The Decoupling Debate: Chinese Scholars Assess U.S. Chip Restrictions

U.S. chip restrictions have split the Chinese scholarly consensus on its tech strategy. One camp argues that American export controls function as a "forcing function" (倒逼) accelerating China's pursuit of autonomous AI capabilities. The other warns that retreating into autarky will sever China from the global innovation networks that built its long-term competitive edge.

This paper examines how Chinese scholars assess China's self-sufficiency prospects under semiconductor constraints—revealing fundamentally different prescriptions for how Beijing should navigate technological competition with Washington. Chinese perspectives on AI self-sufficiency—and consequently the strategic value of Gulf cooperation—vary significantly based on how scholars assess U.S. chip restrictions.

One camp, frames U.S. chip controls as a "forcing function" (倒逼 - dàobī) accelerating China's pursuit of autonomous capabilities across the AI value chain. He Defang et al argues in a piece with the Chinese Academy of Social Sciences that Western "small yard, high fence" (小院高墙) policies are accelerating "de-Sinicization" of global supply chains, with targeted constraints on computing power (算力) representing strategic chokepoints that reinforce "the need for China's integrated semiconductor modernization path."

Writing in Economic Daily, She Huimin portrays U.S. AI chip restrictions as an escalating strategy designed to constrain China to deliberately limited, "second-tier" technologies rather than to exclude it outright. She rejects any form of narrative of technological victimhood and calls external restrictions a "whetstone" (磨刀石) for China to actively sharpen its innovation capacity. In her framing, short-term pain from restrictions, higher costs, and transitional inefficiencies are acceptable—and even necessary—because U.S. pressure forces Chinese firms and the state to invest more deeply into domestic innovation and breaking China's technological dependence on U.S. technologies. In short, U.S. controls strengthening Beijing's resolve to pursue autonomous and controllable innovation pathways and provides the necessary push to accomplish it.

In their China Institute of International Studies commentary, Cai Cuihong and Yin Jiahui contend that the United States has transformed economic sanctions into a long-term, systematized strategy increasingly targeting semiconductors and other high-technology sectors. They argue that Washington is now cementing sanctions policies initially designed as temporary measures into permanent structural constraints. From their view, the U.S. adoption of technology as a permanent feature of competition means that China must also treat chips and artificial intelligence as strategic priority domains which require sustained, state-level interventions rather than conventional market adjustments to thrive.

Meanwhile, Zhao Minghao at Fudan University proposes a "geo-technology" framework, in which he argues that the United States has embedded semiconductors and emerging technologies like AI at the core of its strategic and security policy and uses its multilateral coordination of its allied network as a durable form of power to keep China away from its chips. Zhao's analysis moves beyond cataloguing specific policy tools to explain how chip export controls restructure global technology hierarchies and redefine access, standards, and dependence. His framework positions chips and AI beyond economic sectors vulnerable to sanctions. He portrays these as strategic security domains that great powers embed within their competition because they ultimately shape technological advantage for future military capabilities and the structure of international order.

For the "forced self-reliance" camp, the primary risk is technological dependency. The short-term costs of decoupling pale against long-term vulnerability to foreign chokepoints.

The open-but-secure camp focuses less on dependency and more on the consequences of isolation. They warn against China's over-correction into autarky because the greater danger lies in severing connections to global innovation networks that fueled China's technological rise.

Writing in China-US Focus, Sun Chenghao and He Wenxiang show the accelerating trend toward technological decoupling and observing that this shift transforms China-US relations from competition within a shared system to confrontation between rival systems. They note that comprehensive decoupling threatens to unravel the intricate tech economy entanglement that decades of economic globalization forged, fragmenting the foundations of multilateral technology governance and isolating countries from the frontier innovation ecosystems that long-standing global integration shaped. Sun and He warn that decoupling structurally weakens Chinese firms' cross-border financing, international collaboration capabilities, and access to global markets, pushing companies toward conservative strategies that scale back overseas ambitions. The shift targets not just individual companies but entire supply chains, value chains, and standards systems, fundamentally reshaping the global technological order.

Wu Xinbo from Fudan University treats technology competition as a bounded form of structural rivalry that policymakers must actively manage rather than allow to escalate into full decoupling. Wu argues that over-securitizing economic and technological ties risks isolating both countries from global innovation and production networks. Instead of autarky, he calls for clearly defined limits and continued selective engagement between both countries. He argues that indiscriminate disengagement imposes high economic and innovation costs on both sides without strengthening either country's strategic position.

The gap between these camps are fundamentally different assessments of China's technological position and timeline. The "forced self-reliance" camp see U.S. restrictions accelerating China's path to autonomous AI capabilities, rather than delaying it. The "open-but-secure" camp sees China still embedded in global innovation networks that took decades to build and cannot be rapidly replicated through state mobilization alone.

Both camps converge by placing China's domestic semiconductor capacity at the center. Self-reliance advocates argue China must build complete, autonomous supply chains because chips determine AI sovereignty. Openness advocates warn against decoupling precisely because losing access to frontier chip ecosystems threatens China's competitive position. Neither camp suggests that alternative partnerships, capital flows, or deployment markets can substitute for resolving the semiconductor constraint itself. A decade ago, semiconductors produced in western-aligned supply chains fueled China's technology book. Now, the shift to viewing western-developed semiconductors as the singular irreducible chokepoint reflects several converging developments.

First, U.S. restrictions on China's access to chips demonstrated that China's advances in AI, supercomputing, and telecommunications all depend on a semiconductor access, which could be turned off by U.S. regulations. No amount of innovation in other domains could offset that reliance short of Beijing's own insulated chip manufacturing capacity.

Second, China's own AI breakthroughs reinforced that computational infrastructure determines the limits of AI technology. Every Chinese frontier AI achievement like DeepSeek requires access to compute capacity fueled by advanced chips. Even the forced self-reliance camp's optimism about indigenous innovation assumes China can achieve semiconductor autonomy. It remains to be clear what that timeline looks like, but there is a growing belief that external pressure by U.S. competition will accelerate China's path to producing them domestically.

Third, both camps recognize that semiconductors represent the one domain where China faces disadvantages across multiple critical nodes, including manufacturing equipment (extreme ultraviolet lithography), design tools (EDA software), advanced materials, and process knowledge accumulated over decades. Other tech-adjacent sectors, like mineral processing, refining, and manufacturing, China has a clear lead. But, in advanced semiconductor manufacturing, China trails by what both camps acknowledge as at least a few years. The exact timeline is still a point of debate.

For U.S. policymakers, the consensus on China's semiconductor and AI autonomy matters because it shows the potential limits of chip restrictions as a strategy. Chinese scholars across the spectrum agree on the chokepoint and on the imperative to overcome it. They disagree on whether U.S. restrictions accelerate or delay China's path to self-sufficiency, but both camps treat semiconductor autonomy as non-negotiable. The question facing Washington is not whether chip restrictions can deter China's attempt at semiconductor self-sufficiency—Chinese strategic debates have moved past deterrence—but whether restrictions buy time or lose it.

This question is increasingly valid after Trump's recent decision to allow China's access to some H200 AI chips. More specifically, Trump's December 2025 announcement changed the equation slightly by allowing Nvidia to ship H200s to approved customers in China for 25 percent premium paid back to the U.S. government. This positions Chinese customers to choose between accessing U.S. hardware and inadvertently funding the U.S. government or refraining from U.S. chips and continue relying on Chinese hardware. The move sharpens the domestic debate by exposing how easily AI hardware access can become a bargaining chip. Xi's response, however, slowed the Chinese private sector's rush for chips. First, Chinese customs guidance, following Trump's announcement, reportedly told customs agents the H200s are not permitted to enter China. Second, Chinese authorities are finalizing a licensing regime for H200 chips which controls their inflow into China so that U.S. hardware does not become a dependency and undermine China's own domestic innovation push.

For the forced self-reliance camp, Trump's actions are likely to reinforce the view that semiconductor autonomy for China is non-negotiable, and that any exceptions postpone the ultimate goal of Beijing building its own AI stack. For the open-but-secure camp, the volatility is a stark warning that preserving selective access to global innovation networks is getting harder precisely because the network is being weaponized. However, Xi's willingness to consider allowing even controlled purchases of H200s suggests that there is a sufficient gap in Chinese compute capacity, especially with the big firms. This means China still has a degree of dependency on U.S. innovation, which Trump aims to deepen.

Photo Credit: Reuters, Evelyn Hockstein.