Chip War: Why Semiconductors Determine the Balance of Global Power

Chris Miller reveals how the battle for semiconductor supremacy has become the defining geopolitical struggle of our time, with Taiwan at the epicenter of global vulnerability.

Key Takeaways

• The Soviet Union's failure in semiconductors despite scientific prowess stemmed from lacking consumer markets and international supply chains that drove US innovation
• Taiwan produces 90% of the world's most advanced processor chips and one-third of all new computing power, creating unprecedented global dependence on a single vulnerable location
• China imports more money in semiconductors than oil, spending tens of billions annually to develop domestic chip capabilities in an "unparalleled effort" in technological history
• Modern chip manufacturing requires $20+ billion facilities with machinery costing $150 million per tool, creating insurmountable barriers to entry for most nations
• ASML of the Netherlands holds an absolute monopoly on extreme ultraviolet lithography machines essential for advanced chips, representing a critical Western chokepoint
• Moore's Law has driven exponential progress from 4 transistors in early 1960s chips to over 10 billion transistors in modern iPhone processors
• The industry has specialized into distinct models: fabless design companies, dedicated foundries like TSMC, and integrated manufacturers like Intel
• Xi Jinping's focus on political control over GDP growth signals a shift from economic interdependence theory, increasing risks of military conflict over Taiwan
• Semiconductor equipment manufacturing is dominated by five companies that control lithography, deposition, etching, and metrology processes essential for advanced production

Timeline Overview

• 00:00–15:30 — From Russian History to Chip Geopolitics: Miller's journey from studying Soviet economic history to understanding why the USSR failed at semiconductors despite comparable nuclear and space capabilities, revealing early importance of consumer markets and international supply chains
• 15:30–30:45 — Taiwan Vulnerability and China's Challenge: Discussion of how Taiwan's dominance in advanced chip production creates global fragility, while China's massive domestic investment threatens to reshape both commercial and military balance of power
• 30:45–50:20 — Economic Interdependence vs Strategic Competition: Analysis of why mutually assured economic destruction theory may not prevent conflict, using Russia's invasion of Ukraine as cautionary example of dictators prioritizing politics over economics
• 50:20–70:15 — Moore's Law and Technical Foundations: Deep dive into semiconductor physics, from transistor miniaturization to manufacturing processes involving components smaller than coronavirus particles, requiring unprecedented precision engineering
• 70:15–90:30 — Industry Structure and Specialization: Explanation of how semiconductor ecosystem evolved from integrated companies to specialized players including fabless designers, foundries, and equipment manufacturers with distinct competitive advantages
• 90:30–110:45 — Manufacturing Process Complexity: Detailed walkthrough of chip production from ultra-pure silicon and specialized chemicals through lithography, deposition, etching, and testing processes requiring thousand-step procedures
• 110:45–125:00 — Global Supply Chain Dependencies: Analysis of raw material sourcing, rare earth concentrations, and geographic distribution of critical capabilities from Dutch lithography tools to Taiwanese manufacturing

The Soviet Semiconductor Failure: Lessons for Today's Competition

• Despite comparable scientific expertise and early success in nuclear weapons and space technology, the Soviet Union "could never find a way to miniaturize computing power," creating fundamental economic and military disadvantages
• Soviet scientists won Nobel Prizes in semiconductor physics, demonstrating that "it certainly wasn't scientific expertise that the Soviet Union lacked" but rather systemic structural disadvantages
• The USSR lacked two critical elements that drove US success: first, "a consumer industry that was nearly as big as the US" and second, "an international supply chain" enabling access to advanced components and talent
• US chip firms quickly pivoted from military to consumer applications, with "by the 1970s most chips being sold for corporate computers or consumer devices like calculators," providing vast markets to fund innovation
• Soviet isolation meant relying on "Machine Tools from Bulgaria or Hungary which were just far less advanced than what the US could get from Western Europe or Japan"
• Immigration proved crucial for US innovation: "some of the key innovations in the chip industry came from foreign-born engineers working in Silicon Valley whereas the Soviets had almost no immigration"
• The global focus enabled US companies to "buy components from those countries as well and sell to customers in Europe or Japan the world over," creating scale and specialization advantages

This historical lesson demonstrates how technological competition requires not just scientific capability but also market access, international collaboration, and economic incentive structures that centrally planned economies cannot replicate.

Taiwan's Critical Vulnerability and China's Response

• Taiwan's semiconductor dominance creates unprecedented global vulnerability: "Taiwan produces around one-third of the new computing power we rely on each year" and "90 percent of the most advanced processor chips"
• A conflict in the Taiwan Straits would devastate global technology: "entire segments of Industry would just grind to a halt from Autos to smartphones to high performance Computing"
• This dependency has grown precisely as military risks increased: "we've become more Reliant on Taiwan even as the risk of War around Taiwan is at the highest it's been in 25 years"
• China's response represents historically unprecedented technological mobilization: China "spends more money importing chips than it does importing oil" and is pursuing "an effort that is really unparalleled in history to domesticate chip technology"
• Chinese leadership fears US leverage: "this gives the US leverage over China that can be used both during peace and during war time and so they're desperate to find a way to produce the technology they need domestically"
• Success would fundamentally alter military balance: "China's investment in domestic ship making will have a direct impact on the military balance of power" as both militaries invest heavily in autonomous systems and electronic warfare requiring advanced semiconductors
• The timeline matters crucially: "if China's chip Investments succeed and to some extent they are succeeding...China's military will get more powerful and as China's military gets more powerful the risk of War on the Taiwan Straits grows"

This creates a dangerous feedback loop where China's efforts to reduce dependency on Taiwan paradoxically increase the likelihood of conflict that would disrupt global semiconductor production.

The Limits of Economic Interdependence Theory

• The traditional theory of mutually assured economic destruction assumes leaders prioritize GDP growth and fully understand economic interlinkages, assumptions increasingly questionable under Xi Jinping's leadership
• Miller cites 2022 Russia as proof that "we've already got one good example of a dictator being fed bogus information by his Security Services and launching a disastrous war" despite economic costs
• European reliance on Russian energy demonstrated how "economic interdependence for several decades to restrain the Russians...worked for several decades until it didn't"
• Xi Jinping's leadership represents a departure from previous Chinese prioritization of economic growth: "for four decades after Deng Xiaoping consolidated power in China it was pretty obvious that Chinese leaders were focused on GDP growth above all"
• Recent Chinese policies contradict pure economic rationality: cracking down on private tech sector "in a way that was really costly for financial markets and GDP growth" and pursuing zero-COVID policies with "devastating effects for the Chinese economy"
• The shift suggests Chinese leaders now "focus on their own political power at home or nationalistic Pursuits" rather than maximizing economic outcomes
• This evolution toward leaders who "might be willing to roll the dice, gamble with economic consequences to try to retake the island" undermines deterrence strategies based on economic costs

The implication is that strategies assuming rational economic calculation may prove inadequate against leaders prioritizing political survival or nationalist objectives over GDP optimization.

Moore's Law and the Physics of Exponential Progress

• Semiconductor advancement follows Gordon Moore's 1965 prediction of exponential miniaturization, progressing "from 4 to 10 billion transistors" on processors from early 1960s to modern iPhones
• Each transistor represents "tiny electrical circuits that turn on and off" creating "the ones and zeros that undergird all Computing and all software," making miniaturization directly proportional to computational power
• Three parallel trends enable Moore's Law: "smaller transistors, more cost-effective transistors and more energy efficient transistors" running simultaneously for six decades
• Modern manufacturing achieves precision "smaller than a coronavirus with almost 100% accuracy by the billions," representing "some of the most complicated engineering Feats in human history"
• The design challenge has evolved from hand-drawn layouts to software managing billion-component chips: "if you got a billion or more transistors on ship you need really complicated software to lay it all out"
• Manufacturing tools utilize "the flattest mirrors, some of the most powerful lasers" and can "move individual atoms in some cases, layering on ultra thin layers of chemicals"
• The precision requirement extends beyond laboratory proof-of-concept to industrial reliability: "how can you make a process happen 24 hours a day with basically perfect accuracy surrounded by all sorts of other explosive chemicals"

This exponential progress creates the foundation for all digital technology while simultaneously raising the barriers to entry for new competitors seeking to match leading-edge capabilities.

Industry Specialization and Competitive Dynamics

• The semiconductor industry has evolved from vertical integration toward extreme specialization, with early companies like Texas Instruments producing "their own silicon Wafers, refined their own chemicals in-house, produced their own Machine Tools"
• Modern chip production requires "dozens of different companies" providing specialized components, chemicals, tools, and software, enabling firms to "focus on what they do best and therefore Drive technology forward"
• The fundamental industry division separates design from manufacturing: "fabless firms" like Nvidia and Apple focus solely on chip design while foundries like TSMC specialize exclusively in production
• TSMC's dominance stems from scale advantages: as "the world's biggest producer of processor chips" it becomes the platform around which "the entire chip ecosystem has begun circling"
• Integrated Device Manufacturers like Intel struggle because "it's just hard to be good at everything" while specialized firms achieve superior results in their focused domains
• Samsung faces unique challenges operating both foundry services and consumer electronics, creating customer wariness: "customers are very wary of going to Samsung and saying hey can you produce my smartphone chip"
• The ecosystem effect reinforces TSMC's position: "for the Machine Tool makers it's their biggest customer, for the design software firms they've got to have their design software optimized to work with TSMC"

This specialization creates both efficiency gains and strategic vulnerabilities, as critical capabilities become concentrated in a small number of irreplaceable companies.

Manufacturing Process Complexity and Chokepoints

• Advanced chip production requires materials of unprecedented purity, starting with "Ultra Ultra Pure silicon" from "a couple of firms, one in Germany, one in Taiwan, a couple in Japan"
• The lithography process uses "shooting light in specific patterns at chips" interacting with photoresist chemicals that "react with light in specific ways" to create nanometer-scale structures
• ASML of the Netherlands holds "absolute Monopoly on the production of the most advanced lithography tools" with machines costing "$150 million dollars a piece, making them the most advanced and expensive Machine Tool in human history"
• Chip production involves thousands of steps: "from the time a design reaches a Fab to the time a completed ship exits the Fab there could be a thousand or two thousand different process steps"
• Five companies dominate essential equipment: "ASML, Applied Materials, Lam, KLA and Tokyo electron really have a dominant position such that you really can't operate an advanced Fab today without buying some equipment from all of them"
• The deposition process requires "Ultra Ultra flat layers that are very very very thin" while etching creates "little Canyons in ships that are a couple of nanometers wide" with near-perfect accuracy
• Quality control demands sophisticated inspection: "you've got to be able to see these nanometer scale structures" requiring "really complex microscopes and a lot of computing power" to identify potential defects

This complexity creates multiple potential chokepoints where a small number of specialized suppliers control capabilities essential for the entire global industry.

Capital Requirements and Economic Barriers

• Modern chip manufacturing requires massive upfront investment: "if you want to make an advanced logic chip Fab...it'll cost you around 20 billion dollars plus or minus"
• Half of facility costs go to equipment: "around half will go to purchasing the Machine Tools inside the Fab" which must be updated "every couple of years because there are new generations of tools"
• TSMC's annual capital expenditures exceed military spending: "many tens of billions of dollars every year...more than an aircraft carrier" which "costs 10 billion dollars"
• Once facilities are operational, "the cost of personnel, the cost of the actual silicon, the chemicals are relatively small in comparison" to capital expenditures
• Leading-edge chips represent high-value market segments: "around half of logic chips that you produce in Revenue terms are Leading Edge" because smartphones, PCs, and data centers require latest technology
• The economics favor continuous investment: facilities built in early 1990s "are still online" after depreciating capital costs, producing chips "quite cheaply" at older technology nodes
• Research and development spending matches biotech/pharmaceutical industries as percentage of revenue, with "chip firms spend a ton of money in r&d because they've got to be rolling out new generations of Technology every couple years"

These capital requirements create enormous barriers to entry while rewarding companies that achieve scale, explaining why the industry consolidates around a small number of leading manufacturers.

Looking Forward: The Reshoring Challenge and Strategic Implications

Chris Miller's analysis reveals semiconductors as the ultimate strategic technology, where control over production capabilities determines both economic prosperity and military power. The industry's extreme complexity and capital requirements create natural monopolies and chokepoints that make rapid geographical redistribution of production capabilities extraordinarily difficult.

Future Semiconductor Geopolitics Predictions

• Bifurcated supply chains will emerge as US and Chinese technological ecosystems decouple, creating parallel but incompatible semiconductor industries with significant efficiency losses and innovation constraints
• Taiwan vulnerability persistence will continue despite reshoring efforts, as the island's manufacturing expertise and scale advantages cannot be quickly replicated elsewhere, maintaining global economic exposure to conflict risk
• Technology competition intensification will accelerate as both superpowers recognize semiconductor supremacy as essential for military dominance in autonomous weapons, electronic warfare, and artificial intelligence applications
• Capital requirements explosion will make semiconductor manufacturing increasingly concentrated among a few nations and companies capable of sustaining multi-decade investment cycles exceeding $100 billion
• International alliance restructuring will occur around semiconductor access, with countries aligning based on technology partnerships rather than traditional diplomatic relationships
• Innovation bottlenecks may emerge as extreme specialization creates dependency on single suppliers like ASML, making technological progress vulnerable to geopolitical disruption or natural disasters
• Economic weapon weaponization will expand as both US and China use semiconductor trade restrictions as primary tools of international coercion, potentially triggering broader technological conflicts