An instructive read on microconductors and artificial intelligence is the book Chip War, the Fight for the world’s most critical technology, by Chris Miller (2022). In Portuguese, “The Chip War: the battle for the technology that moves the world”. Chips or microconductors have become a strategic technological resource, both economically and militarily. At the same time, the complex chip production chain reveals the interdependence between countries, suggesting tensions in production regions, such as Taiwan. The importance of chips becomes even more evident considering their incorporation into artificial intelligence systems. This article summarizes the historical background and considerations made by Miller (2022).
Keywords: technology; microconductors or chips; artificial intelligence; fabless business model, geopolitics.
The microconductor industry
In his book, Chris Miller (2022) presents a historical retrospective of the microconductor industry, from the post-World War II era to the present day. While he emphasizes the economic importance of this resource, he reveals why chips are a strategic military resource for many countries.
The book suggests how the factors of knowledge (engineering, physics, computing), institutions (government, regulation) and cultural environment help to explain the dominance of certain countries. Another notable aspect is the interdependence between the different players in this production chain, who design and/or produce microconductors and advanced photolithographic technologies. Miller (2022) warns of a possible conflict in the Taiwan region, one of the most militarily monitored.
The author narrates the development of microconductors and the main countries, companies, entrepreneurs and scientists involved. The actions of these players intertwine political, economic, scientific and military issues. It shows how important these technologies (microchips, photolithography and GPUs)1 are for countries, for data analysis, communication and artificial intelligence.
The demand for these microcomponents has been very noticeable during the pandemic. In mid-2021, several automotive industries stopped production because there were no chips available. The pandemic led to an increase in the sale of computers and other electronic devices, which consumed a lot of chips, and several production chains were paralyzed. Today, production of the most advanced chips is concentrated in Taiwan, particularly at the Taiwan Semiconductor Manufacturing Company(TSMC).
Historical context of chip production
In the 80 years since the Second World War, there have been interesting changes in the US, which has always tried to stay ahead of other countries in terms of technology. At the same time as collaborating with other countries, it used certain economic policy mechanisms to overtake them. However, not only economic mechanisms explain a country’s loss of influence (as in Japan in the 90s). There are internal and external factors related to culture, government, the economy and innovation that can help to understand the course of events.
Chips are considered by some to be the countries’ new oil (Miller, 2022, p. 97). There are logic processing chips (37% of them produced in Taiwan), memory chips (44% in two Korean companies) and analog chips (Japan). In addition, machines capable of producing these chips on a nanometer scale (down to 3 nanometers) are needed. So far, only one company makes ultraviolet lithography machines for chip manufacturing: the Dutch company ASML.
For the author, semiconductors can shape international politics, the structure of the global economy and the balance of military forces (Miller, 2022, p. xxvii). The Second World War was marked by the use of coal and steel in the production of tanks, airplanes, rockets and radar, as well as atomic bombs. However, it required more precise calculations to reach military targets. At that time, the first electric computers used vacuum tubes. A connected tube corresponded to “1” and a disconnected tube corresponded to “0”. But their use was slow and unreliable. In order to conduct electricity better, California-based theoretical physicist William Shockley suggested using semiconductor materials such as silicon and germanium, combined with phosphorus and antimony. Two of his colleagues put his theory to the test and managed to control the electric current in December 1947, creating the first transistors. These transistors replaced vacuum tubes, but wiring thousands of transistors was complex.
Shortly afterwards, Jack Kilby, from Texas Instruments, thought of bringing together multiple components in a single piece of semiconductor material, creating the first integrated circuit, better known as a chip. The first semiconductor company was Fairchild Semiconductor, formed by eight researchers from Shockley’s laboratory. Among them was Gordon Moore, who would later coin the concept of “Moore’s Law”, which explains the exponential growth of computing power.
Semiconductor production advanced when the company became a supplier to NASA’s Apollo program, coinciding with the search for a US response to the Soviet space program. The challenge of mass production was overcome with the arrival of other players, such as Morris Chang in 1958. By the mid-1960s, the US military was already using the first chips in all kinds of weapons, such as satellites, sonars, torpedoes and systems. The Pentagon was an important agency in recognizing the strategic importance of these technologies and funding research with its DARPA agency.
This use by the army ensured the military supremacy of the United States over the Russians at the end of the Cold War. Russia was unable to develop the production of microconductors because it lacked installed capacity and industrial manpower. The Russian government was aware of the need to keep up with this technological development, to the point of creating a city, Zelenograd, as a center for microconductors. The Russians’ strategy was to copy microconductors. But each step in the manufacture of a chip involved specialized knowledge, rarely shared outside the company, a kind of know-how. Therefore, the North American system set the pace and was difficult to copy.
In Japan, physical entrepreneurs like Morita, from Sony, have enabled the country to add their business model to the chip industry. They developed electronic products to meet the demands of end consumers. Their experience in design and marketing allowed the country to mass-market electronic products containing chips, such as radios, calculators and walkmans. It wasn’t long before US companies set up factories in Hong Kong, Taiwan, Malaysia and South Korea.
Among the countries that have established associations with the United States, the most significant case is that of the island of Taiwan. Its Finance Minister, K.T. Li, saw the production of microconductors for the United States as a way of gaining security and political stability in the face of tensions with China. At the time, many Asians trained in engineering and physics were hoping for job opportunities in the region, since Mao’s China limited technological development. Knowing this, Li, also a physicist, invited Morris Chang, from Texas Instrument, to lead the process in Taiwan. This was how the TSMC company was born, which would become the most important chip manufacturer today.
In the United States, two engineers from FairChild founded Intel, a company that would become one of the main suppliers of microprocessors for computers and data centers. In other words, with the end of the Cold War, chip manufacturers turned to the consumer market as their supply focus. In the 1980s, Japan overtook the United States in electronics. Faced with the massive influx of Japanese products, Silicon Valley companies turned to the government, considering taxes and domestic production costs. The US government took action on the matter and applied pressure until the Japanese government limited its exports to the US. In retrospect, in the 1960s Silicon Valley was very dependent on the Pentagon. By focusing on consumers, it became less dependent on the government and gradually grew to the point where the country overtook Japan again.
The fabless business model
Miller (2022) describes a shift in microconductor production towards the fabless business model: chip design has been separated from the manufacturing process. North American companies are in charge of design, while production is the responsibility of industries abroad. In particular, Silicon Valley concentrates the design of new microcomponents, while several Asian countries manufacture them. China was the last to enter this race, only under the government of Xi Ji Ping, but with huge investments and seeking to overcome US dominance.
With the services offered by Taiwan, especially TSMC, the fabless business model comes into operation (geographically separated design and manufacturing). Producing the most complex types of chips is currently a challenge for most countries. In addition, a factory producing advanced logic chips costs around 20 billion dollars.
In the case of Intel, in the 1990s it achieved impressive results by supplying personal computers and building data centers. However, the company missed the opportunity to supply a small company called Apple at the time, turning down the invitation to supply microprocessors for smartphones. Today, irreplaceable components of iPhones are designed in California, assembled in China and manufactured in Taiwan (Miller, 2022, p. 224).
Importance of knowledge and cultural environment
In the collaboration between the countries, scientific knowledge was crucial to the development of the sector and technologies. For example, Taiwan’s Minister of the Economy is a nuclear physicist with a degree from Cambridge; Sony’s Morita had a degree in physics, like many others. Without this scientific background and the closeness between companies, it was almost impossible to have access to this tacit knowledge. In addition, a country’s business environment and its culture of entrepreneurship and innovation are important factors.
However, the influence of governments on the chip production chain cannot be denied. In 2020, for example, the United States prevented the sale of chips to China, in an action called “chip choke” (Miller, 2022, p. 327). In recent decades, it has been graphics processing units, or GPUs, that have gained the most value. Among other uses, they enable training in artificial intelligence and their main developer is the American company Nvidia. To produce them, however, photolithographic reading machines are needed, produced only by the Dutch company ASML. TSMC is the main manufacturer of Nvidia cards. In addition, the US government recently prevented video cards produced by Nvidia from reaching China. Finally, it has supported the Dutch company, suggesting how strategic these technological resources are for the country.
To conclude
The “war” for chips shows that these resources are not only relevant to economic results, but also to military security. New issues, such as cybersecurity and cyber-attacks, have called for increased governance of conflicts based on information and the digital. Miller’s book (2022) shows the concentration of design in the North American context and of manufacturing processes in Asian countries. The Chinese government, for its part, prefers not to depend on other countries. But in general they agree that advanced technologies are important for information management, security and communication.
- Photolithography allows laser access to the most complex chips, which over the years have drastically reduced their size and exponentially increased their processing capacity. GPUs are graphics processing units, capable of handling 3D graphics, and differ from CPUs, computer processing units.
References
Miller, C. (2022). Chip War: the fight for the world’s most critical technology. New York: Scribner.