At the Frontier and Beyond: Competing Theories of Technological Power in U.S.-China AI Competition

by Emanuele Rossi

Two months after the release of DeepSeek-V4, the immediate shock has largely dissipated. Markets have moved on, benchmark comparisons have been absorbed, and the intense debate surrounding China’s latest AI models has faded from the headlines. The strategic questions raised by the episode, however, remain unresolved.

DeepSeek challenged more than assumptions about China’s technological capabilities. It exposed a deeper uncertainty about how AI leadership should be measured. For much of the past decade, assessments of U.S.-China competition have focused on a relatively narrow set of indicators: model performance, computing power, access to advanced semiconductors and research breakthroughs. Through most of these metrics, the United States continues to hold a substantial advantage. American firms dominate frontier model development, control critical segments of the AI ecosystem and benefit from unparalleled access to capital and talent.

Yet AI competition increasingly extends beyond the frontier itself. The deeper issue concerns the relationship between technological achievement and power. Breakthroughs matter, but their strategic value depends on whether they can be embedded within economies, institutions and societies in ways that generate lasting advantages. The central question is therefore not simply who develops the most capable models, but how technological capabilities translate into economic strength, industrial capacity, military effectiveness, and geopolitical influence.

Viewed through this lens, the United States and China appear to be pursuing different theories of technological power. Washington continues to prioritise leadership at the technological frontier, operating on the assumption that superior capabilities will generate broader strategic advantages. Beijing is pursuing a wider strategy centred on deployment, industrial integration and ecosystem expansion.

Two Different Paths to AI Leadership

The contrast between the American and Chinese approaches reflects deeper differences in how each country understands the relationship between innovation and power.

The United States remains heavily concentrated on frontier development. Public policy, private investment, and corporate strategy are largely aligned on maintaining leadership in advanced semiconductors, large-scale computing infrastructure, and cutting-edge AI models. This approach has produced remarkable results. The world’s most influential AI firms remain American. The leading frontier models are overwhelmingly developed in the United States. The country continues to attract a disproportionate share of global AI talent and investment.

Underlying this strategy is a straightforward assumption: technological leadership generates downstream advantages. Superior models strengthen military capabilities, attract investment, reinforce technological ecosystems and create opportunities for economic gains. Maintaining a lead at the frontier, therefore, becomes a strategic objective in its own right.

China places greater emphasis on a different stage of the technological cycle. While Beijing continues to invest heavily in research and development, its broader priority is deploying AI across the national economy. Industrial policy increasingly focuses on adoption in manufacturing, logistics, infrastructure, public services and emerging sectors such as robotics.

This distinction matters because technological revolutions rarely depend on invention alone. The largest gains often accrue to societies that succeed in embedding new technologies across productive sectors. Productivity growth emerges through adoption, organisational adaptation and industrial transformation as much as through scientific breakthroughs. The United States and China are therefore pursuing different pathways towards the same objective. Neither approach has yet demonstrated clear superiority.

Why the United States Still Holds the Advantage

Despite growing attention to China’s progress, the balance of capabilities remains tilted in Washington’s favour.

The strongest pillar of American leadership remains the semiconductor ecosystem. Advanced computing power underpins frontier AI development, and the United States occupies critical positions in chip design, software tools and computing architectures. Yet this advantage is often described too narrowly. American strength does not rest solely on domestic capabilities. It is embedded within a broader network of allied technological power.

Taiwan dominates advanced fabrication. The Netherlands remains indispensable in lithography. Japan and South Korea occupy critical positions across materials, equipment and memory technologies. Together, these actors form a  semiconductor coalition that continues to shape the global distribution of advanced computing capabilities. The effectiveness of U.S. export controls derives from this wider ecosystem rather than from American power alone. 

Capital constitutes a second structural advantage. The scale of investment flowing into the American AI ecosystem remains unmatched. Venture capital markets, institutional investors and technology firms have mobilised hundreds of billions of dollars for research, infrastructure and deployment. This financial depth provides a degree of flexibility that competitors struggle to replicate.

A third source of strength lies in the interaction between universities, research laboratories, technology companies and investors. The American innovation system continues to attract a disproportionate share of global talent and remains exceptionally effective at transforming scientific research into commercially viable technologies.

Taken together, these advantages explain why the United States remains ahead. China’s progress is real and increasingly significant. Yet it has not displaced the United States from the commanding heights of frontier AI.

China’s Alternative Strategy

China’s position becomes more interesting when viewed through a different lens. Much of the debate surrounding AI competition assumes that Beijing’s objective is to replicate the American model. In reality, China’s strategy increasingly focuses on areas where scale, coordination and industrial capacity can compensate for disadvantages at the frontier.

One element of this approach is diffusion. Chinese policymakers have consistently emphasised the integration of AI across productive sectors rather than limiting attention to leading laboratories. The objective is to embed AI within industrial processes, manufacturing systems and infrastructure networks on a national scale.

China’s manufacturing base offers particular opportunities in this regard. Unlike many advanced economies, China combines digital capabilities with an extensive industrial ecosystem. The convergence of artificial intelligence, advanced manufacturing and robotics could create advantages that are difficult to capture through conventional AI benchmarks.

The rise of Chinese open-source models represents another important development. While American firms continue to dominate the frontier, Chinese companies have demonstrated an ability to produce increasingly capable models at lower cost. These systems may not always match the performance of the most advanced American counterparts, but they are often more accessible and easier to deploy. This dynamic could prove influential across emerging markets, where affordability and adaptability frequently matter more than frontier performance.

Energy and critical minerals further reinforce China’s position. As AI systems become more energy-intensive, the ability to expand power generation and infrastructure will acquire growing strategic importance. China has invested heavily in both energy capacity and the supply chains that support advanced technologies. Its influence across critical mineral markets provides additional leverage in sectors that underpin both digital and energy transitions.

Collectively, these factors point to a strategy centred on scale, deployment and industrial integration rather than direct competition at every layer of the technological frontier.

The Competition Is Moving Beyond the Lab

The next phase of AI competition is likely to be shaped by factors that receive less attention than model releases or benchmark scores. Productivity growth, infrastructure development, energy availability and supply-chain resilience may prove just as important as advances in model capabilities.

This shift is already visible in policy debates. Governments increasingly view AI through the lens of industrial competitiveness, economic security and strategic resilience. Questions of data-centre construction, electricity generation and critical-mineral access have moved closer to the centre of national AI strategies.

The most revealing development in recent months may not have been the release of another Chinese model, but the growing willingness of U.S. policymakers to treat frontier AI systems as strategic assets. For much of the past decade, efforts to preserve technological advantage focused primarily on hardware: semiconductors, manufacturing equipment and compute infrastructure. Policymakers are now beginning to confront a different possibility: that frontier AI models themselves may require forms of control traditionally associated with strategically sensitive technologies.

As advanced systems gain greater relevance to cybersecurity, intelligence, and military applications, the distinction between commercial technology and national security capabilities becomes increasingly difficult to sustain. Competition with China may therefore push the United States towards a more interventionist approach to technology governance, one in which the state plays a more direct role in determining how strategically sensitive capabilities are developed, deployed and accessed.

Artificial intelligence is also becoming increasingly relevant to cybersecurity, intelligence analysis, military planning and information operations. As these applications expand, competition will extend further into domains where economic, technological and security considerations intersect.

From Frontier Leadership to Strategic Power

The conventional narrative of U.S.-China AI competition remains centred on frontier capabilities. That focus is understandable. The United States retains substantial advantages in semiconductors, capital, research and frontier model development.

Yet the competition is increasingly being shaped by factors that lie beyond the frontier itself. China’s strategy places greater emphasis on the mechanisms through which technological capabilities are translated into economic and industrial outcomes. Deployment, infrastructure, manufacturing integration and ecosystem expansion occupy a more prominent place in Chinese thinking than they often do in Western assessments.

The implications extend beyond the balance of power between the two countries. Competition with China is increasingly influencing how the United States itself approaches technology. Measures once focused on hardware and supply chains are gradually expanding towards models, algorithms and access to advanced capabilities. The underlying logic of technological competition is beginning to reshape the relationship between innovation and state power.

This may prove one of the most consequential effects of the AI race. Much of the debate has focused on whether China can narrow the technological gap with the United States. An equally important question is whether prolonged competition will change the way the United States governs strategic technologies. The boundary between commercial innovation and national security is becoming less distinct. In that sense, AI competition is not only redefining the balance of power between Washington and Beijing. It is also reshaping the institutions through which technological power is exercised.

*****

Emanuele Rossi is an international affairs specialist whose work centres on the Mediterranean and its strategic links to the wider world. He is Diplomatic Editor at Formiche and a senior analyst at Decode39, and he collaborates with a range of international think tanks and media outlets.

This article can be reproduced provided the author and the Politeia Research Foundation receive due acknowledgement, and it is not used for commercial purposes. The copyright will remain with the Politeia Research Foundation ©️PRF.