan (14 percent), Taiwan (12 percent), and China (6 percent). One notable feature of the industry is that China is the predominant sink of integrated circuit chips, importing $260 billion worth in 2017. Integrated circuit chips made up 36 percent of Taiwan’s exports in 2017, followed by 21 percent from the Philippines, 19 percent from Malaysia, 17 percent from Singapore, and 15 percent from South Korea in the same year.
Each of the key nations is trying to increase its control over the supply chain and get a degree of autonomy by having fabrication facilities within its borders. Each one is also willing to offer liberal subsidies to monopoly capital:
Another remarkable aspect of Chip War is the way the idea of actual war is normalized, so much so that one may almost miss that it has anything to do with blood and gore and lives of people. In fact, technology-driven modern warfare involves significantly more lives (and deaths). The book begins with war-like confrontation in the Taiwan Strait in 2020, then constructs a scenario wherein TSMC is hit by a Chinese missile. Miller lays out how the interests of the military establishment and Silicon Valley have been intertwined from the beginning. In 1965, the U.S. Department of Defense bought 72 percent of all integrated circuits, though, within three years, as many chips were being sold to the personal computing industry as to the Pentagon. Miller claims that most of the bombs dropped during the Vietnam War missed their targets, but claims one great positive outcome for the industry—that it was a “successful testing ground for weapons that married microelectronics and explosives in ways that would revolutionize warfare and transform American military power.”¹⁰ According to Miller, in the 1991 Iraq War, semiconductors were the “war hero” guiding the U.S. missiles that flattened Iraq, and, by proxy, Moscow’s technology sector, hitting targets with precision. However, if we look at contemporary accounts at the time of the war, there were serious questions raised regarding the accuracy of U.S. missiles in Iraq, with a large number of civilian losses as a consequence.¹¹
The focus of Chip War and, perhaps, the source of its title, is an examination of the rivalry between China and the United States, and the serious threat that the Chinese semiconductor industry apparently poses to the preeminence of Silicon Valley. Almost a third of the book is devoted to this subject exclusively.
In 1979, China had hardly any commercially viable semiconductor production, and only 1,500 computers in the entire country. However, the country made rapid strides and, by the late 1980s, Huawei started producing DRAMs that were similar to those of the early ’70s. Given that China’s attempts to find its way came with their own twists and turns, it would be difficult to make any linearized account of its struggle to find a toehold. Miller provides the example of Grace Semiconductor, Shanghai, started in 2000, which was cofounded by Jiang Mianheng, son of President Jiang Zemin, and Winston Wang, son of a Taiwanese billionaire. Grace Semiconductor also hired Neil Bush, brother of President George W. Bush, to advise on “business strategies” for $400,000 annually. The company failed miserably. At the same time, Richard Chang, formerly of Texas Instruments, set up Semiconductor Manufacturing International Corporation, also based in Shanghai, with a total of $1.5 billion in investments from Goldman Sachs, Motorola, and Toshiba. It hired hundreds of foreigners, including substantial numbers from Taiwan. Soon, they were receiving job offers from foreign chipmakers, and it was listed on the New York Stock Exchange in 2004.
China has been aggressive in its effort to acquire and develop semiconductor technology. Given that China remains the biggest buyer of semiconductors, chip companies cannot ignore it, and hence are willing to transfer some parts of their technology to the Chinese semiconductor companies and invest in subsidiaries in China. The Chinese state, in turn, has been more than willing to raise capital for these companies. Several global semiconductor corporations, such as Qualcomm, Intel, Advanced Micro Devices, Arm, and IBM, have been part of this process. Another notable attempt has been the development of Tsinghua Unigroup, which was spun out of Tsinghua University in Beijing in the 1990s. In 2013, Tsinghua Unigroup spent many billions of dollars in buying China’s most successful fabless chip design companies. By the following year, the group had struck a deal with Intel to make smartphone processors. It also attempted to buy a stake in TSMC, as well as Micron and other U.S. chip companies. In 2017, Tsinghua Unigroup was planning new investments to the tune of around $22 billion, funded by Chinese state-controlled financial institutions.
Perhaps no other case better represents China’s aggressive semiconductor pursuit than the rise of Huawei, which has been at the center of all sorts of geopolitics in recent years. Miller quickly dismisses the accusation that the rise of Huawei is due to intellectual property theft and points to its $15 billion annual research and development budget, one of the largest of any technology company in the world. Though it had massive state support, Huawei was willing to learn quickly from global business practices, then outcompeted many of the largest Western corporations, such as Nortel and Alcatel-Lucent. By the end of the 2010s, Huawei’s HiSilicon unit was designing some of the world’s most complex chips for smartphones, and had become TSMC’s second-largest customer (after Apple). In Miller’s assessment, if this trend would have continued—without political intervention from the United States and other powerful states—“Huawei looked likely to play a bigger role in the construction of 5G networks than any other company” in the world, overtaking Ericsson and Nokia.¹² By 2030, China’s chip industry could rival Silicon Valley’s influence. Quoting Miller, “This wouldn’t simply disrupt tech firms and trade flows. It would also reset the balance of military power.”¹³
The United States and Silicon Valley have gone into overdrive to stop China from controlling more chunks of the global semiconductor chain, beginning with the Barack Obama administration. The great contradiction that they face is the fact that for every major chip firm, the Chinese semiconductor industry also constitutes a huge market, often a bigger customer than any other. Hence, Washington and the U.S. chip industry are caught between trying to limit the Chinese industry and maintaining trade relations. Miller cites several examples of U.S. attempts at curtailing China:
Miller stops tracing these developments somewhere in 2020, but since then, the U.S. government has, if anything, redoubled its efforts to block China:
In Miller’s assessment, there has been little retaliation by Chinese firms, and Beijing appears to have accepted the role of a “second-rate technology player.”¹⁶ However, since the end of Miller’s story, there have been concerted attempts by multiple Chinese firms to respond to the U.S. ban. A few of them are listed below, the first of which is discussed in Chip War, but included here to provide a larger sense of the diversity of Chinese efforts to overcome the U.S. ban:
In the end, I will flag four further issues in Chip War:
Though we learn a great deal about the semiconductor industry from Chip War, Miller does not question the fact that there are limits to exponential growth in the real world. Hence the substantial questions raised by human suffering on a finite planet, faced with the immediate catastrophe of the planetary crisis, ought to be asked beyond Chip War, as we cannot expect their resolution from a system determined by the logic of monopoly capital and imperialism that Miller takes for granted.
Subscribe to the Monthly Review e-newsletter (max of 1-3 per month).
 2024 Monthly Review Foundation. All Rights Reserved
Monthly Review | Tel: 212-691-2555
134 W 29th St Rm 706, New York, NY 10001

source