Reading Chip War by Chris Miller made me realise how much the advances in chip technology had passed me by. These advances were made by a combination of brilliant engineering to develop the technology and the cutthroat competition to sell the products at a profit. The “war” is not only the commercial one between the chipmakers but also the geopolitical one between countries.
Small beginnings
The pioneers of chip technology in the 1960’s were Fairchild Semiconductors and Texas Instruments (TI). The challenge was to move from individual transistors wired together to integrated circuits etched onto germanium and silicon wafers.
Fairchild got an initial break with contracts to supply NASA (for Apollo 11) and the Pentagon (for Minuteman missiles) (p21). Both realised that the key to developing a market was in perfecting manufacturing techniques rather than the science of chip design. The progress in this fledgling industry was driven by a few ambitious, as well as brilliant, individuals.
America encouraged the development of Japanese industry after the war but, after a while, American companies felt disadvantaged by the subsidies and low interest rates available to Japanese chip makers. In contrast, American firms worked together. Defence expert William J. Perry got his chips from Intel boss Robert Noyce who sang in the same Palo Alto madrigals choir as Perry (p74).
The Japanese edge on quality of chips was confirmed publicly by Hewlett Packard (p90). But it was not only chips where Japanese quality won. Precision lenses were needed for photolithography to etch circuits onto chips. American companies like CGA, run by maverick Milt Greenberg, could not compete on quality with Japan’s Nikon. Also, the USA could not compete with the cheap labour markets in South East Asia whereas Japan could with its loyal but underpaid workforce.
Japan can say no
In the 1980s, Sony co-founder and boss Akio Morita wined and dined American CEOs and demonstrated that Japan could say no to the mighty USA’s protection.
US semiconductor firms began to go bust in the face of cheap quality product from Japan. Only brutal management decisions kept the likes of Intel alive. The US sought to partner with Korean firms like Samsung, who could compete with the Japanese. But Samsung had their own ideas. Korea was slowly overtaking Japan in chip production.
The Russians followed a policy of “steal and copy”, but this was no real threat to America as Russia would always be several years behind. By 1990, the Russians acknowledged they’d lost the chip war as well as the Cold War.
Sony made a critical innovation in developing specialised chips for image sensors still used in digital cameras today. But most Japanese firms were locked into memory chips and didn’t foresee that microprocessors would be needed to feed the rise of the personal computer (PC) (p156).
The rise of Taiwan
In 1985, Taiwan persuaded TI engineer Morris Chang to set up semiconductor production in Taiwan with the radical idea to make chips to customers’ designs. Soon, Taiwan Semiconductor Manufacturing Company (TSMC) was making inroads the industry (pp163-9).
Meanwhile, China’s engineers were sent to hoe fields. The Cultural Revolution deemed electronics anti-socialist. Only when Mao Zedong died and the new leaders decided semiconductors were essential to China’s rebuild, was Richard Chang, with his TI and TMSC experience, allowed to build fabrication facilities in China. But China had a long way to go to catch up with the rest of SE Asia which had invested heavily in such facilities.
Lithography and EUV
In order to cram more processing power into a chip, the lithography (etching) had to use light of shorter wavelengths. Extreme Ultraviolet (EUV) won out over other techniques like X-rays.
The US was worried that the leaders in lithography, Nikon and Cannon, were Japanese and preferred to use a Philips spin-off, Advanced Semiconductor Materials Lithography, in the Netherlands (pp183-9).
Chip architecture
By the 2000s, PCs were powered by Intel’s chips with x86 architecture. Although less efficient than rival RISC architecture, the x86 prevailed. Servers in the booming data centre sector used x86. The mobile market, however, was more suited to the RISC architecture, and a British company, Advanced RISC Machines Ltd, licenced its architecture to power chips for mobile phones and other devices.
Fabrication units were stupendously expensive at $20bn. so companies like Nvidia concentrated on Graphics Processor Units (GPUs) to create realistic 3-D graphics.
Chip design vs fabrication
Mobile phones needed to cram as many calls as possible into a wavelength. Qualcomm founder, Irwin M. Jacobs, hit on the idea of sneaking one call into gaps in other calls. And it worked! Qualcomm went on to design chips for many applications but left the fabrication to “foundries” like TSMC.
The 2008 financial crisis caused many firms to cut workforce and investment. When TSMC founder Morris Chang saw his successor doing this, he sacked him and started a program of investment. Abu Dhabi had a similar vision and set up GlobalFoundries with massive capital. EUV became a reality, but GlobalFoundries bailed out, leaving just Intel, TSMC, and Samsung in the EUV race.
Intel also abandoned EUV in favour of their GPU and AI business. And then there were two: TSMC and Samsung. [Ed: It’s interesting to note that in 2026, Intel are back in the EUV game!].
China awakes
China was worried it was always behind the curve, so President Xi announced an assault on core technology. China had cut some favourable deals with foreign countries but needed make a step change in its strategy.
Domestically, a company called Huawei had built a presence in many sectors, especially telecommunications. Its founder, Ren Zhengfei, followed the Samsung model of making products that could compete globally on quality and price. Huawei acquired US technology – sometimes legally – and invested massively in research and development. It certainly had government backing but whether it was run by Huawei is in doubt.
The current 5G technology uses not only every nanosecond of transmission (using the gaps as Qualcomm did), but also the concept of “beam forming” – transmitting to the target in a narrow angle rather than 360°. This processing power, developed for mobile phones, is now used in many other devices, especially automobiles like Tesla.
China was also seeking to dominate in the military sector with intelligence-driven weapons. One tool is that of predictive maintenance, first developed to enable reliable microprocessor manufacture, to fix a component before it goes wrong.
US tries to choke China
In 2018, the first Trump administration replaced the laissez-faire of Obama and banned some exports to China. ZTE (partially owned by the Chinese state) stole from the USA’s Micron, so America banned export to several Chinese firms.
Companies like Fujian Jinhua, starved of imports, went bust.
SMIC and Huawei imports from China were restricted. This paradoxically boosted China’s drive to excel and compete globally. China refused to shut down YMTC – based in Wuhan, the centre of the Covid outbreak – instead pressing their advantage. However, building its own EUV capability was a road too far for China.
Where to now?
Most processors use propriety architecture like x86 (from Intel) or RISC (like British company arm). RISC-V, an open architecture version of RISC, is also now gaining popularity.
Covid indirectly caused supply chain problems due mainly to the auto industry overreacting to Covid and prematurely cancelling orders. This illustrates the fragility of the complex supply chain needed to produce chips.
The biggest threat in this “war” is now that a military war may develop between China and the US over Taiwan. Neither side wants this, but the risk must be acknowledged. Taking out Taiwan’s TSMC would reduce global semiconductor capacity by 37% which would be globally detrimental.
Despite this very real threat, the market demand for ever more capacity is there as is the technology. So will computing power continue to increase exponentially according to Moore’s Law?
One thing is certain; the few years since this book was written have shown that the chip war still rages as hot as ever.
Postscript : Who would have thought that the Korean chip industry would be threatened by the US/Israeli war on Iran. Yet The Carnegie Endowment for Peace just published this (13/3/26): The Iran War Is Also Now a Semiconductor Problem.
Comments? We want to hear them. Please write to editor@westenglandbylines.co.uk
Friends of Bylines Network
There has never been a greater need for grassroots journalism that investigates the stories that really matter, holds power to account and champions the voices of everyday citizens. We are proudly powered by volunteers but what we do isn’t free.
STAND WITH US for independent, citizen-led journalism that makes democracy stronger, and you will even get some exclusive benefits.
