Geopolitics & Locational Competitiveness
Export strength, but dependency on advanced semiconductors and manufacturing equipment; export controls represent a geopolitical vulnerability.
Where does China stand, and where might it break?
China is the world's largest exporter of goods and dominates numerous supply chains. In advanced semiconductors, however, it remains import-dependent. Since October 2022, the United States has tightened export controls on high-end AI chips and manufacturing equipment on several occasions, most recently at the end of 2024. The measures bear primarily on EUV lithography — essential for cutting-edge chips and anchored in the Dutch company ASML (CSIS / Congressional Research Service). China's largest contract manufacturer, SMIC, has been producing chips at the 7-nanometre node without EUV since 2023, as demonstrated in Huawei's Mate 60 smartphone. Yields remain limited, however, and the step to even smaller geometries without modern equipment is difficult. The state has responded with substantial support: the third 'Big Fund' for semiconductor development was launched in May 2024 with approximately 344 billion yuan (roughly $47.5bn). The self-sufficiency target of 70 per cent set under 'Made in China 2025' has been clearly missed for advanced chips; domestic production covers only a fraction of demand depending on the segment. In return, China is leveraging its own market power: it has imposed export controls on critical raw materials such as gallium, germanium, and rare earths, whose processing it controls to the tune of around 90 per cent (industry reports).
AI value creation depends on computing power, and computing power depends on advanced chips. As long as China cannot manufacture these autonomously, its AI base remains vulnerable to export controls. A race is emerging between two forces: Western containment, which restricts access to leading-edge hardware and manufacturing equipment, and China's self-sufficiency policy, which drives the development of domestic chips, domestic equipment, and domestic models at any cost. If China achieves a breakthrough in lithography and manufacturing, it will attain technological sovereignty and emerge as an equal AI pole. If it does not, the computing base will remain its most vulnerable point, constraining the scaling of the AI economy. At the same time, mutual dependence over raw materials and markets is escalating into an instrument of geopolitics. The outcome of this race will shape the distribution of power in the economy ahead through to 2040.
Paths for China
China is pooling sovereign funds, talent, and domestic demand to build the entire supply chain at home — from design through lithography to manufacturing. This matters because only full sovereignty ends the vulnerability to export controls. The third Big Fund and SMIC's progress at the 7-nanometre node point in that direction. The cost is considerable: the gap in EUV lithography spans years and vast sums, initial yields will remain low, and a purely domestic build-up foregoes the efficiency of international division of labour.
Trade-off: Technological sovereignty versus cost and efficiency
Feasibility: Realistic only as a decade-long project requiring heavy capital deployment
China is focusing on the mass production of mature chips, an area in which it is already strong. It is deploying its dominance in rare earths, gallium, and germanium as a counterweight to Western controls. This is attractive because mature chips meet the bulk of industrial demand and raw-material power is a real source of leverage. The cost: mature chips do not serve the AI frontier. And deploying the raw-material weapon prompts countermeasures and diversification in the West that will erode China's leverage over time.
Trade-off: Short-term leverage versus long-term dependence at the frontier
Feasibility: Buys time, but does not replace the build-out of sovereignty
China is compensating for its hardware deficit through more efficient models, algorithms, and system architectures. The aim is comparable AI performance with less and older computing power. This matters because Chinese foundation models have recently closed much of the performance gap with leading Western models (Stanford AI Index 2026). The cost: efficiency gains cannot indefinitely overcome the physical constraints of older chips. And training the largest models continues to require cutting-edge hardware that remains subject to controls.
Trade-off: Near-term competitiveness versus long-term scalability
Feasibility: A valuable bridge, but no substitute for manufacturing sovereignty
Trigger-based phases
Trigger: Western export controls tighten, restricting access to leading-edge AI chips and EUV equipment.
China is running into hardware constraints when scaling its largest models. The state is launching multi-billion-dollar investment funds and deploying its own raw-material export controls as a counterweight.
Trigger: SMIC and domestic equipment makers approach advanced process nodes; yields rise gradually.
China is combining mature-node mass production, raw-material power, and model efficiency while working towards manufacturing sovereignty. Progress in lithography will ultimately determine the pace.
Trigger: China achieves sovereign production of advanced chips, or remains import-dependent for leading-edge AI hardware.
The outcome will determine whether China becomes a peer AI power or whether its compute base remains a critical strategic vulnerability. The result will shape the geopolitical distribution of power through to 2040.
Real-world pilot programmes
Status: Active (approximately 344bn yuan in semiconductor funding)
Launched in May 2024, the third national semiconductor investment fund is the largest to date, at approximately 344bn yuan (around $47.5bn). It pools capital for design, manufacturing and equipment. The test case: can the lag in advanced chips be closed through state capital?
Status: In production (limited yield without EUV)
China's largest contract chipmaker, SMIC, has been producing chips at the 7-nanometre node since 2023 without EUV lithography, as demonstrated in the Huawei Mate 60. The project illustrates how far China can progress with existing equipment — and where the limit lies without advanced lithography.
Status: In force (geopolitical counterweight)
China controls approximately 90 per cent of rare earth processing and has imposed export controls on gallium, germanium and other critical raw materials. These represent China's real leverage against Western chip controls — a test case for mutual dependence.
What we are watching
This indicator measures how independently China can produce its computing base. The threshold matters because only high-volume advanced manufacturing ends vulnerability to export controls. If reached, China becomes an equal AI pole; if not, hardware constraints limit the scaling of its AI economy.
This indicator identifies the most critical bottleneck in the supply chain. The threshold matters because lithography is the chokepoint of chip manufacturing and precisely where Western controls are targeted. A breakthrough would pierce the containment strategy; its absence sustains the dependency.
This indicator measures the resilience of China's geopolitical counterweight. The threshold matters because the raw materials lever only functions as long as the West has no viable alternatives. Should diversification succeed, China loses a central instrument of pressure in the technology conflict.