20 May 2026, Geneva

Executive Summary

Taiwan has become one of the most consequential chokepoints in the global economy. Its semiconductor ecosystem, led by Taiwan Semiconductor Manufacturing Company, is not merely a supplier of commercial electronics. It is a strategic keystone supporting advanced logic manufacturing, AI packaging, cloud computing, defense systems, autonomous platforms, and the digital operating layer of modern states.

Taiwan accounts for more than 60% of global foundry revenue and more than 90% of leading-edge chip manufacturing. It also holds a substantial share of legacy node production, the older but still essential chips that run automotive systems, medical devices, aerospace electronics, and industrial infrastructure worldwide. This dependence is a strategic risk problem, not merely a question of market concentration. The world's most advanced digital capabilities, and much of its everyday functioning, rely on a production ecosystem located within one of the most actively contested geopolitical environments on earth.

Situation Overview

Recent acceleration in AI infrastructure investment has deepened global dependence on advanced semiconductors and on Taiwan's singular ability to produce them at scale. The more that governments and firms build national AI strategies, frontier models, and autonomous systems, the more they depend on chips whose production remains concentrated in a single, contested geography.

That concentration is not simply a function of market share. It reflects decades of accumulated process maturity, engineering knowledge, yield discipline, and a dense supplier ecosystem that cannot be quickly replicated elsewhere. Supply chains can sometimes be rerouted. Taiwan's advanced semiconductor role cannot.

The exposure runs deeper than the leading edge. Taiwan's strength in legacy nodes, chips manufactured at 28nm and above, means that a sustained blockade could severely disrupt automotive, medical device, aerospace, and industrial supply chains within weeks. These sectors depend on basic microcontrollers, not advanced AI accelerators, and they carry almost no inventory buffer against sustained disruption. Global automotive and medical suppliers operate on lean, just-in-time logistics with minimal safety stock downstream — meaning the weeks-to-impact timeline is structural, not speculative.

Efforts to diversify are underway. U.S. CHIPS Act-linked investments have reached nearly $450 billion across 25 states. But resilience is not the same as replacement, and the pace of AI-driven demand is outrunning the pace of diversification.

Why This Matters

AI dependence is deepening the chokepoint. Every national AI strategy, every frontier model deployment, every expansion of cloud infrastructure increases exposure to a contested geography already under sustained pressure. The chokepoint has already expanded beyond wafer fabrication into advanced packaging, where high-end AI accelerators are entirely dependent on specialized Taiwanese assembly capacity. Leading chips cannot reach deployment without advanced packaging processes such as CoWoS that remain concentrated in Taiwan regardless of where the underlying wafer was fabricated.

Geopolitical risk is directly embedded in the supply chain. A blockade, quarantine, cyber campaign, or even partial disruption of shipping, power, or logistics around Taiwan could produce cascading effects far beyond the immediate conflict zone. Full invasion is not required for severe global impact.

The chokepoint extends beyond Taiwan. ASML's EUV lithography systems, Japanese materials, and South Korean memory and high-bandwidth-memory capacity are all load-bearing nodes in the same network. Pressure on any of them degrades the whole.

A secondary chokepoint is forming in China. Beijing has spent the last several years heavily subsidizing trailing-edge semiconductor capacity through SMIC, Hua Hong, and affiliated fabs. In a crisis scenario, China could gain substantial coercive leverage over the legacy chips that run everyday global infrastructure, from hospital equipment to vehicle systems to grid controls, while Taiwan remains central to the leading edge. That is not a single chokepoint. It is a two-stage coercive structure.

What is often framed as an economic risk is, more precisely, a security risk. Disruption would degrade military readiness, AI development, intelligence systems, telecom infrastructure, and the digital sovereignty of democratic states.

Key Judgments

1. Taiwan is the world's most important semiconductor chokepoint, defined not by chip volume alone, but by dominance across both leading-edge manufacturing and legacy node production where alternatives remain limited.

2. Diversification will reduce risk, but not eliminate it. Advanced-node production maturity and yield reliability cannot be replicated on any near-term policy timeline.

3. AI growth is increasing, not decreasing, global exposure. The rapid expansion of AI infrastructure intensifies demand for the very chips most dependent on Taiwan's ecosystem.

4. The most plausible disruption may be below the threshold of war. A blockade, quarantine, cyberattack, logistics disruption, or coercive pressure campaign could produce severe consequences without full-scale invasion. This is the scenario current planning most consistently underweights.

5. China's trailing-edge buildup creates a parallel vulnerability. If Beijing achieves effective sufficiency in legacy node production, it gains coercive leverage over global everyday infrastructure independent of any action against Taiwan.

6. Successful diversification carries its own strategic tension. Taiwan's semiconductor centrality functions as a form of deterrence, the silicon shield, whether by design or structural consequence. The logic is that the world cannot afford to allow Taiwan to be seized or blockaded. If diversification proceeds far enough to reduce that dependency, Taiwan's geopolitical leverage diminishes alongside the risk.

Indicators to Watch

• Increased PLA activity simulating blockade or port denial around Taiwan.

• Cyber intrusions targeting Taiwanese power, ports, telecom, or semiconductor suppliers.

• Shifts in advanced chip inventory behavior, extended GPU lead times, or insurance premium spikes on Taiwan-linked shipping.

• Acceleration of Chinese trailing-edge fab capacity, new SMIC or Hua Hong production lines, or PRC export controls on legacy chips to third-party markets.

• Disinformation framing Taiwan's semiconductor sector as unstable, designed to erode market confidence ahead of or alongside coercive action.

• Acceleration of allied semiconductor subsidies or expanded public-private planning around chip security.

ISRS Assessment

Taiwan's semiconductor chokepoint combines geopolitical tension, supply-chain fragility, cyber vulnerability, industrial concentration, and AI acceleration into a single point of systemic exposure. The risk is not that the world depends on Taiwan for chips in a general sense. The risk is that the world has built much of its future economic, military, and technological trajectory around a highly specialized ecosystem located inside a contested security environment, and that the dependency runs from the most advanced AI accelerators down to the microcontrollers embedded in cardiac monitors and aircraft systems.

The emerging Chinese legacy node buildup adds a second dimension. A future crisis may not present as a single chokepoint but as a coordinated structure in which access to the leading edge and access to everyday infrastructure are controlled by adversarial parties simultaneously.

There is also a structural paradox that diversification policy cannot resolve cleanly. Taiwan's semiconductor centrality functions as a form of deterrence, whether by design or structural consequence. An adversary that seizes or blockades Taiwan disrupts the global economy along with it, and that consequence has long shaped how the risk is understood in strategic terms. But the logic cuts both ways: as allied diversification succeeds, that insurance weakens. There is no clean resolution to this tension, and any serious assessment of Taiwan semiconductor risk has to account for it.

An adversary does not need to destroy Taiwan's fabs to weaponize that dependence. Threatening market confidence and operational continuity may be sufficient to achieve significant strategic effect.

Conclusion

Taiwan's semiconductor chokepoint is not primarily a trade story or a technology story. It is a security story, one in which the infrastructure of the digital future has been concentrated, through decades of market logic, inside a geography that is now a primary theater of great-power competition. That concentration spans the full stack, from the most advanced AI chips to the trailing-edge components that keep hospitals, vehicles, and power grids running.

When a critical capability becomes sufficiently concentrated, it ceases to function only as an economic asset. It becomes leverage for those who hold it, and a vulnerability for those who depend on it. In Taiwan's case, that leverage has always cut both ways.

For democratic societies, the Taiwan semiconductor chokepoint is not only a supply-chain vulnerability. It is a test of whether open economies can preserve technological independence, institutional resilience, and strategic autonomy in the face of authoritarian pressure and coercion. The issue is not simply where chips are made. It is whether democratic states can preserve trusted access to the infrastructure on which their security, prosperity, and public trust now depend.

"The question is not whether Taiwan is a chokepoint. It is whether the world has built enough of its future there that the chokepoint has become the strategy."

— Dr. Dave Venable, Chairman, ISRS