Book Overview
Chip War was one of the few English-language books on China I could find in Phuket, so I picked it up recently. I wanted some background and a quick overview before my trips to China. I’m particularly interested in tech, and I hoped this book would shed some light on recent developments.
Unfortunately, there’s barely any useful signal, it’s more focused on a historical overview of the semiconductor industry. It does an okay job at that, but you’ll only find it interesting if you know absolutely nothing about the topic.
Chris Miller seems more focused on crafting a propaganda narrative than giving an objective, honest look at the history and latest trends in the semiconductor industry. The book is full of “boomer takes”, like equating Nazi and Soviet rule, and spreading smears about China. In this story, American companies are the only good guys, and everyone else, including their direct allies in Asia, isn’t playing fair.
The book is generally careful about making predictions, but the author can’t help himself and suggests that the US and its allies will remain unchallenged, with a 10-year lead in semiconductors. That’s a bold prediction. Let’s see how it’s holding up at the beginning of 2026.
Energy
The US generates about 4.45 TWh of electricity, while China generates roughly 10.27 TWh. Clearly, China is in a stronger position for leadership in computing. Deploying hardware at scale is no easy task, and the US is actually falling behind China in this race.
CPUs
The book correctly notes that the 2020 sanctions on Huawei dealt a heavy blow to a leading Chinese CPU designer, but it failed to kill the company. As of 2026, Huawei is back, and its CPUs are competitive enough to be sold to discerning smartphone buyers.
Huawei switched to a domestic SMIC fab. It’s not cutting-edge, but it’s certainly not 10 years behind TSMC either. Their latest mobile SoC is roughly three years behind the global frontier in raw performance. We can safely say that China has achieved self-sufficiency and will likely become globally competitive in the coming years.
RAM
China can produce its own globally competitive RAM. Its current global market share stands at about 6%, which isn’t even enough to meet domestic demand, but the technology is there and can be scaled quickly. CXMT’s latest DDR5 offerings are on par with the global competition. Neither the US nor its proxies hold a technological advantage here, and they should prepare to lose market share.
Storage
The last SSD I bought was a Lexar, guess who owns it and where it’s made? Chinese SSDs are highly competitive and already hold about 23% of the global market.
“AI”
LLMs need vast amounts of energy and world‑class researchers, and China has both. DeepSeek was a “Chinese Sputnik” moment, and most LLM benchmarks show Chinese models are highly competitive. Some argue US models are slightly ahead, but even if true, we’re talking a gap of weeks, not 10 years.
Conclusion
Chip War is essentially a historical overview laced with cope and US propaganda. It’s still a decent read if you want to understand the American perspective on this issue.
My personal, normative opinion is that no single nation should hold a decisive edge in semiconductors. A global monopoly is the worst possible state for the world, so I’m glad to see it being challenged. Ideally, I’d like more nations to prioritize self-sufficiency and reduce their dependence on foreign powers.