Chip War: A Factory Is Not Technological Independence

Morris Chang promises his customers that he will not design chips of his own. For the founder of a chipmaking company, this sounds almost like giving up the profession. In Chip War, Chris Miller shows the opposite: that promise helps make TSMC a place where other firms can bring their designs without handing them to a direct competitor.

Previously, a small company without a factory could look for spare capacity at a larger manufacturer. But its orders would remain secondary to the manufacturer’s own products. The partner also saw the customer’s design while competing in the chip market. Chang offers a different bargain: you design, we manufacture—and we do not compete for the same buyer.

The most interesting part is what success produces. Designing new chips becomes accessible to more companies, while manufacturing the most complex ones becomes concentrated in fewer fabs. Democratisation at one stage intensifies concentration at another. That is the contradiction worth looking for in the book, rather than a ready-made answer to who will win the next technology war.

A cheaper entrance, a more expensive centre

In chapter 29, Miller does not present TSMC as the triumph of a solitary entrepreneur. Behind Chang stand Taiwanese government funding, technological participation from Philips and connections to the American industry. Customer trust matters, but it needs a factory that can actually fulfil orders.

Higher volumes allow the manufacturer to spread enormous costs across many customers and improve its production process. Designers no longer have to build everything themselves; the manufacturer receives more work and accumulates experience. A convenient solution to an individual firm’s problem gradually creates an industry-wide dependency.

Chapter 37 adds another layer. Chang’s “Grand Alliance” brings together designers and suppliers of equipment, materials and intellectual property. TSMC does not simply accept a finished file. Its processes become a point of coordination for different participants. Moving an order is therefore not necessarily like forwarding it to another email address.

Miller’s strength here is a change of question. Instead of “Which country owns the chip?”, a more precise question emerges: “Who in this system can stop the next step?”

What sits behind ASML’s logo

In chapter 39, the answer begins with a droplet of tin. Producing light for EUV lithography—a way of transferring extremely fine patterns onto a wafer—requires striking it precisely with a laser. Special mirrors must then direct the light. An ordinary lamp and lens will no longer do.

Miller brings together California’s Cymer, Germany’s Trumpf and Zeiss, and the Netherlands’ ASML within this machine. The important detail is not simply the international supplier list. The required components take years to develop and must work together; the assembled machine then has to operate reliably enough for production. A logo on the casing does not mean that one company can independently reproduce everything inside.

This is a useful corrective to a simple idea of independence: buy a machine, install it in your own building and declare the problem solved. The machine brings requirements for servicing, materials, specialists and future generations of equipment. Owning the building does not remove those connections.

The next chapter, number 40, shows how the chain can narrow even without sanctions. In 2018, GlobalFoundries halts its advanced manufacturing-process programme while EUV equipment is already being installed at its facility. In Miller’s account, continuing would require expenditure that the company did not consider economically justified at its scale. Leaving the race can be a rational decision for the firm—and still reduce the alternatives available to its customers.

What has been checked since publication

Chip War appeared in 2022. Its market shares, production plans and assessments of military scenarios cannot quietly be carried into the present. A narrower check matters for this article: does dependency still exist within the supplier of complex equipment itself?

In its 2025 annual report, dated 25 February 2026, ASML writes that the complexity and uniqueness of EUV components often make sourcing from multiple suppliers economically impractical.[1] The same report identifies Carl Zeiss SMT as the sole supplier of critical optics; ASML says a prolonged loss of that supply would effectively prevent it from conducting its business.[1]

This does not establish all of Miller’s geopolitical predictions. It does support the point needed for his industrial history: dependency does not end at the familiar company name. Another irreplaceable participant may stand behind it. A factory’s new address does not, by itself, change that.

For a reader in Ukraine, this offers a practical framework for discussing a critical component: which operation depends on a single supplier, who services the equipment, and what would need to change for an alternative to exist? These are questions for a specific project, not a conclusion that Ukraine should build a particular type of fab. The book does not calculate Ukrainian costs or choose industrial policy on the reader’s behalf.

Four stops rather than the entire race

If dependency is your interest, rather than a complete history of semiconductors, read selectively: chapter 29, “We Want a Semiconductor Industry in Taiwan”; chapter 37, “Morris Chang’s Grand Alliance”; chapter 39, “EUV”; and chapter 40, “There Is No Plan B”. The first explains the bargain between designer and fab, the second the power of a coordinated network, and the last two its technical and economic cost.

A short account can convey the thesis. This route adds what makes the book itself worth reading: decisions by people and companies that made sense individually but together produced a system with few substitutes. The broader military and political history remains outside this verdict.

The thought to retain: technological independence is not an inventory of factories you own, but an answer to whose absence would stop them working.

Book source and scope

Chris Miller, Chip War, Scribner, 2022. The reading covers the introduction, chapters 29, 36, 37, 39, 40 and 54, and the conclusion. The article examines the division of industrial capabilities; it does not verify every historical or military claim in the book.

Sources

[1] https://ourbrand.asml.com/m/6ea363f69344ebd4/original/asml-2025-annual-report-based-on-ifrs.pdf — ASML Annual Report 2025 (EU-IFRS), 25 February 2026