Chip War: A review

As part of my winter reading, I took on “Chip War” by Chris Miller. I saw the book recommended in the NYTimes list of notable books for 2022, and it caught my attention.

One of the things that stood up from the second section of the book “The Circuitry of the American World” was that it comes across as a mini-history of Silicon Valley (SV). I arrived in the Valley in 2015, and thought that its prominence in technology could be traced back to the 2000’s and the dot com boom. The book showed me that wasn’t the case. It turns out many small startups started in the Valley right after the second world war(e.g., HP). I seem to remember seeing the “Fairchild” name in different places in the Valley. I never knew what that meant. Turns out it was one of the first semiconductor companies established there.

The book also makes an interesting point on how U.S. technology policy shaped the destiny of places in Asia. We all now know that Taiwan is a key place in the world of chips. It wasn’t always like that. It’s rise to dominance in the field can be traced to some Taiwanese Ph.D. Students at Stanford who convinced Texas Instruments to set up a plant over there. Japan was also shaped by a great amount of technology transfer from the US. Ironically, the Japanese eventually overtook SV in producing more chips at higher quality rates and a lower price.

The book has echos to the present. In the 1980’s, Japan produced the vast majority of chips in the world, and had an increasing technological edge. SV companies suffered in the wake of Japanese competition, and many of them disappeared. The military establishment in the U.S. fretted about this. A lot of military equipment, from long-range precision bombs, to fighter jets, relied on chips made in Japan. If the U.S. lost its edge in the silicon race, it could relinquish its dominant military position. Sounds familiar? This sounds very similar to the recent developments we’ve witnessed in the U.S. recent foreign policy changes regarding China’s semiconductor industry.

The tide momentarily turned for the U.S. in the 90s thanks to the increasing importance of microchips. The 80s were dominated by companies who could sold cheap memory chips. The 90s were different. With the rise of personal computers, microprocessors became more important, and U.S. companies like Intel had an early competitive advantage in this field leading to their dominant position throughout the decade.

The book is a fascinating description on how companies with an early-advantage in the microprocessor business eventually went bust because they failed to innovate rapidly enough. Intel is the poster-child example of this. Having been an early mover on Extreme Ultraviolet Litography (EUV), it butchered its attempts to break into that market. It also failed to diversify its business from CPUs, and realize that GPUs were gaining traction for AI. A cautionary tale for many SV companies that keep an eye on short term marginal improvements, and lose sight of the long-term trends shaping the industry.

Intel seems to have failed to the same trap in the early two-thousands, when Steve Jobs asked the Intel CEO to develop a chip for the iPhone. Intel politely declined concluding that the market for cell-phone processing units wasn’t big enough to invest in it. A colossal mistake from which Intel is still reeling.

One of the eye-opening moments of the book is the realization of how the chip industry has been consolidating over the years, and now there are only a few companies that can produce the most sophisticated chips. These companies in turn rely on a small set of other companies for the spare-parts that make the machines that produce the nano metric chips.

The fact that the chip industry is so concentrated has been used as a geopolitical weapon by the U.S., which has used its clout to slow down chip development in China, and consequently China’s march towards more powerful AI algorithms.

The geopolitical advantage that the U.S. has now, can also be a dangerous double edge-sword. Taiwan sits on an area that is prone to big earthquakes. Should one hit the island and take down the places that produce these chips, the entire world would suffer the consequences of what could be a months-long chip shortage.

I had no idea of all these complexities in the chip industry, and the book was definitely interesting. It is well written, and the author seems to have done his research well. If you’re interested in geopolitics, the chip industry, or the technology industry, this book definitely has worthwhile lessons to learn.