CHIP WAR
Very few companies make chips but they can't do it alone. You can't find a better example of an interconnected, globalized supply chain than this. Chip War is an excellent account of how research, the Cold War, actual wars, capitalism, communism, government policy and individual people shaped the industry that now powers everything from phones and PCs to AI systems and modern weapons.
Semiconductor supply chains are so intertwined that perhaps it's this "silicon shield" that is protecting Taiwan from a Chinese assault. Even if TSMC, which accounts for more than 90% of advanced chip production, fell into China's hands, China wouldn't be able to scale or replicate every aspect of it. The most advanced lithography machines, the devices used to imprint transistors onto chips, are 100% controlled by ASML in the Netherlands, which in turn needs Zeiss for optical lenses and many other companies to build a single $100+ million machine. Fun fact: TSMC only prints the chips. What to print and how is decided by its customers, like Google (Tensor), Apple (Apple Silicon), NVIDIA (GPUs) and Qualcomm (Snapdragon), who in turn rely on other companies' software, from California to Israel, to design every aspect of these chips.
The current crop of companies dominating this industry (Samsung and SK Hynix in DRAM, Intel and AMD in PCs, and NVIDIA in GPUs) owe their dominance to their place in this intertwined, globalized supply chain, which was shaped by cost optimization, cold wars, tariffs and export controls.
The integrated circuit was invented in the late 1950s at American companies like Texas Instruments and Fairchild, and found its first customers in NASA and the US military, including in the Vietnam War. As companies like Texas Instruments and later Intel looked for customers outside government, they landed on consumer products like radios and calculators. Japanese companies like Sony seized on this, licensing many of the technologies to build useful end products for consumers. This was a win for America, keeping Japan in its sphere of influence during the Cold War; a win for post-WWII Japan, which used the technology to lift people out of poverty and modernize its society; and a win for early US companies, which got cost-effective mass production of chips.
This turned out to be counterproductive for US companies in the 1980s, because these chips became so commoditized that almost all US companies lost the DRAM market to Japan. Their only way out was to innovate, with Intel moving to microprocessors and the industry making transistors ever smaller. Other East Asian countries like Korea and Taiwan then followed the Japanese model of subsidizing the cost of production, land and equipment and using tariffs to promote homegrown companies, which led to the rise of Samsung, SK Hynix and TSMC in the 1990s. This also led companies to focus on one part, like design, and delegate chip production to specialized foundries like TSMC so that the unit economics work.
Geopolitics and government policies have had a huge impact on this industry and continue to do so. The Soviet Union lost the race because its industry served only the military and could not find other customers, while China and the US are now engaged in tariff, export-control and IP wars to win this race.
My thinking was semiconductor supremacy was mostly about who could design and innovate fastest. I came away realizing that no one really makes an advanced chip alone: not a company, and increasingly, not even a country. What makes the book impressive is how Miller weaves together technological history, economics and geopolitics to explain how this industry came to be.
Overall, this book is an excellent account of who, how and what made present-day Silicon Valley, and of the companies, countries and people involved in making the tiny processor in your mobile phone work.