The phrase “technology decoupling” entered mainstream policy vocabulary somewhere around 2018, when the United States placed ZTE on the brink of corporate death with a single export restriction. But the rupture between the American and Chinese technology ecosystems did not begin with ZTE, and it will not end with whatever executive order or tariff announcement arrives next. It is a structural reorganization a decade in the making, still incomplete, and with consequences that extend well beyond the two countries at its center.
Understanding how it happened requires going back further than most coverage does. The earliest institutional signals appeared in 2015, when China released Made in China 2025 — a state industrial roadmap targeting global leadership in semiconductors, robotics, aerospace, and nine other advanced technology sectors. The document was explicit about its ambitions and explicit about what it saw as China’s vulnerability: dependence on foreign technology, particularly in chips. Washington read it, registered concern, and largely moved on. The Obama administration’s final year produced incremental Entity List additions and tighter satellite export rules, but nothing that suggested a fundamental rethinking of the technology relationship.
The rethinking came fast. Between 2017 and 2019, the Trump administration assembled the legal and regulatory infrastructure that would define the next decade. The Export Control Reform Act gave Commerce new authority over emerging and foundational technologies. FIRRMA transformed CFIUS from a deal-screening body into a strategic investment gatekeeper. The Section 301 tariffs applied blunt economic pressure. And the Huawei Entity List designation — announced in May 2019 — demonstrated that the United States was prepared to use export controls not as a compliance tool but as a weapon capable of severing a company from its supply chain entirely.
Huawei was the hinge event. The company was, at the time, the world’s largest telecommunications equipment manufacturer and a serious competitor to Apple and Samsung in premium smartphones. Its HiSilicon division was designing chips that rivaled Qualcomm’s best. When the Commerce Department cut it off from American components, software, and — critically, through the foreign direct product rule — from any chip manufactured anywhere in the world using American equipment or software, the message was unambiguous. No Chinese technology company was too large, too global, or too integrated into Western supply chains to be beyond reach.
The foreign direct product rule is worth understanding in some detail because it is the mechanism that gives American export controls their extraterritorial force. Any semiconductor manufactured using US-origin equipment or designed using US-origin software tools is treated as a US product for export control purposes, regardless of where it was physically made. Since American companies — Applied Materials, Lam Research, KLA, Cadence, Synopsys — dominate the equipment and software layers of chip manufacturing, the rule effectively means that the United States controls who can buy advanced chips regardless of whether the transaction involves an American company. When TSMC announced it would stop taking Huawei orders in May 2020, it was not making a voluntary commercial decision. It was complying with US law.
The Biden administration retained every major Trump-era restriction and added systematically to them. The CHIPS and Science Act of 2022 was not primarily a subsidy program, though it contained $52 billion in manufacturing incentives. Its strategic function was to embed decoupling into industrial policy through a guardrails provision barring recipients from expanding advanced chip manufacturing in China for a decade. Companies that accepted US government money to build fabs were legally required to choose sides.
Then came October 7, 2022. The Commerce Department’s comprehensive semiconductor export control rule was the most sweeping technology restriction since Cold War-era controls, and it went further in one critical respect: it targeted people, not just products. The US persons rule required American citizens and green card holders working at Chinese chip companies to choose between their jobs and their legal status. Within weeks, American engineers and executives were resigning from Chinese semiconductor firms. The rule acknowledged something that prior export controls had not: that the knowledge required to build advanced chips is embedded in the workforce, not just the equipment, and that workforce mobility is itself a technology transfer vector.
The multilateral dimension of the October 2022 rule was not announced but quickly became visible. The Netherlands restricted ASML’s advanced lithography exports within months. Japan followed with controls on 23 categories of semiconductor manufacturing equipment in early 2023. The coordination was the product of intensive US diplomatic pressure, and it closed the gap that unilateral American action had always left open: if only the United States controlled exports, Chinese companies could source from elsewhere. With ASML and Tokyo Electron aligned, there was no elsewhere. The three countries together supply essentially all of the advanced equipment required to manufacture leading-edge chips. China was effectively excluded from the advanced semiconductor supply chain.
China’s response has been multidimensional. On critical minerals, it has wielded supply dominance methodically: gallium and germanium export controls in 2023, graphite restrictions later that year, rare earth expansions in 2025. China produces the majority of the world’s supply of many materials that are irreplaceable in defense electronics, clean energy systems, and advanced manufacturing. The controls have not yet produced acute shortages in Western supply chains, but they have injected strategic uncertainty and accelerated diversification investments that will take years to mature.
On chips, China has pursued self-sufficiency through a combination of massive state investment and engineering ingenuity. The Huawei Mate 60 Pro, launched in August 2023, carried a 7nm processor manufactured by SMIC using deep ultraviolet lithography equipment — tools that were supposed to be incapable of producing chips at that geometry. The technique involved multiple exposure passes to simulate the single-pass precision of extreme ultraviolet tools China cannot obtain. It was slower, more expensive, and lower-yield than what TSMC produces. It was also not supposed to be possible.
The Mate 60 episode was a warning. The semiconductor containment strategy rests on the assumption that leading-edge manufacturing requires tools that China cannot access. That assumption is eroding, not because China has solved the hard problem of advanced manufacturing, but because it has found partial workarounds that, combined with years of incremental progress and effectively unlimited state capital, are compressing the gap faster than the policy framework anticipated. CXMT’s DDR5 memory production milestone in early 2026, achieved using non-American equipment, suggests the trajectory will continue.
The current state of play is a technology relationship that is neither coupled nor decoupled but stratified. In advanced logic semiconductors — the chips that power AI training, data centers, and high-end devices — China is excluded from leading-edge production and, through export controls on Nvidia and equivalents, from easily purchasing the most capable chips. In mature-node chips — the workhorses of automotive, industrial, and consumer electronics — China is now the world’s dominant producer, subsidized to price competitors out of the market. In critical materials, it retains leverage it has barely begun to exercise. In software, AI, and cloud infrastructure, two distinct ecosystems are consolidating around different standards, platforms, and regulatory requirements.
For the global technology industry, this stratification creates a compliance architecture of extraordinary complexity. Multinationals must manage product lines, workforce compositions, data flows, and investment structures against the overlapping requirements of US export controls, CFIUS investment screening, the outbound investment rule, Chinese data security and counter-sanctions laws, and the connected vehicle and cloud compute restrictions that extend controls into new categories. The cost of compliance is significant. The cost of getting it wrong — as ZTE, Huawei, and the 140 companies added to the Entity List in December 2024 have discovered — is existential.
The decade ahead will be defined by which side of this divide advances faster: American efforts to maintain a technology lead through controls, investment, and alliance coordination, or Chinese efforts to close the gap through state-directed investment, engineering adaptation, and the leverage of critical mineral and mature-node supply dominance. Neither outcome is assured. What is certain is that the integrated global technology economy that characterized the first two decades of this century is not coming back. The integrated global tech economy is gone, and here’s where you can read exactly how it came apart.
Leave a Reply