Somewhere in the Dutch countryside, in a town most people couldn’t find on a map, sits a factory that the entire digital age depends on.
ASML, based in Veldhoven, Netherlands, is the sole manufacturer of extreme ultraviolet (EUV) lithography machines, the $380 million tools required to etch the nanoscale circuits onto the most advanced semiconductors on earth [1]. There is no alternative supplier. No workaround. No “Plan B.”
If ASML stopped shipping tomorrow, the production of every cutting edge chip in the world from the processor in a smartphone to the guidance system in a hypersonic missile would grind to a halt within months.
This is the chokepoint that makes the semiconductor industry the most strategically consequential supply chain in human history. It is the physical manifestation of a new kind of warfare. The “chip war” between the United States and China is not a metaphor for trade tension; it is an actual, kinetic struggle over the substrate of intelligence itself.
It is a conflict over who gets to build the future and who gets locked out.
The Architecture of Control
The modern semiconductor supply chain is a marvel of hyper specialization and extreme fragmentation. In the 20th century, power was measured in oil and steel; in the 21st, it is measured in transistors and lithography.
The architecture of this control is a global relay race. The United States holds the “brain”: the high level chip design and the Electronic Design Automation (EDA) software that makes complex architecture possible. Taiwan, through TSMC, holds the “hands”: the manufacturing capacity to fabricate roughly 90% of the world’s most sophisticated chips. South Korea, via Samsung and SK Hynix, controls the memory that allows these processors to function. Japan supplies the high purity chemicals and specialized materials, and the Netherlands, through ASML, holds the absolute monopoly on the light used to print them [2].
No single country controls the entire chain, but this fragmentation is precisely what enables modern American statecraft. The United States realized a fundamental truth of the 21st century economy: you do not need to own the entire supply chain to control it. You only need to control the most critical, irreplaceable nodes. If you control the “instructions” (software) and the “tools” (lithography), you can effectively dictate the technological ceiling of your adversaries.
This realization birthed the most consequential technology policy of the modern era. In October 2022, the Biden administration imposed sweeping export controls targeting China’s ability to acquire advanced AI chips and the equipment to make them [3]. These controls were not mere trade protections; they were attempts to engineer a technological ceiling. As the years progressed, these restrictions were tightened in October 2023, expanded in December 2024, and further aggressively expanded by the Trump administration in March 2025, which blacklisted dozens of additional Chinese entities ranging from AI startups to heavy industrial manufacturers [3].
The logic is mathematically unassailable: AI capabilities scale with compute. To build Artificial General Intelligence (AGI), one requires massive clusters of interconnected, high performance chips. To build those chips, one requires EUV lithography. To control the lithography is to control the speed at which a rival can achieve AI parity.
The Enforcement Problem: The Grey Market and the Shell Game
On paper, the strategy is an elegant exercise in containment. In practice, it is a high stakes game of cat and mouse played out in the shadows of global trade.
The semiconductor industry is notoriously difficult to police. These components are small, highly fungible, and can be easily disguised through complex logistics chains. The enforcement challenge is structural: once a piece of advanced equipment enters a country like China, verification becomes a diplomatic nightmare. Foreign technical teams often require months of bureaucratic approval from Chinese authorities just to perform a routine maintenance visit, during which time “upgrades” can be performed via unauthorized hardware modifications [6].
This has given rise to a sophisticated “grey market” of smuggling and deception. Huawei, the Chinese telecommunications giant that has become the central figure in Beijing’s drive for technological self sufficiency has already tested the limits of these controls. Reports indicate that Huawei used a network of shell companies to trick TSMC into manufacturing 2 million chiplets for its Ascend 910 AI processor [4]. While TSMC halted production upon discovery, the incident highlighted a terrifying reality: the components are already out there, being packaged into the next generation of Huawei’s 910C processors.
The methods are becoming increasingly surreal. In 2024, a procurement network operated through a shell company named “Luxuriate Your Life”, a name that sounds like a satire of consumerist excess, to purchase $390 million worth of servers from Dell and Supermicro containing banned Nvidia GPUs [4].
Even the architects of the technology are caught in the crossfire. Nvidia, facing a massive loss of the Chinese market, began designing “downgraded” chips, specifically tailored to meet U.S. regulatory requirements while still remaining useful to Chinese customers. In 2024, China legally imported approximately 1 million of these modified chips [5]. This regulatory arbitrage creates a perverse incentive: the more stringent the controls, the more the industry works to engineer around them, creating a constant, shifting baseline of what is legal versus what is strategic.
The Production Gap: Brute Force vs. Scaling Laws
Despite the constant leaks, the primary objective of the U.S. strategy appears to be working: it has successfully prevented China from reaching the frontier of AI compute at scale.
The disparity in production is stark. According to congressional testimony by U.S. Commerce Secretary Howard Lutnick, China’s ability to manufacture its own high end AI chips remains severely throttled; Huawei is projected to produce only 200,000 AI chips in 2025 [5]. Industry analysts at SemiAnalysis suggest that China’s total Ascend production, the cornerstone of their domestic AI push, will hit roughly 800,000 units in 2025, but will drop to as low as 250,000 units in 2026 due to bottlenecks in high bandwidth memory (HBM) [7].
Compare these figures to the United States and its allies. Nvidia alone is expected to produce 3.67 million B300 equivalent chips in 2025, a scale of production that dwarfs the entire Chinese domestic ecosystem [7].
This is not just a difference in volume; it is a difference in yield. In semiconductor manufacturing, yield is the percentage of functional chips on a silicon wafer. While U.S. allied fabricators like TSMC and Intel achieve yields north of 90% for advanced nodes, Chinese manufacturing yields for comparable processes hover between 30% and 50% [7]. This is an economic and temporal death knell. If your cost per chip is three times higher than your rival’s, you cannot build the massive data centers required to train the next generation of models.
This is where the scaling laws of AI become a geopolitical weapon. Large Language Models (LLMs) and other frontier AI systems do not just require better code; they require more compute. The relationship between compute power and intelligence is non linear. A country that is stuck at the good enough tier of 7nm or 5nm chips might be able to run existing AI models, but it will never be the one inventing the next paradigm of intelligence.
The gap between good enough and frontier is the gap between a consumer of technology and a master of it.
The Allied Alignment Problem: The Weaponization of Interdependence
The efficacy of the U.S. strategy rests on a foundation that is increasingly shaky: allied cooperation.
The Foreign Direct Product Rule is the most potent tool in Washington’s arsenal. It allows the U.S. to claim jurisdiction over any product manufactured anywhere in the world, provided that product was made using American origin technology or software.
This rule effectively turns the global supply chain into an extension of U.S. foreign policy.
However, leverage is not the same as alignment. The U.S. is currently asking its closest allies, the Netherlands and Japan, to absorb massive economic pain to serve American strategic goals. For ASML and Tokyo Electron, China has historically represented a massive portion of their revenue and growth. When the U.S. imposes restrictions, it is asking European and Japanese governments to subsidize the economic loss of their own most important technology companies to ensure American security.
This tension reached a fever pitch with the introduction of the MATCH Act in 2026. This proposed legislation would move the U.S. from entity specific restrictions (targeting specific companies like Huawei) to a countrywide ban on all advanced semiconductor manufacturing equipment to China [6]. This would effectively compel allies to stop selling even the less advanced, but still critical, lithography and etching tools. It is the ultimate expression of the weaponization of interdependence: if you want to play in the American led digital ecosystem, you must accept our embargoes, even if they damage your own national prosperity.
The Taiwan Singularity: The Silicon Shield and the Target
All roads in the semiconductor war lead to a single, highly volatile point on the map: Taiwan.
The concentration of advanced manufacturing in Taiwan is the most significant single point of failure in human civilization. Taiwan Semiconductor Manufacturing Company (TSMC) produces the overwhelming majority of the world’s sub five nanometer logic chips.
There is no alternative.
This concentration has given rise to the Silicon Shield theory, the idea that Taiwan’s indispensability to the global economy makes an invasion by China suicidal for both sides. If China attacks Taiwan, the global economy doesn’t just slow down; it ceases to function.
However, the silicon shield is a double edged sword. To the United States, the shield is a strategic vulnerability that must be mitigated. This is why Washington is aggressively incentivizing TSMC to build fabs in Arizona and why the U.S. is working to engineer away its dependence on the island.
The uncomfortable truth is that the U.S. is simultaneously relying on Taiwan’s irreplaceability for its own security and working to dismantle that very irreplaceability to ensure its own strategic autonomy. This creates a precarious paradox: the more the U.S. succeeds in its goal of onshoring chip production, the more it erodes the very shield that prevents a kinetic conflict in the Taiwan Strait.
The Global South and the Digital Divide
While the great powers fight over the available compute resources, a third, often ignored dimension is emerging: the impact on the Global South.
The world is splitting into two distinct technological spheres. One is the Frontier Sphere, powered by high end, US aligned hardware, capable of generating advanced AI for medicine, weather modeling, and intelligence. The other is the Legacy Sphere, comprised of countries that lack the capital or the political alignment to access the frontier.
The risk is a new form of Digital Colonialism.
Countries across Africa, Southeast Asia, and Latin America are becoming increasingly dependent on AI for essential services. If these nations are forced to choose between a Western AI stack or a Chinese AI stack, they are not just choosing a software provider; they are choosing a regulatory, ethical, and surveillance framework.
A country that does not own its own compute, or even the chips required to run it, will essentially be a vassal state in the digital age, consuming whatever intelligence the compute rich nations decide to export.
China’s Self Sufficiency Bet: The Red Queen Effect
Beijing’s response to these encirclement efforts has been a massive, state subsidized sprint toward self sufficiency.
China is currently engaged in a brute force approach to semiconductor development.
Lacking the ability to reach the absolute frontier through legal means, they are pouring hundreds of billions of dollars into domestic alternatives. SMIC, China’s largest foundry, has achieved 7nm chip production, a feat that caught Western analysts off guard and proved that, with enough capital and state direction, China can bypass certain levels of technological denial [3].
While Chinese made chips are currently less efficient, more expensive, and harder to scale than their Western counterparts, they are good enough for many applications. For the Chinese military, good enough may be sufficient to maintain parity in many domains of electronic warfare and autonomous systems.
This creates what biologists call the “Red Queen” effect, a state of constant, rapid evolution where you must run as fast as you can just to stay in the same place. The U.S. builds a better chip; China builds a bigger factory. The U.S. tightens an export rule; China subsidizes a new domestic supplier. The gap between good enough and frontier is the only thing keeping the current peace, and that gap is narrowing every year.
VIII. Conclusion: The Cost of the Wall
The chip war is often framed as a trade dispute, but it is actually a struggle over the very substrate of 21st century power. To control the chip is to control the ability to process information, to automate intelligence, and to exert influence over the physical world through digital means.
The cost of building the wall around China is immense. It fragments the global market, increases the cost of technology for everyone, and accelerates the very self-sufficiency it was meant to prevent.
By cutting China off from the leading edge, the West is essentially forcing China to build its own ecosystem, an ecosystem that will eventually be independent, un-monitored, and entirely beyond Western influence.
The question is not whether the export controls will work in the short term. They are working. They are successfully slowing China’s ascent into the AI frontier. But the real question is what happens when the wall stops mattering. If the goal of these controls is merely to buy time, the world must ask: what are we doing with that time? Are we using it to build a durable, sustainable technological lead, or are we simply delaying the moment when the chip war becomes a permanent, bifurcated reality for the entire human race?
Everything else is downstream.
Verified Sources & References
Primary Sources & Data Points:
[1] ASML Lithography Technology: For detailed specifications on EUV (Extreme Ultraviolet) lithography and the manufacturing complexities of $380M machines. (Reference: ASML Annual Reports/Technical Documentation).
[2] Semiconductor Supply Chain Analysis: On the division of labor between US (EDA/Design), Taiwan (Foundry), South Korea (Memory), Japan (Chemicals), and Netherlands (Lithography). (Reference: SIA - Semiconductor Industry Association).
[3] US Export Control Frameworks: On the October 2022 and October 2023 BIS (Bureau of Industry and Security) rules regarding advanced computing and semiconductor manufacturing items. (Reference: U.S. Department of Commerce/BIS).
[4] Huawei/TSMC & Shell Company Incidents: On the use of third-party procurement and the “grey market” for advanced GPUs. (Reference: Reuters/Financial Times investigative reports on semiconductor smuggling).
[5] AI Compute and Production Projections: On the production volumes of Nvidia H200 vs. Chinese domestic alternatives like Huawei Ascend. (Reference: Congressional Testimony/Industry Analyst Reports).
[6] Enforcement and Verification Challenges: On the difficulties of inspecting semiconductor fabs in restricted jurisdictions. (Reference: CSIS - Center for Strategic and International Studies).
[7] SemiAnalysis/Yield and Production Data: On the comparative yields and production volumes of SMIC/Huawei vs. TSMC/Nvidia. (Reference: SemiAnalysis Industry Intelligence).
Contextual References for Further Reading:
The “Silicon Shield” Concept: Originating from discussions by Taiwanese strategists regarding the economic deterrent of TSMC.
Weaponization of Interdependence: Theoretical framework by Henry Farrell and Abraham Newman regarding the use of global networks as tools of statecraft.
The Foreign Direct Product Rule (FDPR): The specific legal mechanism used by the US to exert extraterritorial jurisdiction over non-US companies.