Abstract

Since October 2022, the United States has restricted Chinese access to advanced semiconductors and the equipment used to produce them, extending those measures through coordination with the Netherlands and Japan. This paper asks how those controls have affected international trade and how the World Trade Organisation should respond. The analysis is descriptive and comparative, examining trade restrictions, supply chain reorganisation, corporate and state responses, and patterns of allied corporations and conflicts. It applies the framework of weaponised interdependence which holds that states with authority over concentrated intersection points in global production networks can convert that position into political leverage. Three comparisons structure the evidence: the Cold War CoCom regime set against the present arrangement; the different measures adopted by the United States, the Netherlands and Japan; and the American restriction of semiconductors alongside China’s restriction of critical minerals. The paper finds that the controls have narrowed Chinese access to leading edge technology without halting its development and have instead accelerated domestic substitution and provoked retaliation at other chokepoints. Effectiveness depends less on the formal financial limits of the measures than on the target’s integration into the world economy and the internal cohesion of the restricting coalition. Compliance costs have risen and supply chains are fragmenting along political rather than commercial lines. The paper concludes that the security exception under GATT Article XXI requires procedural discipline rather than abolition. Members should be obliged to justify security measures and to demonstrate a genuine connection to the threat claimed, so that the exception does not absorb the rules-based trading system.

Introduction

Government export controls on advanced technology, such as semi-conductor and artificial intelligence (AI) chips are affecting global trade and economic governance. Demand for chips that can process large amounts of data and run/train AI models is growing so rapidly that many governments are finding it necessary to step in. Advanced chips are required in nearly every industry imaginable, such as defence, cloud computing and nearly every hardware or software company. Access to these chips is both an economic issue and a matter of national security. The companies that produce these chips, on the other hand, are few and far between because of the highly specialised manufacturing steps involved. Not only are they extremely hard to produce, but the companies that do essentially hold monopolies on them. Because these companies are sparse, the country of the company that produces chips sometimes has the ability to limit or stop rival nation’s access to advanced computing power and AI chips.

Since the first Trump administration, the United States has steadily tightened export controls on advanced semiconductor technology to China, reflecting growing national security concerns over China’s access to advanced computing and AI capabilities. It is the hypothesis of this paper that the restrictions mentioned have limited China’s access to the most advanced chips, but have not fully stopped its technological development. Instead, the controls may have encouraged greater investment in domestic semiconductor production and alternative supply chains. At the same time, the growing use of export restrictions may be increasing fragmentation in global technology trade and weakening the principles of open international trade that institutions such as the World Trade Organisation (WTO) are meant to protect.

Literature Context

The first thing to realise is that the process of producing AI chips is no longer done by a single company in a single building; this idea is founded on an outdated principle. Instead, what occurs in modern society is an interconnected global value chain, with certain development steps acting as chokepoints, limiting other nation’s access. This operates with different countries taking up specific roles in the development of the chips. For example, the US is behind the most valuable process within chip-making: the blueprint design and producing the software required to enable chip-making. China, by contrast, is where much of the testing, packaging and transmission into final products – such as smartphones – usually occurs. Essentially, countries employ teamwork on a global scale to optimise this industry, making the processes faster and cheaper. Different companies often divide roles between designers and manufacturers. The designers choose the intended functionality of the chips, along with its features. The manufacturing process follows and requires specialised factories called fabs in order to complete it. This requires highly advanced technology to effectuate, the most recent of which is extreme ultraviolet (EUV) lithography. This utilises a specialised type of light to draw out the small and powerful circuits necessary for AI chips. Companies in charge of manufacturing are known as foundries. The design process is completed by fabless companies. Primary examples include the fabless company Nvidia and the foundry TSMC. TSMC, in Taiwan, controls the massive $10 billion factories that actually build the products. EUV lithography, however, has scarce availability and is not readily accessible to every manufacturer. Factories that do not possess this latest technology instead rely on deep ultraviolet (DUV) machines. These machines would take multiple uses to create the high level of detail used for advanced chips that EUV machines can do in a single try. However, access to EUV is pivotal to advanced chip manufacturing because multiple attempts with DUV machines is far less efficient and significantly more costly. 

The most central principle thus far was the interdependence of countries in industry, the fundamental framework upon which the nations rely on for global trade. This presents both a major strength in terms of reliable teamwork and a weakness in the reliance of countries on one another. Because of interdependence, any one nation cannot survive in the global economy on its own. While the nations working together makes things cheaper, there remains the underlying vulnerability of each relying on one another. Because certain specialised jobs, such as producing cameras or software, only occur in one or two places, those nations have chokepoint power. This is the ability of one country to use the fact that it has control over an important and irreplaceable part of the interdependent supply chain to place political pressure over other nations. For example, the US controls the design software, and the Netherlands controls lithography because ASML is the only firm capable of producing the machines that can print the world’s most advanced circuits. Japan adds to this concentrated power by providing more than 50% of the world’s essential chemicals and materials needed inside these factories. With this concentration of specialised activities, if any one chokepoint is shut down, the entire network of global production could wind up frozen. This is how the United States could remotely control huge parts of manufacturing in other countries from a distance. For instance, Taiwan and South Korea can be controlled because their fabs would not run without the software received from America. This set of chokepoints exploded in value as a result of the rise of AI. Training the latest AI models requires thousands of high-performance chips working in unison. These processors are “dual-use technologies”, meaning the exact same chip used for a commercial cloud service is also essential for military surveillance, weapons development and intelligence gathering.

Another business that clearly underscores similar effects of interdependence is the global semiconductor industry. Semiconductor production is split among several countries with distinct roles. American countries often design the chips, Taiwanese foundries manufacture them, Dutch companies provide essential lithography equipment and Japanese firms supply key materials. This specialisation makes production more efficient, but it also creates chokepoints. If one country restricts access to a crucial technology or component, the entire supply chain may be disrupted. Semiconductors therefore demonstrate how global trade can generate both economic cooperation and political vulnerability.

Relatedly, there was a fundamental shift on 7th October 2022. The Bureau of Industry and Security (BIS) of the United States stopped only sanctioning particular companies. Instead, it went for degrading the entire Chinese infrastructure for the next generation of AI. This is the Small Yard, High Fence policy: household appliances will have a large yard where basic chips will be traded freely, but the most powerful AI technology will be enclosed within a small yard surrounded by a high fence of strict controls. The United States then tried to implement this policy worldwide through the Foreign Direct Product Rule (FDPR), a tool that applies American law to foreign-made products if they use American technology at any stage.

The industry initially responded with a cat-and-mouse game of workarounds. For example, Nvidia designed modified chips like the A800 and H800, which were partially less efficient, to stay just below the banned speed limit while still supporting AI functions. But in 2023 and 2024, the United States again tightened the screws, closing loopholes, lowering speed thresholds and banning high-bandwidth memory (HBM) to keep a competitor’s AI two generations behind. However, these restrictions backfired in an event known as the “Huawei effect”. Instead of stopping technological progress, the bans inspired Chinese companies to become more self-sufficient. For example, Huawei began developing its own computer chips and operating systems to replace Western technology. It had the opposite effect from what the US intended. Huawei surprised the world by developing its own indigenous chips and operating systems to replace what it could no longer buy from the West. China also responded aggressively. It restricted exports of important minerals, such as gallium and germanium, which are essential for producing advanced technology. 

These actions have caused significant problems for the WTO. Countries are normally expected to follow international trade rules, but GATT Article XXI allows them to break those rules if it’s for reasons of international security. The United States argues that each country should decide for itself when this exception applies. However, if every country used national security as an excuse to protect its own industries, the global system of open trade could become seriously weakened. Overall, there’s a clear limitation in that production depends on a small number of specialised countries. With this system, if one important part of a supply chain is disrupted, production becomes heavily disrupted on a global scale. This remains an unresolved weakness in global trade.

Methodology

To elaborate on this back-and-forth between China and the US, this paper will analyze how the US-led export controls have affected advanced-technology trade with China and the international semiconductor supply chain since 2022. The analysis focuses on four main areas: trade restrictions; changes in supply chains; company and country responses; and international cooperation or conflict. This approach will break apart the conflict and reveal its inner working layers, to not rely on simple import and export values alone. Restrictions can reduce direct sales while also encouraging firms to redesign products, shift production, seek alternative markets or reorganise their supply chains.

The primary method of restricting trade to China by the US is an elaborate system of exportation controls and licensing requirements. Instead of completely shutting down all semiconductor trade, US policy targets specifically the most advanced chips, equipment required to produce them and companies that are seen as national security concerns. As far as actual practice and enforcement goes for these policies, the US Department of Commerce’s Bureau of Industry and Security (BIS) handles all administration regarding exportation. Starting in 2018, and rapidly continuing after 2022, the US has expanded and broadened restrictions, aiming to limit China’s semiconductor and AI field as a whole (Sutter, 2025). In particular, the October 2022 controls added restrictions on certain advanced computing chips, integrated circuits and semiconductor manufacturing equipment, while later rules expanded and adjusted these restrictions.

Arguably the most important mechanism is the entity list, which places licensing requirements on particular foreign companies or organisations. Earlier iterations followed the same company by company method, specifically targeting groups of interest like Huawei, but the US later moved to a more broad spectrum, with restrictions placed not only on individual companies but also on advanced chips as a whole and the equipment required to make/produce them. Another major mechanism used by the US is Foreign-Directed Product Rule (FDPR). It essentially extends already existing rules to those of products actually produced outside of the US, if they relied on US software or equipment. This allows the US to have control over products not made within their borders, and is especially important for the semiconductor industry which has production across numerous countries, not just the US. 

While these mechanisms somewhat work by themselves, a large factor in their success is international coordination. American influence over the semiconductor supply chain is broad and significant, but does not extend everywhere. The Netherlands contains ASML, the only producer of the most advanced EUV lithography systems, while Japanese firms also supply important semiconductor manufacturing equipment. Because alternative foreign suppliers can weaken a purely American restriction, the United States has pushed allies such as the Netherlands and Japan to implement similar controls. Allied cooperation is extremely relevant to the success of American foreign export policy (Branstetter, 2024).

For the purpose of the paper, each mechanism can be evaluated through several key questions: How broad is restriction? How far beyond the regulating country does it reach? How closely are allied controls coordinated? And finally, how easily can the targeted country or its firms adapt? These questions provide a common framework for comparing the modern US-led controls with other case studies such as CoCom, differences among American, Dutch and Japanese policies, and China’s use of restrictions on critical minerals.

Theoretical Framework

The most important working theory used to understand these policies is weaponised interdependence. Developed by Henry Farrell and Abraham Newman, their argument challenges the idea that economic interdependence distributes power evenly between countries. International networks often become concentrated around a concise, limited number of important firms, technologies, financial systems or other central sites. States or countries that have legal authority over these nodes can sometimes use that position for political or strategic purposes. The most important, and most destructive, form of this power is known as the “chokepoint effect”, in which a state can restrict another’s access to an important international network (Farrell & Newman, 2019).

The semiconductor industry closely fits this framework, if not being completely explained by it. Semiconductor production is international, but the individual stages of production are not equally distributed, much like the theory underlines. While American firms hold strong positions in areas such as chip design software and semiconductor manufacturing equipment, the Netherlands dominates the most advanced lithography systems through ASML, and more advanced fabrication is concentrated among a relatively small number of firms in East Asia. This means that control over only one part of the supply chain can potentially influence production occurring thousands of miles away. However, while the United States accounts for only a small share of global semiconductor fabrication, foreign manufacturers still depend heavily on American design software, equipment and other inputs, which can fall under FDPR supervision (Branstetter, 2024).

The FDPR then demonstrates this theory in practice. Because foreign semiconductor producers depend on American technology, the US can use control over those technologies to influence whether foreign producers are allowed to supply to China. The power therefore comes not from the American market, but from the position American technology occupies within the wider production network. Allied cooperation strengthens this chokepoint further by limiting possible substitutes.

However, when countries use weaponised interdependence as a strategy, it can backfire severely. If a country realises that dependence on a foreign technology can be used against it, it has a strong incentive to reduce that dependence. It is very likely that American export controls are what influenced Huawei to replace controlled foreign components with Chinese alternatives, and that Huawei accomplished some of this substitution much faster than expected. It is also worth noting that continuously widening sanctions may eventually face declining effectiveness (Branstetter, 2024). This creates a central friction or tension: export controls can create strong short-term leverage, while their extended use may encourage domestic Chinese production, alternative supply chains and a more divided international technology market, which violated free-trade policy and the goals of the WTO.

Institutional Framework

The most prominent institution responsible for governing international trade conflicts like these is the WTO. WTO rules and policy generally hope to create predictable and consistent trade between countries. However, those rules also recognise that governments may sometimes restrict trade in order to protect national security. The specific document is Article XXI of the General Agreement on Tariffs and Trade (GATT), commonly called the security exception for its governing outline. 

Article XXI allows states or countries to take specific actions considered necessary to protect “essential security interests”. The article specifically refers to areas of high interest (but not fully limited to), including arms and military supplies, nuclear materials and actions taken during war or another emergency within the context of international relations. The difficulty, and underlying problem with this framework, is essentially where to draw the line. How much freedom should governments have when defining those security interests? If governments could label any economically useful restriction a national security measure, the article could be abused to work around WTO obligations. On the other hand, allowing external organisations to decide what represents an actual security threat could interfere with decisions that states consider fundamental to their sovereignty.

This disagreement directly affects the semiconductor conflict. In December 2022, China opened a WTO dispute, which challenged US measures in advanced computing chips, semiconductor production equipment, supercomputer items and other related technologies. China argued that the measures appeared inconsistent with obligations under WTO agreements. The United States stated that the measures involved national-security issues that were not suitable for WTO review or resolution. This particular dispute has remained at the consultation stage rather than advancing to a panel, leaving the legal conflict unresolved.

Comparative Case Studies

This section compares three cases: CoCom and the controls introduced after 2022; the different measures adopted by the United States, the Netherlands and Japan; and the American restrictions on semiconductors alongside China’s restrictions on critical minerals. Each comparison changes one condition and keeps the others fixed: the period, the implementing state or the actor using the instrument. This makes it possible to separate the effects of the controls themselves from the conditions around them. The section closes by asking why measures of this scale have developed with so little multilateral adjudication.

The CoCom Precedent

Policymakers often cite CoCom as evidence that coordinated technology controls can work. The historical record suggests otherwise. Lind and Mastanduno (2025) compare CoCom with the current controls on China and find that the same three problems are appearing again, this time under less favourable conditions: controls that start out narrow tend to grow wider over time; CoCom slowed Soviet access to technology but did not prevent it; and cooperation between allies weakened even though conditions were more favourable than they are today.

Two differences separate the periods. The first concerns the target economy. The Soviet Union was largely closed off from Western trade, so blocking technology at the border was a realistic goal. China is deeply integrated into the world economy and is the largest trading partner of most of the states now being asked to restrict it, which makes the same approach far more difficult. The second concerns institutional form. CoCom was a standing body. The alliance gathered in Paris regularly and negotiated a common list of controlled technologies. The current regime began as a unilateral American rule in October 2022 and spread later through bilateral persuasion (Sutter, 2025). It has no common list and no shared mechanism, so coordination rests on voluntary alignment rather than on any agreed obligation. It is difficult to separate the effect of the controls from the other factors. The Western technological lead may have owed as much to the inefficiency of the Soviet economy, its weakness in developing civilian technology, as to the embargo itself (Lind & Mastanduno, 2025).

Divergence Among Allies

The three governments reached an understanding in January 2023 and then implemented it differently. American rules target China directly through the entity list and through the FDPR, which extends United States jurisdiction to goods made abroad with American technology (Sutter, 2025). Dutch rules operate through a country-neutral licensing system that does not name China at all, although the most advanced lithography systems are barred outright. Japan restricted equipment exports on similar lines. Allen and Goldston (2025) find that allied measures are weaker than the American regime in three respects: they do not restrict the activities of Dutch or Japanese nationals in China; they do not name the target; and the legal powers behind them are not equivalent. The reluctance to name China recalls CoCom, which never acquired treaty status because several members preferred not to appear openly discriminatory (Lind & Mastanduno, 2025).

The divergence reflects unequal exposure and unequal legal capacity rather than differing commitment. ASML is a single firm of great importance to the Dutch economy and its sales to China were high, so restriction carried a direct national cost. The United States did not face the same problem because its design firms remained dominant even without the Chinese market. Dutch authority was also narrower. As Allen and Goldston (2025) show, allied governments did not have export control powers equal to those of the United States because their systems had not been built for this purpose. The same commitment therefore carried unequal costs and the national measures diverged accordingly.

The trade effects are measurable but need careful reading. China’s share of ASML’s net system sales fell from about 36% in the fourth quarter of 2025 to about 19% in the first quarter of 2026, while the company’s total net sales and its full-year guidance both rose (ASML, 2026). Nvidia recorded no sales of restricted processors to Chinese customers across several quarters, yet its total revenue grew by 85% year on year with 81.6 billion dollars (NVIDIA, 2026). The cost of restriction therefore fell mainly on the target rather than on the coalition. This was not the case under CoCom, where allied firms carried significant and politically visible losses, and it helps explain why the current coalition has held together despite much weaker institutional machinery. The burden was also unevenly distributed across sectors within the coalition. Lithography and equipment suppliers in the Netherlands and Japan bore the revenue loss, while American chip designers did not.

Chips and Rare Earths

The third comparison sets the American restriction of semiconductors alongside China’s restriction of critical minerals. Both rest on control of a single point in the supply chain. The United States and its partners control design tools and lithography equipment; China refines roughly 90% of the world’s rare earth elements, although it mines a considerably smaller share, closer to 60% (International Energy Agency, 2025). The methods are also similar. China now requires a licence for any foreign product containing 0.1% or more of Chinese-origin rare earths, which extends its authority beyond its own borders in much the same way the American FDPR does. The two positions differ in how long they last. Equipment chokepoints rely on technical capability built up over decades and cannot be replicated quickly. Refining capacity rests on investment and regulatory tolerance, and could in principle be rebuilt elsewhere within about a decade. Chinese controls, however, take effect immediately and across a much wider range of industries while the effects of chip controls accumulate slowly. The case also shows that restriction produced escalation rather than compliance. China did not moderate its position. It expanded its mineral controls in response to successive American measures and instructed its own technology firms to stop buying certain American processors, a step that serves both as industrial policy and as a bargaining position (Woods, 2025). Evidence of circumvention points the same way, although it must be treated with caution: estimates of smuggled processing capacity reaching China by the end of 2025 range from roughly 290,000 to 1.6 million high-end equivalents, a spread of more than five times (Juniewicz, 2026). Even the lowest figure suggests that control of a chokepoint converts into practical denial far less reliably than its concentration implies.

Findings

Three findings hold across the comparisons. First, the formal strength of a control regime is a poor guide to its effect. CoCom had a standing multilateral body, an agreed list and an isolated target, and still managed to achieve delay rather than denial (Lind & Mastanduno, 2025). Second, coalitions fracture along lines of unequal cost. The gaps in the Dutch and Japanese measures were not oversights but reflections of what restriction was worth to each government. They have proved durable rather than temporary (Allen & Goldston, 2025). Third, control of one chokepoint invites retaliation at another. Rather than producing compliance, the American measures prompted China to develop its own instruments (Woods, 2025), which is the pattern the literature on weaponised interdependence would predict where more than one actor holds a central position in the network (Farrell & Newman, 2019). A fourth observation cuts across all three. Each case involved substantial restriction of trade justified on security grounds, and none were resolved through multilateral adjudication. China’s complaint has remained at the consultation stage since December 2022 without a panel being established (WTO, 2022) and the legal category at issue remains, as Svetlicinii and Su (2025) put it, an unsettled battleground.

Discussion

US-led export controls have succeeded in the short term but are less promising in the longer term. The restrictions have reduced China’s access to the most advanced American chips, design software and manufacturing equipment. They have also delayed the expansion of some Chinese semiconductor firms and made it more difficult for China to produce leading-edge chips at the same pace as the United States and its allies. However, this does not mean the complete denial of technology. China remains able to obtain some restricted technology through workarounds, alternative suppliers, stockpiling and illegal diversion, while also increasing its domestic production capacity.

The cautious view above is supported by the experience of CoCom. Technology controls do slow the pace of development of a competitor, but they rarely prevent it from developing the technology entirely. In China’s case, the task is especially difficult because the country is deeply connected to the global economy and remains an important market for many of the governments and companies involved. The United States therefore needs the cooperation of allies such as the Netherlands, Japan, South Korea and Taiwan. Yet these countries do not face identical economic interests or possess the same legal powers. As a result, allied controls have not always been equally strict, and companies have had incentives to search for gaps in the system.

This leads to differences in the effects of the policy over short and longer horizons. On the one hand, in the short term, such controls boost the United States’ negotiating position because the United States and its partners control several choke points that Chinese firms will find very difficult to substitute: high-end lithography equipment, electronic design software and high-performance processors. Chinese companies cannot easily replace advanced lithography equipment, electronic design software or high-performance processors. In the long term, however, the same restrictions encourage China to invest more heavily in domestic substitutes. The Huawei example demonstrates this problem clearly: the bans limited Huawei’s access to Western technology, but they also encouraged the company to develop alternative chips and operating systems. According to one strand of literature, a restrictive regime leads to technological substitution rather than simple dependence.

The controls have indirect effects on trade beyond the direct US-China nexus. Firms must now spend more time checking the origin of components, applying for licences and ensuring that their products do not violate American, Dutch, Japanese or Chinese rules. This increases compliance costs and makes supply chains less efficient. Companies may also choose suppliers based on political reliability rather than price or quality. The consequence is a move away from one highly integrated global network towards a handful of regional or politically aligned networks.

On the other hand, these costs are not shared evenly: China has suffered the greatest loss of access to advanced technology, but companies in the United States and its allied countries have also lost customers and revenue. Dutch and Japanese equipment producers, in particular, face the difficulty of restricting sales to China while protecting their own industries. The United States has been in a stronger position because its leading design companies can often replace some lost Chinese business with demand from American cloud providers, governments and other international customers. This explains why cost asymmetry makes allied cooperation possible but not necessarily stable.

Again, this shows that economic interdependence can be used as leverage by both sides. The United States used its control over design software, AI chips and manufacturing technology to restrict Chinese access. China responded by using its influence over critical minerals and by encouraging domestic production. This is consistent with the theory of weaponised interdependence, which explains that global networks promote cooperation but that countries at the heart of these networks can convert economic dependence into political leverage. After one country exploits such a chokepoint, other countries have incentives to create alternative routes and chokepoints.

The WTO’s most intractable problem lies here. Article XXI must leave governments enough freedom to respond to genuine national-security threats. Advanced chips can contribute to military intelligence, weapons development and surveillance, so it would be unrealistic to treat every export-control measure as an ordinary commercial restriction. However, allowing states to decide on the applicability of Article XXI on a case-by-case basis makes this exception far too broad. Protectionist industrial policies could be easily cloaked as security measures, and little meaningful review would be possible.

This leaves the WTO with no easy way forward. It does not need to examine who is right or wrong about underlying security threats in disputes between the United States and China, but it can look into whether the challenged measure has at least some nexus to national security, was applied in a consistent manner and was no broader than necessary.

Conclusion

US-led allied export controls have significantly changed international trade in advanced technology since 2022. They have restricted China’s access to leading-edge chips and manufacturing tools, strengthened the bargaining power of the United States and its allies, and delayed parts of China’s advanced semiconductor development. Nevertheless, the controls have not completely stopped China’s technological progress. Instead, they have encouraged domestic substitution, increased circumvention and motivated China to use its own control over critical minerals as a response.

Rather than fully denying China access, these measures have only limited it. While these policies have, in the short term, yielded national security gains, they have promoted a more fragmented future for global trade in technology. Manufacturing has become costlier and more politically divided, while firms must now operate in an increasingly complex world of competing export controls and licensing regimes. The very interdependence that allowed the semiconductor industry to become efficient has, at the same time, created the chokepoints through which governments may coerce each other.

The WTO should accept that national security grounds can validly be invoked in restricting some advanced technologies – especially those with obvious military uses. However, Article XXI should not be interpreted as an absolute free pass from the reach of trade law. The WTO should build a regime of transparency that requires members to justify their security measures and examine whether they are related to a genuine security concern rather than just protecting domestic industries. This will not remove the tension between national security and free trade; it will prevent this tension from engulfing the whole of the rules-based trading system.

Ultimately, export controls will only slow a rival down; they will never guarantee that one side will remain ahead indefinitely in terms of technology. They could, in fact, encourage the formation of competing supply chains and ultimately increase the difficulty of international cooperation. There is, therefore, no best answer to the dilemma posed by unrestricted trade on the one hand and economic decoupling on the other: it lies in carefully targeted controls combined with close allied coordination, a clear WTO role and ongoing efforts to maintain openness in the broader multilateral trading system.

Allen, G.C. and Goldston, I. (2025) Understanding U.S. allies’ current legal authority to implement AI and semiconductor export controls (Washington, DC: Center for Strategic and International Studies).

ASML Holding N.V. (2026) Q2 2026 financial results (Veldhoven: ASML Holding N.V.).

Branstetter, L. (2024) Export Controls and U.S.-China Technology Competition in an Interdependent World. Brookings Institution [Working Paper]. <https://www.brookings.edu/wp-content/uploads/2024/05/20240528_ES_Sanctions_Branstetter_Final.pdf>

Bureau of Industry and Security (2022) ‘Implementation of additional export controls: Certain advanced computing and semiconductor manufacturing items; supercomputer and semiconductor end use; Entity List modification’, Federal Register, 87(198), pp. 62186–62215.

Farrell, H. and Newman, A.L. (2019) ‘Weaponized interdependence: How global economic networks shape state coercion’, International Security, 44(1), pp. 42–79.

Harithas, B. and Schumacher, A. (2024) Where the chips fall: U.S. export controls under the Biden administration from 2022 to 2024 (Washington, DC: Center for Strategic and International Studies).

Hrynkiv, O. and Lavrijssen, S. (2024) ‘Not trading with the enemy: The case of computer chips’, Journal of World Trade, 58(1), pp. 61–86.

International Energy Agency (2025) Rare earth elements. International Energy Agency [online]. <https://www.iea.org/reports/rare-earth-elements/executive-summary>

Juniewicz, I. (2026) Diversion and resale: estimating compute smuggling to China. Epoch AI [online]. <https://epoch.ai/publications/chip-smuggling>

Lind, J. and Mastanduno, M. (2025) ‘Hard then, harder now: CoCom’s lessons and the challenge of crafting effective export controls against China’, Texas National Security Review, 8(4).

NVIDIA Corporation (2026) NVIDIA announces financial results for first quarter fiscal 2027. Press release, 20 May. Santa Clara, CA: NVIDIA Corporation. <https://nvidianews.nvidia.com/news/nvidia-announces-financial-results-for-first-quarter-fiscal-2027>

Oxford Economics (2023) Implications of U.S. semiconductor export controls for China. Oxford Economics Research Briefing [online]. <https://www.oxfordeconomics.com/resource/implications-of-us-semiconductor-export-controls-for-china/>

Pinchis-Paulsen, M., Saggi, K. and Mavroidis, P.C. (2024) ‘The national security exception at the WTO: Should it just be a matter of when members can avail of it? What about how?’, World Trade Review, 23(3), pp. 271–295.

Semiconductor Industry Association (2016) Beyond borders: The global semiconductor value chain (Washington, DC: Nathan Associates Inc).

Sutter, K.M. (2025) U.S. Export Controls and China: Advanced Semiconductors. Congressional Research Service, R48642. <https://www.congress.gov/crs_external_products/R/PDF/R48642/R48642.5.pdf>

Svetlicinii, A. and Su, X.J. (2025) ‘The unsettled governance of the dual-use items under Article XXI(b)(ii) GATT: a new battleground for WTO security exceptions’, World Trade Review, 24(1), pp. 75–100.

Woods, D. (2025) ‘Escaping a weaponized network: China’s reaction to the United States escalating technology controls’, Asian Review of Political Economy, 4.

World Trade Organization (2012) Article XXI: Security Exceptions. GATT Analytical Index. <https://www.wto.org/english/res_e/booksp_e/gatt_ai_e/art21_e.pdf>

World Trade Organization (2022) United States — Measures on Certain Semiconductor and Other Products, and Related Services and Technologies (DS615). <https://www.wto.org/english/tratop_e/dispu_e/cases_e/ds615_e.htm>

* LLMs were used in some areas for skeleton essay structure (ideas) and grammar/spelling.