Beyond the Frontier

When China unveiled its large language model DeepSeek in January 2025—trained using less powerful GPUs and significantly less computing power than the leading American models—it generated anxiety in Washington foreign policy circles. Commentators compared it to the “Sputnik moment” of the early Cold War space race, when U.S. policymakers were shocked that the Soviet Union had taken a decisive technological lead in the contest for global power.

When analysts consider great power competition and the role technology plays in success or failure, they primarily think about the frontier—the race to develop leading-edge technologies that can confer economic or military advantage and be translated into strategic success. This is how the United States is approaching the development of artificial intelligence. It is engaged in a headlong rush to capture the leading edge of AI development before China, its only real peer competitor.

However, history tells us that focusing on the technological frontier alone is not enough. Recent work by Jeffrey Ding at George Washington University has emphasized that it is the long-term diffusion of new technology throughout society that is decisive in economic competition between leading states in moments of international power transition, where one great power supplants another in shaping the international system.

Previous Industrial Revolutions as Guidepost

As competition to master and shape the cluster of technologies associated with the Fourth Industrial Revolution proceeds apace, looking back to the previous three industrial revolutions is illustrative. How did they shape the international balance of power in their respective eras? Ding’s argument in Technology and the Rise of Great Powers is that, although scholars and policymakers have tended to equate great power dominance and hegemony with controlling the leading edge of innovation, that view pays insufficient attention to the more prosaic question of how new technologies are adopted and embedded in the productive process.

It is only through the more mundane process of technological diffusion (unfolding over many decades) that innovation makes a decisive difference in great power transitions. Although innovation may generate an initial burst of economic growth, the more far-reaching gains arrive later in the technology’s developmental cycle, as it becomes embedded in an interconnected technological system.

The states that ultimately emerged dominant from the first three industrial revolutions were not necessarily those that pioneered the new technologies, but those that most effectively disseminated them throughout their economies. Britain’s advantage over its European rivals during the first industrial revolution, the United States’ eclipse of Britain and Germany during the second, and America’s lead in the Information Age (the third) all depended less on invention than on successfully embedding new technologies in systems, new infrastructures, changes to firm’s organizational structures and workforce composition. If the United States and its allies are to make the most of the new technologies of the fourth industrial revolution, they also need to build the long-term infrastructures of diffusion to carry them throughout their societies.

The Competition to Build AI

The current competition to build AI systems is illustrative. A stock market boom driven by firms such as NVIDIA, Microsoft, and Alphabet reflects transformative expectations surrounding AI. Yet the first wave of actual transformation is working its way through selected working practices as early adopters integrate AI into coding, logistics, and industrial automation. The much longer and harder work of building the necessary societal and material infrastructure (including data centers, electricity generation, transmission networks, and computing capacity) is still to come. Only then can AI spread across the wider economy, as manufacturers automate production, logistics firms optimize supply chains, and public services embed AI into everyday administration. Such diffusion also needs to be accompanied by a broader reorganization of firms and labor markets, supported by skills infrastructure developed through education and training. AI should not be viewed as a standalone frontier technology, but as a complex ecosystem or stack.

A key ingredient  in the successful translation of technological mastery into international influence and power is the way in which new technologies are diffused throughout both national and international societies (the latter is perhaps a crucial and overlooked point). This requires both material infrastructure and an understanding of the systems within which clusters of technologies co-evolve and are deployed. Building such infrastructures is a long term, intergenerational project that unfolds over decades—as do its economic and geopolitical rewards. It is a project that must involve an alliance of other states, as resources, supply chains, and markets sit outside of the great power’s homeland.

Although the United States may be pushing furthest toward the technological frontier in many technologies and technological clusters (frontier artificial intelligence, advanced semiconductor design, quantum computing, aerospace, and biotechnology) it is in long-term planning and construction of infrastructures of diffusion where its current model struggles.

China’s Advantage

Infrastructures of diffusion are where China is pulling ahead, due to two distinct policy choices made decades ago and pursued with long-term patience and strategic stability. The first is its conscious choice to use the state as an engine of development and diffusion, subordinating the market to national goals (embodied in its five-year industrial plans). The second is its flagship multidimensional and transnational infrastructure undertaking, the Belt and Road Initiative (BRI), launched in 2013 and subsequently written into China’s constitution. The BRI acts as a hinge, linking the domestic and international infrastructures of technological diffusion needed to build future forms of political and economic order.

In his study of four decades of official Chinese documents, speeches, and memoranda, Rush Doshi, director of the China Strategy Initiative at the Council on Foreign Relations, found that China developed a remarkably consistent strategy following the death of Mao Zedong in 1976. He called this strategy the “long game,” a decades-spanning plan to first blunt and then displace American power.

At first, in the decades after Mao’s death, China grew its economy within the confines of the American-led liberal international order. President Xi Jinping’s rise to power in 2012 and the announcement a year later of the BRI marked a decisive shift in emphasis— a more confident and expansive direction within that long-term strategy. The BRI is the umbrella term for the immense set of transnational infrastructure projects designed to connect large swathes of Afro-Eurasia together—and to China’s center of gravity—through investments in ports, roads, railways, energy, digital connectivity, cities, and urban corridors. Although it has produced a mixed record of successes and failures, more than a decade after its launch, China remains committed to the initiative.

There are several notable features in the context of this discussion. The BRI is just that kind of long-term, intergenerational project designed to connect domestic industrial policy with foreign policy. The former diplomat and author Bruno Maçães has described it as the first transnational industrial policy. It builds the material exoskeleton upon which future forms of Chinese-led international order may eventually be overlain. The BRI harnesses the capabilities of a great power state to direct investment in infrastructure across transnational spatial scales and long-term temporal scales. It offers a vision of Chinese international leadership to potential allies and seeks to build the material foundations required to sustain it. In conjunction with successive five-year industrial plans, it also demonstrates that China is thinking strategically about how to secure the far-flung supply chains needed to power its technological development.

In the long-term horizon, its transnationalism and its use of the state to shape markets toward national objectives, China is demonstrating a steady and coherent strategic approach. And although China may not yet have captured the furthest edges of the frontier of technological development, the BRI can be understood as an ever-evolving and vast infrastructure of technological diffusion.

This is diffusion as a deliberate act of statecraft. Will it give China a decisive edge in the struggle for international leadership now underway? Currently, the United States lacks this kind of long-term geostrategic vision and planning. It also lacks institutional mechanisms that might allow it to muster its own vast resources of capital and talent and apply them to building the infrastructures of diffusion that could carry its own technological models and values.

To Own the Future, You Must Build It

The United States once knew the importance of infrastructures, but has now seemingly forgotten it. Its own domestic success was built on grand infrastructure projects: highways, railroads, cities, dams, and power plants. The first half of the twentieth century saw successive waves of state-backed infrastructure development that transformed the productive capacity of American society. Rural electrification, the Tennessee Valley Authority, the expansion of ports and rail networks, and later, the interstate highway system, created the substrates through which American industrial innovation diffused across the continent. Infrastructure was a strategic foundation of national power in this period. And the United States did not limit itself to domestic infrastructures. It rebuilt a devastated Europe in the wake of the World War II through the Marshall Plan. It also built the architecture of international order, including the current multilateral system. Its model of society and its vast capital investments went on to shape infrastructures and urban spaces around the world, carrying with them a liberal disposition.

But today, the United States has become lost in short-termism. Four-year political cycles, short-term investment cycles, and ephemeral media cycles have supplanted long-term vision. But technological and infrastructural systems are by their nature long-term undertakings. If the United States fails to lift its gaze toward the temporal horizon, its citizens may end up living in a future that China builds.

What can be done? Lessons can be learned from history, which tells us that it is not enough to race to the frontier or occupy the commanding heights of technological innovation. Lessons can also be gleaned from China’s approach, which effectively harnesses the power of the state for national and transnational projects.

To begin, democratic societies will need to show that they can match China’s ability to play the long game, developing a culture of strategic planning that thinks across decades and generations. This is currently a significant weakness when it comes to competing with a one-party regime whose leadership is unruffled by the regular political upheavals that can hamper democracies.

If infrastructure is one of the main ways in which decisions made in the present reach into the future to shape it, then this is the place to start building bipartisan consensus and long-term perspective. Once again, history is a guide. The United States once built its century of leadership and success on the material foundations it constructed in the mid-twentieth century. These are rare moments of world-building. China appears to be engaged in such a moment of world building with the BRI as its mechanism. If the United States is to partake in the leadership of the next cycle of international order, then it will need to muster all the resources at its command.

Success in the Cold War required an intergenerational strategic vision and collective effort. The containment of the Soviet Union was a long-term undertaking. China need not be contained in this way, but building the infrastructures of tomorrow is a similar intergenerational undertaking, one for which the United States currently seems ill-equipped. If China is playing the long game, then the United States is barely in the game. Its divided and partisan political environment needs mechanisms to unite it around long-term strategic thinking about the requirements of material infrastructure.

The United States will need to build its own twenty-first century infrastructures of technological diffusion in a purposeful way, both domestically and internationally. But private capital and free markets alone will not do it. The United States has vast resources of capital and talent and a culture centered on free thought and innovation. Yet these resources are running up against institutional blockages when it comes to infrastructure investment. A more active state and industrial policy are required. Just as the federal state was once instrumental in shaping the domestic and international environments that gave the United States eight decades of international leadership, so it will need once again to develop policy instruments that enable it to shape infrastructure effectively.

A Sovereign Infrastructure Institution as Catalyst

Proposals have been made for a sovereign infrastructure institution that can unlock structural blockages and unleash the vast productive energies and investment capacities of the United States that are now confined. There have been several proposals about the form that such an institution might take. A National Infrastructure Bank, a Sovereign Infrastructure Fund, or a Strategic Investment Fund have all been proposed.

Each option has its supporters and detractors. But the key point is that they can offer the missing institutional mechanism needed to unlock investments in long term infrastructure and technology projects—clean energy, data centers, grid upgrades—that markets might eschew for reasons of risk or return horizon. And they can allow the federal government to set strategic infrastructure objectives for the long term and at scale.

Leading in the race to the frontier needs to be accompanied by institutions that help firms adopt those advances. An estimated initial investment of a trillion dollars of transformative infrastructure might be quickly unlocked in this way—a figure that matches investments in the BRI over its first decade. This could build the transformative infrastructures of technological diffusion that the United States will need to get the most out of its current innovation wave, and to build connecting transnational infrastructures to its own technological systems.

China’s BRI should focus minds on the fact that successful strategic competition today must be transnational. If an enlargement of the temporal imagination is needed, so too is an enlargement of the spatial imagination. This will necessarily involve allies. The BRI is vast in its scope.  In some sense it spans more than two-thirds of the world’s nations, shaping infrastructures and urban spaces far from China’s borders, providing evolving conduits for China’s AI and digital offerings as they develop, and tying vast swathes of the world’s resources, supply chains, and peoples into a China-centred system. This evolving meshwork of infrastructural connectivity is what the United States is competing with.

Technological Diffusion as a Shared Endeavor

Under the previous administration, the United States had begun to think in such terms.  It placed the structural transition of U.S. energy systems at the core of its Inflation Reduction Act, thought hard about reindustrialization, and began to develop transnational infrastructures along the lines of the BRI. Two flagship projects to that end are the Lobito Corridor in Africa (led though the G7 Partnership for Global Infrastructure Investment) and the India-Middle East-Europe Economic Corridor. The administration recognized that transnational infrastructures are a key feature of international strategic competition and influence—and that the United States had no equivalent of the BRI.

It also recognized that the world was splintering into distinctive socio-technical systems, and that the United States needed to play this game more effectively. Doing so must transcend administrations and partisan politics. In other words, the long game.

If transnational scale is needed for success, then the United States will need a less transactional and extractive approach to overseas allies going forward. It will need to see them as part of its own socio-technical sphere of influence. In a world shaped by infrastructures of technological diffusion and widely dispersed global supply chains, the ability to think about where the transnational edges of your socio-technical and infrastructural system lie will be crucial.

No great power, no matter how technologically advanced, can go it alone. Trusted allies will also need to be engaged in ways that relate to infrastructures of technological diffusion. Their value needs to be calculated not simply in the older terms of “balance of power,” but in terms of a joint project aimed at creating shared markets, shared technological standards, and shared infrastructures—as a sphere of technological diffusion.

A central task, then, is to build a sphere of technological diffusion as a shared endeavor. These can be geopolitical and technological ecosystems that scale technologies and allow interoperability with trusted allies through shared infrastructures, standards, and investment networks. Here the United States already works closely with Japan, South Korea, and Taiwan when it comes to advanced manufacturing, semiconductors, and chip fabrication, the Netherlands for crucial semiconductor equipment, the UK in AI research, and Australia in critical minerals.

This is why it is vital to look beyond the technological frontier. As in past moments of geopolitical competition over the nature of international order, the key to success is not simply who invents the next generation of technologies, but who can build the widest sphere of technological diffusion: networks of allies, shared infrastructures, standards, and institutions through which innovations spread and become embedded in societies.