Artificial Intelligence, Sovereignty, and the New Geopolitics of Power

-An examination of how artificial intelligence is reshaping the global balance of power by turning models, chips, data, and platforms into strategic assets contested by states, corporations, and emerging AI blocs.

Every fracture in institutional life eventually reaches the border. Earlier essays in this series explored how artificial intelligence transforms labor markets, corporate structures, governance frameworks, and the human mind. Yet each of these transformations unfolds within a geopolitical landscape where nations compete, cooperate, and negotiate around the technologies that define economic and military advantage. The interface is not merely a technological platform. It is a geopolitical resource.

Control over advanced models, computational infrastructure, and the data ecosystems that feed them increasingly resembles control over strategic assets such as energy reserves, trade routes, or rare minerals. Ian Bremmer calls this a “technopolar” order, one in which a handful of technology companies rival states for geopolitical influence, exercising a form of sovereignty over digital space that previous centuries reserved for nations alone. Artificial intelligence is becoming embedded within the architecture of international power, and the companies building it are not passive instruments of state policy but active participants in the reordering of that architecture.

AI as Strategic Infrastructure

Throughout history, certain technologies have acquired strategic significance beyond their immediate economic value. The steam engine enabled naval dominance. Telegraph networks accelerated military coordination. Nuclear technology redefined deterrence and diplomacy. Artificial intelligence belongs to this category of strategic infrastructure, but it arrives with a distinctive feature: the private sector, not the state, controls most of the critical capability.

Training frontier AI models requires enormous computational resources, advanced semiconductor manufacturing, and large-scale data processing capabilities. These components form a supply chain that is geographically concentrated. Semiconductor fabrication depends heavily on a handful of companies and regions. Advanced chips rely on specialized equipment produced by even fewer manufacturers. Cloud computing infrastructure requires billions of dollars in capital investment. Henry Kissinger, Eric Schmidt, and Daniel Huttenlocher go further: AI represents a transformation comparable to the Enlightenment in its implications for how states understand and exercise power.

Because these resources are unevenly distributed across the world, the capacity to build and deploy advanced AI systems becomes a marker of geopolitical capability. Nations increasingly treat AI infrastructure as a strategic asset worthy of industrial policy, export controls, and national security review. Yet Mariana Mazzucato’s research complicates any clean division between state and corporate responsibility. Much of the foundational work underlying today’s AI industry was publicly funded. The relationship between government investment and private-sector innovation is symbiotic rather than adversarial, and the nations that recognize this are best positioned to lead.

The Emerging AI Blocs

As AI development accelerates, distinct geopolitical groupings appear to be forming around different regulatory philosophies and technological ecosystems. The landscape is more complex than a simple tripartite division, but three primary models have emerged.

The United States continues to rely on a model in which private-sector innovation, driven by venture capital and large technology firms, leads development. Government involvement focuses on research funding, export controls on advanced semiconductors, and national security oversight. The July 2025 White House AI Action Plan framed the export of the American AI technology stack as a mechanism for sustaining technological leadership. The pattern is older than AI. Government funds the foundational research. Corporations drive commercialization and deployment. The partnership works until the question of who governs what the partnership produces can no longer be deferred.

The European Union has pursued a regulatory model emphasizing privacy, transparency, and risk classification. The AI Act, which entered into force in August 2024 with provisions phasing in through 2027, represents the most comprehensive attempt to govern AI deployment through law. It classifies AI systems into four tiers of risk: unacceptable, high, limited, and minimal. Each tier carries distinct obligations for both the companies that build the systems and the organizations that deploy them. The Act’s extraterritorial provisions mean its influence extends well beyond EU borders.

China has adopted a model that blends state direction with corporate execution. Major technology companies operate within strategic frameworks established by national industrial policy, and AI development is closely aligned with state objectives ranging from economic modernization to social governance. Kai-Fu Lee sees China’s position differently than many Western analysts do: its advantages in data scale, entrepreneurial intensity, and government mobilization make it a formidable competitor in applied AI, even where it trails in foundational research.

Yet the tripartite framing captures only part of the picture. A growing constellation of “swing states” is shaping the AI landscape in ways that resist simple alignment with any single bloc. India, the United Arab Emirates, Israel, Japan, South Korea, Singapore, and Brazil are each pursuing sovereign AI strategies calibrated to their own strategic positions. India seeks technological partnerships with Western and non-Western powers simultaneously, refusing the binary. The Gulf states are investing heavily in AI infrastructure as part of broader economic diversification. Brazil launched its national AI strategy in 2024, framing AI development as a matter of sovereignty in terms its twentieth-century industrialists would have recognized. The Global South, far from being a passive recipient of technology developed elsewhere, is emerging as an arena of active strategic negotiation.

The resulting landscape may resemble the early internet, when different regions developed distinct regulatory and technological ecosystems before partial convergence occurred. Hegel would recognize the pattern. Competing models do not simply coexist; they provoke and shape one another. The American emphasis on speed and scale forces Europe to ask whether regulation inhibits or enables trust. The European emphasis on rights and transparency forces American firms to consider whether self-governance is sustainable. China’s integration of AI into state capacity forces both to reconsider the relationship between public authority and private innovation. Yet unlike a Hegelian dialectic that resolves into synthesis, these competing models may produce not convergence but a landscape of sustained and productive tension. Whether that tension remains productive depends on choices that governments and corporations are making now.

Compute and the New Arms Race

The geopolitical competition around AI increasingly centers on computational capacity. Training large models requires enormous amounts of processing power, and the chips capable of delivering that power are produced through highly specialized supply chains.

This reality has already shaped international policy. Export controls on advanced semiconductor technology aim to slow the diffusion of cutting-edge computational capability to rival states. New alliances such as the Chip 4 grouping (the United States, Japan, Taiwan, and South Korea) have formed specifically to coordinate semiconductor strategy. At the same time, governments are investing heavily in domestic semiconductor production to reduce reliance on foreign manufacturing. The Stargate initiative, announced in early 2025, envisioned $500 billion in AI infrastructure investment over five years.

The dynamic resembles earlier strategic competitions over industrial capacity. Just as steel production and oil supply once served as indicators of national strength, computational capacity may become a central metric of technological sovereignty. Thucydides observed that it was the rise of Athens and the fear this inspired in Sparta that made war inevitable. Graham Allison’s modern revival of the concept as the “Thucydides Trap” frames the structural instability that arises when a rising power threatens to displace an established one. The US-China competition over AI infrastructure carries this structural tension: not because conflict is inevitable, but because the perception of shifting capability can drive preemptive action, export controls, and alliance formation even in the absence of hostile intent.

Yet the analogy has limits. Unlike natural resources, knowledge spreads through networks. Even when hardware is restricted, ideas and algorithms continue to circulate through academic collaboration and open research communities. Acemoglu’s caution is worth holding here: the economic benefits of AI-driven growth may not distribute as broadly as proponents assume. Nations must consider not only whether they can build the technology but whether its benefits will reach their populations in ways that sustain social cohesion.

The Multipolar Innovation Landscape

Despite geopolitical competition, the development of artificial intelligence remains deeply international. Researchers collaborate across borders. Open-source models circulate freely. Scientific publications remain accessible to global audiences.

This creates a paradox that no export control can resolve. The same week a government restricts chip shipments, its universities publish research that advances the field for everyone. Nations restrict hardware while encouraging academic partnerships. Companies compete fiercely while sharing findings through conferences and preprint servers. The knowledge moves even when the silicon does not.

The result is a multipolar innovation landscape rather than a simple technological arms race. Progress emerges through a combination of rivalry and collaboration, and the technology companies driving much of this innovation operate as bridges between national ecosystems as much as they serve as instruments of any single government’s strategy.

The Cold War offers a partial precedent. Scientific advancement continued across bloc lines even at the height of strategic rivalry. Soviet and American physicists read each other’s papers. International conferences persisted. The present landscape, for all its tensions, retains more channels of cooperation than the Cold War analogy typically suggests. That may be its saving feature.

Governance Without a Global Regulator

Unlike nuclear technology or global finance, artificial intelligence currently lacks a comprehensive international governance framework. There is no equivalent of the International Atomic Energy Agency overseeing model development, nor a global treaty establishing enforceable limits on AI deployment.

Instead, governance emerges through overlapping national regulations, corporate policies, and voluntary international initiatives. The OECD AI Principles, first adopted in 2019 and updated in 2024, now count 47 adherents and have been incorporated into the legislative frameworks of the European Union, the United States, and other jurisdictions. The United Nations has convened advisory bodies and launched the Global Dialogue on AI Governance. UNESCO’s Recommendation on the Ethics of Artificial Intelligence represents the first global standard on AI ethics. Yet enforcement remains thin.

This fragmented governance environment reflects the difficulty of regulating a technology that evolves rapidly and whose applications span nearly every sector of the economy. But the absence of a single regulatory authority does not necessarily imply chaos. Elinor Ostrom’s Nobel Prize-winning research demonstrated that polycentric governance systems, in which multiple overlapping institutions share responsibility for managing shared resources, can prove more adaptive and resilient than centralized authorities. The emerging AI governance landscape may follow this pattern: messy, distributed, and partially redundant, but capable of evolving alongside the technology it seeks to govern.

The Security Dimension

Artificial intelligence also carries direct implications for national security. Autonomous systems, cyber defense tools, intelligence analysis platforms, and strategic simulations all rely increasingly on machine learning.

Military planners recognize that AI may influence future conflict in ways comparable to earlier revolutions in communications or precision weaponry. Cyber defense systems already use machine learning to detect and respond to network intrusions at speeds no human team can match, and intelligence agencies process satellite imagery and signals data through models that compress weeks of analysis into hours. Yet the broader strategic implications remain uncertain. Some analysts warn of destabilizing arms races in autonomous weapons systems, while others argue that AI may enhance defensive capabilities and reduce human error.

International dialogue on military AI governance has begun but remains tentative. The challenge is balancing transparency and restraint without revealing sensitive capabilities.

The geopolitics of AI therefore unfold simultaneously in commercial markets, academic research networks, and national security institutions. The technology companies that build these systems occupy all three spaces at once, a reality that makes their role in governance discussions not merely useful but essential.

Digital Sovereignty

A growing number of governments frame AI development in terms of digital sovereignty: a nation’s ability to control the digital infrastructure and data ecosystems on which its economy depends.

Countries lacking domestic AI capacity may become dependent on foreign platforms for critical services ranging from language models to industrial automation tools. This dependency raises concerns about economic leverage, data privacy, and cultural influence. Hannah Arendt drew a sharp distinction between power, which arises from the capacity of people to act in concert, and the instruments through which force is exercised. AI infrastructure is an instrument. But the question digital sovereignty raises is whether nations that depend on foreign instruments retain the capacity for collective action that Arendt considered the foundation of genuine political life. A country whose critical systems run on models it did not build, trained on data it does not control, hosted on servers it cannot access, faces a deficit not merely of technology but of agency. The African Union declared AI a strategic priority in 2025, reflecting the recognition that digital sovereignty is not only a concern for major powers but an urgent question for the developing world as well.

In response, some governments are investing heavily in national AI strategies designed to foster domestic research, infrastructure, and talent development. The European Union launched the OpenEuroLLM initiative to develop large language models across its 24 official languages. Latin American nations launched Latam-GPT to build regional AI capacity. Others seek to negotiate international standards that preserve interoperability while protecting national interests.

Digital sovereignty therefore represents an emerging axis of geopolitical negotiation, one in which the partnerships between governments and technology companies are as consequential as the competition between states.

Glass Half Full

Technological revolutions have historically reshaped the balance of power between nations, but they have also created opportunities for cooperation. The same scientific networks that fueled competition during the twentieth century helped establish norms around nuclear safety, space exploration, and global telecommunications.

Artificial intelligence may follow a similar trajectory. While strategic competition will remain a defining feature of the AI era, shared challenges such as climate change, pandemic preparedness, and global economic stability require collaborative problem solving. DeepMind’s AlphaFold project, which predicted the three-dimensional structures of nearly all known proteins and made the results freely available, demonstrated what becomes possible when AI capability meets a commitment to open science. Climate modeling consortiums are integrating machine learning to improve prediction accuracy. Medical research collaborations are using AI to accelerate drug discovery across national boundaries.

AI’s greatest potential may lie in its ability to help nations address problems that transcend borders. Environmental forecasting, medical research, disaster prediction, and food security all benefit from analytical systems capable of integrating vast global datasets. The technology companies building these systems have both the capability and, increasingly, the institutional incentive to support cooperative efforts alongside competitive ones.

The emerging global order of the interface will therefore reflect both rivalry and cooperation. Nations will compete to develop capabilities, yet they may also recognize that the most consequential applications of those capabilities require collective effort.

Artificial intelligence will not abolish geopolitics.

It will redefine it.

And the choices nations and corporations make today will determine whether the interface becomes a source of division or a platform for shared progress.

Further Reading

Ian Bremmer, “The Technopolar Moment: How Digital Powers Will Reshape the Global Order,” Foreign Affairs (2021); updated in “The Technopolar Paradox,” Foreign Affairs (2025).

Bremmer coined the concept of a “technopolar” order in which technology companies exercise geopolitical influence rivaling that of nation-states. His 2025 update examines how the fusion of tech power and state power has deepened, creating a bifurcated digital world. Essential for understanding the corporate dimension of AI geopolitics.

Ian Bremmer and Mustafa Suleyman, “The AI Power Paradox,” Foreign Affairs (2023).

A joint essay by the Eurasia Group founder and the co-founder of DeepMind (now CEO of Microsoft AI) arguing that AI’s creators are themselves geopolitical actors whose sovereignty over the technology entrenches the technopolar order. Engages the question of whether policymakers can catch up to the pace of AI development.

Kai-Fu Lee, AI Superpowers: China, Silicon Valley, and the New World Order (Houghton Mifflin Harcourt, 2018).

The definitive account of US-China AI competition from an executive who led AI divisions at Apple, Microsoft, and Google before founding Sinovation Ventures. Lee’s framing of data as a strategic resource and his analysis of China’s four building blocks for AI supremacy remain influential.

Henry Kissinger, Eric Schmidt, and Daniel Huttenlocher, The Age of AI: And Our Human Future (Little, Brown, 2021).

A former Secretary of State, a former Google CEO, and an MIT dean examine AI as a civilizational transformation comparable to the Enlightenment. Their treatment of AI’s implications for statecraft, deterrence, and epistemology connects the technology to the deepest traditions of geopolitical thought.

Elinor Ostrom, Governing the Commons: The Evolution of Institutions for Collective Action (Cambridge University Press, 1990).

Ostrom’s Nobel Prize-winning work on polycentric governance systems provides the theoretical foundation for understanding how overlapping institutions can manage shared resources without a single central authority. Directly relevant to the fragmented landscape of international AI governance.

Mariana Mazzucato, The Entrepreneurial State: Debunking Public vs. Private Sector Myths (Anthem Press, 2013).

Mazzucato’s research demonstrates how much foundational technological innovation originates in public investment rather than private enterprise alone. Her framework complicates any narrative that treats AI development as purely a private-sector achievement and argues for recognizing the symbiotic relationship between government and corporate innovation.

Daron Acemoglu and Simon Johnson, Power and Progress: Our Thousand-Year Struggle Over Technology and Prosperity (PublicAffairs, 2023).

Acemoglu and Johnson examine the historical pattern in which technological breakthroughs concentrate wealth unless institutional countermeasures ensure broad distribution. Their analysis of AI’s labor-market and inequality effects informs the question of whether AI-driven geopolitical competition will benefit populations or primarily serve elite interests.

Goldman Sachs, “The Generative World Order: AI, Geopolitics, and Power” (2023).

A major research report identifying the concept of “geopolitical swing states” in AI development, including the UK, UAE, Israel, Japan, and India. Useful for understanding how the AI landscape extends beyond the US-EU-China tripartite and how innovation blocs may form around differentiated capabilities.

Atlantic Council, “Eight Ways AI Will Shape Geopolitics in 2026” (January 2026).

A forward-looking assessment of sovereign AI strategies, AI stacks competition, and the institutional architecture emerging around global AI governance. Captures the acceleration of national AI initiatives and the growing role of bilateral technology partnerships.

“Digital Disintegration: Techno-Blocs and Strategic Sovereignty in the AI Era,” International Organization, Vol. 79, Issue S1 (December 2025).

An academic treatment of how states are asserting strategic digital sovereignty through unilateral measures and techno-bloc formation, with detailed analysis of the semiconductor value chain and the role of private actors in performing functions traditionally associated with state sovereignty.

Thucydides, History of the Peloponnesian War (c. 431 BCE); Graham Allison, Destined for War: Can America and China Escape Thucydides’s Trap? (Houghton Mifflin Harcourt, 2017).

Thucydides identified the structural instability that arises when a rising power threatens an established one. Allison’s modern revival of the concept applies the framework to US-China relations across multiple domains, including technology. Directly relevant to understanding how perceptions of shifting AI capability drive preemptive policy action.

Hannah Arendt, The Human Condition (University of Chicago Press, 1958).

Arendt’s distinction between power (the capacity of people to act in concert) and violence (the exercise of force through instruments) provides the philosophical foundation for understanding digital sovereignty as a question not merely of technology but of political agency. Her framework illuminates what nations risk when critical infrastructure depends on foreign systems.

G. W. F. Hegel, Phenomenology of Spirit (1807); The Philosophy of Right (1820).

Hegel’s dialectical method, in which competing positions provoke and shape one another toward synthesis, provides the structural logic for understanding how rival AI governance models interact. The Interface Series draws on Hegel throughout as a framework for thinking about institutional transformation through sustained tension rather than simple resolution.