Geopolitics & Locational Competitiveness
Leadership in AI, compute, and semiconductor design; sovereignty over key technologies largely secured.
Where does the USA stand, and where does it break?
Through companies such as Nvidia, the major cloud providers and the leading AI laboratories, the United States commands a degree of sovereignty over AI hardware and frontier models that other nations are striving to match (Stanford AI Index). With the CHIPS and Science Act of 2022, the federal government committed approximately $52.7bn for semiconductor manufacturing and research to bring production back onshore. The US dominates chip design, electronic design automation software and foundation models. The manufacture of leading-edge chips takes place overwhelmingly in Taiwan at TSMC. This strategic dependence could, in the event of conflict, affect the entire AI value chain. In response, TSMC is building several facilities in Arizona with up to $6.6bn in CHIPS Act funding, with initial production already under way. At the same time, the United States is deploying export controls on leading-edge chips as a deliberate geopolitical instrument to restrict China's access to AI computing power (BIS). The competition with China for AI leadership has elevated compute, chips and talent to the status of national priorities. The US leads in frontier models and commercialisation, while China is closing the gap in patent volumes and manufacturing scale.
In the economy ahead, control over compute, chips and foundation models is tantamount to control over the new value creation. The US is heavily dependent on Taiwan for the most advanced chip manufacturing. A conflict in the Taiwan Strait would immediately jeopardise the supply of cutting-edge chips and, with it, AI leadership. Technological strength rests on a supply chain that is both geographically concentrated and geopolitically exposed. Domestic political vulnerability compounds this. Deep polarisation, shifting foreign and trade policy, and a federalism that allows states to drift apart all reduce the predictability on which allies and investors depend. Export controls are an effective instrument, but they provoke countermeasures and accelerate China's drive to develop its own chip capacity. Whether the US can secure its lead depends on how quickly it reshores manufacturing or shifts it to reliable partners, and on how consistent its policy remains.
Paths for the USA
Via the CHIPS Act and subsequent programmes, the US is reshoring the manufacture of advanced chips and shifting production to trusted partners, with the aim of reducing dependence on Taiwan. This strikes at the heart of the fault line, since fabrication is the most vulnerable point in the value chain. TSMC Arizona, Intel and Samsung are already building capacity. The price is steep: fabrication plants cost tens of billions of dollars, take years to complete and require specialist workers — and the most advanced manufacturing remains concentrated in Taiwan for the time being.
Trade-off: security of supply versus cost and construction pace
Feasibility: Results will only become apparent over a full decade
The US maintains its lead through export controls on advanced chips and manufacturing equipment, coordinating these with allies including the Netherlands, Japan, and Taiwan. This is significant because it delays China's access to frontier compute and anchors Western standards. The downside: controls provoke countermeasures, drive China towards indigenous development, and weigh on the revenues of US chipmakers. Their effectiveness depends entirely on alliance cohesion.
Trade-off: technological lead versus trade and revenue risks
Feasibility: Only enforceable with a united alliance
The United States is securing its leadership by attracting the world's foremost AI talent through immigration and elite universities, while expanding public research investment. This matters because top-tier human expertise is the scarcest resource in the AI economy, and the United States has traditionally led in this regard. The capital-talent cycle of Silicon Valley represents a distinctive competitive advantage. The risk lies in restrictive immigration policy, which could impede the inflow of talent, and in brain drain should other countries become more attractive.
Trade-off: innovation leadership versus the immigration policy debate
Feasibility: Restrictive immigration would undermine this advantage
The United States is improving the reliability of its foreign, trade and technology policy in order to bind allies and investors more closely. This addresses the domestic political vulnerability that, alongside dependence on Taiwan, represents the most significant weakness. Predictable alliances and trade rules reduce costs for all parties involved. The structural constraint lies in deep polarisation and federalism, both of which make consistent, cross-party, long-term commitments inherently difficult.
Trade-off: alliance stability versus domestic political reality
Feasibility: Difficult to steer, as it hinges on domestic politics
Trigger-based phases
Trigger: The AI competition with China makes compute, chips, and talent a national priority; export controls become an instrument of policy.
The United States is consolidating its lead through models, computing power and supply-chain control. At the same time, dependence on Taiwanese fabrication is recognised as a strategic risk, and the build-out of reshoring begins.
Trigger: New fabs in Arizona and at partner locations come on stream, whilst export controls and countermeasures reshape global supply chains.
Manufacturing diversifies gradually, but remains costly and slow. The cohesion of the technology alliance and the political predictability of the United States will determine how stable the new order proves to be.
Trigger: It becomes clear whether the US has reduced its manufacturing dependency and secured its leadership, or whether geopolitical shocks have left it exposed.
If reshoring and alliance management succeed, the US will remain the centre of global compute. If dependency on Taiwan stays high and policy remains unpredictable, that leadership will remain vulnerable. The course set during these years will shape the country's position through to 2040.
Real-world pilot programmes
Status: Under implementation (approximately $52.7bn for semiconductor manufacturing and research)
The CHIPS Act is the central US industrial programme to bring chip manufacturing back to the United States and reduce dependence on Asia. It is the test case for whether state industrial policy can durably secure the most vulnerable point in the AI value chain — fabrication.
Status: Under implementation (initial production under way, further facilities planned)
TSMC is building several plants in Arizona with up to $6.6bn in CHIPS Act funding, manufacturing cutting-edge chips at scale in the United States for the first time. The project is the concrete test of whether dependence on Taiwan can be meaningfully reduced.
Status: In force (continuously updated restrictions on advanced compute exports to China)
The Bureau of Industry and Security restricts the export of state-of-the-art AI chips and manufacturing equipment. As a geopolitical instrument, it illustrates how the US seeks to protect its compute lead — whilst also revealing the limits of such controls and the countermeasures they provoke.
What we are watching
This measures the degree to which critical manufacturing remains geographically concentrated. The indicator matters because sustained concentration perpetuates the geopolitical vulnerability of the AI value chain. If it does not shift, leadership remains exposed. For strategic security, diversification is what counts.
This lead measures whether the US holds a commanding position at the source of AI value creation. The threshold matters because any loss of leadership in models and compute undermines the country's broader competitive advantage. Should competitors close the gap, the balance of power shifts. For positioning through to 2040, this indicator is central.
This measures how united the Western alliance is in applying export controls. The indicator matters because the effectiveness of those controls depends entirely on that unity. If the alliance frays, the controls lose their force. For geopolitical strategy, partner cohesion is what counts.