Three Ways to Manage US-China Tech Rivalry: Cooperation, Decoupling, and Targeted Controls

US policymakers debating China and critical technology are not choosing between engagement and isolation; they are choosing among three distinct strategic models—managed cooperation, broad decoupling, and targeted controls—each with different consequences for innovation, national security, and supply chain stability. Understanding what each model actually does, and what it sacrifices, is the prerequisite for evaluating any specific policy move, from tariff reductions to export controls on semiconductor manufacturing equipment.

Introduction: Navigating US-China Tech Engagement

The current moment contains elements of all three models at once. Reporting on the September economic and trade consultations describes the two economies moving toward structured cooperation and a reciprocal tariff-reduction framework covering roughly $30 billion in goods on each side, aimed at improving predictability for businesses, investors, and global supply chains. At the same time, the US maintains export controls on semiconductor manufacturing equipment, and analysis from the Carnegie Endowment argues that "national security" itself has been interpreted expansively—Congress's Export Control Reform Act of 2018 declares that US national security requires leadership in science, technology, engineering, and manufacturing, including foundational technology essential to innovation.

The Atlantic Council frames the underlying contest as a spectrum running from ordinary economic competition to geopolitical competition over strategic alignment. These models are not mutually exclusive in practice, but they pull in different directions, and their trade-offs deserve separate examination.

China's Drive for Technological Self-Sufficiency

Any US engagement model is a response to a deliberate, long-term Chinese strategy. An evaluation of Made in China 2025 by the US-China Economic and Security Review Commission concludes that Chinese industrial policy fostered what it calls the world's most advanced industrial commons—a collective resource base that Chinese firms can exploit to advance technological capabilities—with interlocking innovation flywheels spanning a dense network of interrelated technologies. On the Commission's reading, technical advances in one sector can rapidly catalyze breakthroughs in adjacent industries, and these spillover benefits increasingly advantage Chinese firms in commercializing future technologies.

The Commission identifies the most enduring legacy of the policy in sectors where advances come from "process innovation" on the factory floor, or where China has built large clusters of overlapping end-product manufacturers, component suppliers, and R&D capabilities. Yeling Tan's working paper for the Peterson Institute adds that China has adopted a growth model in which national security logic increasingly underpins economic policy, moving beyond traditional export- or investment-led growth. This context matters for US strategy: the target of American policy is not a static set of sectors but a self-reinforcing innovation ecosystem.

Model 1: Managed Cooperation

The cooperation model aims to strengthen global trade stability and predictability through structured economic engagement. Its tools are reciprocal tariff reductions and standing dialogue channels. Reporting on the September consultations notes that China and the US advanced work on the China-US Board of Trade alongside the tariff framework. The historical precedent is the Strategic Economic Dialogues and Strategic & Economic Dialogues, during which—according to an Asia Society Policy Institute analysis—China's current account surplus fell from 9.8% of GDP in 2007 to 1.7% in 2016, bringing it within IMF external metrics.

The model's innovation logic is that continued access to Chinese markets and collaborative exchange benefits both sides. Its supply chain logic is friction reduction: fewer protectionist disruptions, more predictable inputs. The risk profile is the mirror image. The same Asia Society analysis observes that China's technological capabilities, talent base, and supply chain strengths make it difficult and potentially self-defeating for US firms to disengage—meaning cooperation deepens interdependence that a competitor pursuing self-sufficiency may exploit. The Atlantic Council's framing of competition as a spectrum implies that economic engagement can shade into strategic alignment gains for Beijing, which is precisely what cooperation's critics worry about.

Model 2: Broad Decoupling

The decoupling model seeks to reduce US reliance on China across strategic technology areas. Writing for the Carnegie Endowment, Jon Bateman argues the primary effort should be "offensive": new investments and incentives to bolster and diversify innovation pathways, supply chains, talent pipelines, and revenue models in strategic technology areas. His reasoning is that the United States has far more influence over its own technological strength than over China's, and such investments act as a hedge against multiple scenarios—preparing America for full-scope decoupling at lower cost if necessary, or positioning US firms to compete in a still-globalized marketplace.

Decoupling in practice also involves defensive restrictions. Bateman notes that because the US maintains a significant leadership position in semiconductor design software and some segments of manufacturing tools, controls on entities like Huawei amount to a broad-based ban—though both the Trump and Biden administrations allowed Huawei to continue receiving billions of dollars of less-sensitive chips, with chips destined for Huawei's 5G systems presumptively denied. The trade-off is stark: broad decoupling could stimulate domestic innovation through investment and reduced Chinese competition, but it risks cutting US firms off from the Chinese market and raising costs across supply chains. Bateman's own conclusion is that there is no reliable way to know how the costs and benefits of interdependence will shift, so policymakers should "play for time" while strengthening options.

Model 3: Targeted Controls and Chokepoints

The targeted controls model does not attempt to disengage broadly; it identifies specific chokepoint technologies where China cannot easily substitute, and restricts those alone. The leading example, discussed by Saif Khan in a ChinaTalk interview, is DUV immersion lithography equipment for semiconductor manufacturing. Two US allies, Japan and the Netherlands, are the sole producers of these tools. Khan's assessment is blunt: "It is really the most effective chokepoint we still have available. China can't make it, and it is the hardest tool in the supply chain to indigenize." Based on historical data and the tools' cost and complexity, the underlying report argues China likely will not have a commercial indigenous tool until something like the mid-2030s—probably the last tool it will indigenize.

The model's appeal is precision: it impeded advanced chipmaking without banning all trade, and the administration's AI Action Plan identified closing loopholes in semiconductor manufacturing equipment controls as a priority, with no rollback of existing controls. The model's vulnerability is time. As one ChinaTalk commenter observed, the difficult question is how long export controls can preserve that advantage before they accelerate the effort to develop alternatives. Targeted controls buy a window, not a permanent moat.

Comparing US-China Tech Engagement Models

Three strategic models evaluated against a consistent set of criteria
Model	Primary Goal	Key Tools	Impact on US Innovation	Impact on Supply Chains	National Security Risk Profile
Managed Cooperation	Strengthen global trade stability and predictability through structured engagement	Reciprocal tariff reductions; dialogue channels such as the Strategic Economic Dialogues and the China-US Board of Trade	Potentially fosters innovation through continued market access and exchange, assuming mutual benefit	Improves predictability by reducing trade friction	Risks deepening interdependence that a competitor pursuing self-sufficiency may exploit, reducing US leverage
Broad Decoupling	Reduce US reliance on China across strategic technology areas	"Offensive" investments in innovation pathways, supply chains, and talent; broad restrictions such as the Huawei controls	Could stimulate domestic innovation through investment, but may limit access to global markets and talent	Diversifies supply chains but raises costs and disrupts established links	Reduces exposure to Chinese leverage, but Bateman cautions the future costs and benefits of interdependence cannot be reliably predicted
Targeted Controls	Impede China's advanced technological development at specific chokepoints	Export controls on narrow technologies, e.g., DUV immersion lithography tools produced only by Japan and the Netherlands	Preserves most commercial engagement while protecting chokepoint leverage	Minimal disruption outside the controlled technology	Effective but time-limited: China likely cannot indigenize these tools until the mid-2030s, but controls may accelerate indigenous efforts

Read horizontally, the table shows why no single model dominates: cooperation maximizes stability but concedes leverage; decoupling maximizes autonomy at the highest economic cost; targeted controls maximize leverage per unit of disruption but expire as indigenization proceeds. Read vertically, it clarifies which model aligns with which priority—maximizing innovation, ensuring supply chain resilience, or slowing a competitor's advance.

Evaluating Strategic Trade-offs: Innovation, Security, and Supply Chains

On innovation, the models diverge on a single question: does US innovation benefit more from access or from insulation? Cooperation's proponents point to market scale and exchange; decoupling's proponents, following Bateman, argue the US controls its own technological strength and should invest offensively regardless of what China does. Targeted controls sit between them, protecting specific capabilities while leaving most commercial activity intact.

On supply chains, cooperation reduces friction but deepens dependence on a state whose economic policy, per Tan's analysis, is increasingly driven by national security logic. Decoupling diversifies but is, in the Asia Society's phrase, potentially self-defeating for US firms given China's capabilities and market size. Targeted controls impose the least supply chain disruption—but only because their scope is deliberately narrow, and they do not address dependence outside the controlled node.

On security, the decisive variable is time. Khan's mid-2030s estimate for indigenous DUV immersion tools gives targeted controls a measurable horizon, while Bateman's framework treats uncertainty itself as the reason to hedge. The Atlantic Council's recommendation to use coercive tools sparingly—reserving them for genuine national security cases, as with the Huawei 5G campaign—reflects the same logic: coercion is a scarce resource whose overuse invites substitution.

Strategic Choices for US-China Tech Engagement

Policymakers must weigh the trade-offs between fostering global stability through cooperation, achieving self-reliance via broad decoupling, or strategically impeding competitor advancement through targeted controls. The evidence suggests these are not rival ideologies so much as instruments with different time horizons: cooperation manages the present relationship, decoupling hedges against long-run scenarios, and targeted controls exploit a window—on current estimates, until roughly the mid-2030s for DUV immersion lithography—that indigenization will eventually close. The observable signal will be the specific policy actions taken by the US government regarding export controls, investment restrictions, or renewed bilateral dialogues in critical technology sectors. Watch whether the September tariff framework expands, whether semiconductor equipment controls tighten or loosen, and whether the mid-2030s indigenization timeline holds—each is a measurable test of which model is actually governing US-China tech engagement.