AI governance is becoming infrastructure governance

For much of the past decade, discussions on AI governance have focused on algorithms. Policymakers, regulators, and researchers have debated ethics, transparency, bias, accountability, and human oversight, seeking to ensure that AI systems remain trustworthy and aligned with fundamental rights.

These questions remain essential, but they are no longer sufficient to explain the direction of AI policy.

A new reality is emerging. Increasingly, governments are recognising that the ability to govern AI depends not only on regulating algorithms but also on securing the infrastructure that makes them possible. Data centres, advanced semiconductors, cloud computing, electricity grids, submarine cables, digital public infrastructure, and skilled workforces have become as strategically important as the AI models themselves.

The shift has been gradual rather than abrupt, yet it is now evident across national strategies, international organisations, and multilateral discussions. AI governance is expanding beyond software governance to become infrastructure governance.

AI runs on infrastructure, not algorithms alone

The rapid adoption of generative AI has often obscured a simple reality that every AI system rests on a vast physical and institutional foundation.

Large language models require enormous computing power, specialised chips, cloud infrastructure, reliable electricity, high-speed connectivity, trusted datasets, secure digital environments, and engineers capable of developing and maintaining increasingly complex systems. None of these components can be built overnight, and few can be developed without substantial public and private investment.

As a result, AI capability is increasingly determined by access to infrastructure, not just software.

Countries may have ambitious AI strategies, talented researchers, or innovative start-ups, but without advanced computing capacity, modern data centres, resilient cloud infrastructure, and skilled human capital, those ambitions become difficult to realise. Governments are therefore beginning to treat AI infrastructure as a strategic national asset rather than simply a commercial resource.

Industrial policy is becoming a pillar of AI policy

This shift is increasingly reflected in public policy.

The European Union has complemented its landmark AI Act with broader industrial initiatives, including the European Chips Act, the AI Factories initiative under EuroHPC, and, most recently, plans to develop AI Gigafactories capable of supporting the next generation of AI models. The recently adopted Digital Omnibus on AI also strengthens the role of the European Commission’s AI Office while simplifying implementation aspects, illustrating that regulation, institutional capacity, and infrastructure investment are evolving together.

The United States has pursued a different approach, combining export controls on advanced semiconductors with large-scale investment in domestic chip manufacturing and AI infrastructure. China continues to expand its national computing centres, cloud capacity, and state-backed AI ecosystems as part of its long-term industrial strategy.

Although these approaches differ politically and economically, they increasingly share the view that governing AI requires building the infrastructure that enables it.

AI policy, in other words, is beginning to resemble industrial policy.

Infrastructure has become a development issue

The infrastructure shift is equally visible in developing countries, where the conversation is moving beyond access to AI applications towards the ability to build AI capacity locally.

Throughout the World Summit on the Information Society (WSIS) Forum 2026, discussions repeatedly highlighted that meaningful digital transformation depends on foundational infrastructure. African policymakers, researchers, and technical experts argued that the continent suffers not primarily from a shortage of digital ambition, but from limited access to computing capacity.

Speakers pointed out that Africa has only a tiny share of global AI compute and data centre capacity, forcing many researchers and innovators to rely on infrastructure located elsewhere. The challenge, they argued, is no longer simply connectivity but compute sovereignty, the ability to build, train, and deploy AI systems locally.

The same message emerged during discussions on digital sovereignty. Rather than focusing solely on access to foreign AI models, participants argued that exporting raw data while importing AI services mirrors older economic patterns in which countries exported raw materials while importing higher-value finished products. Building local AI capability, therefore, requires investment in data centres, energy systems, cloud infrastructure, skills development, and trusted data ecosystems.

These discussions suggest that AI infrastructure is becoming an increasingly important component of development policy.

Infrastructure is also becoming geopolitics

The growing strategic importance of AI infrastructure extends far beyond economic development.

Competition over semiconductor manufacturing, cloud services, critical minerals, advanced computing facilities, and energy has become a defining feature of international relations. Governments increasingly view these assets through the lens of economic security, technological resilience, and geopolitical influence.

Submarine cables illustrate this evolution particularly well. Once regarded primarily as telecommunications infrastructure, they are now recognised as essential to AI systems that depend on massive volumes of cross-border data traffic. Recent international efforts to strengthen submarine cable resilience reflect how infrastructure once considered largely invisible has become central to digital governance.

Electricity is undergoing a similar transformation. As AI models become increasingly computationally intensive, access to affordable and reliable energy is becoming a key factor determining where data centres can be built and where AI innovation can flourish. This has encouraged governments to align AI strategies with broader industrial, energy, and climate policies.

In this environment, AI governance increasingly overlaps with trade policy, investment screening, industrial strategy, critical infrastructure protection, and national security.

A broader understanding of AI governance

This evolution does not diminish the importance of ethical AI principles or risk-based regulation. Questions surrounding transparency, accountability, bias, safety, and human rights remain fundamental.

Rather, it broadens the understanding of what AI governance entails.

Traditional AI governance focuses on how AI systems should be designed, deployed, and used responsibly. Infrastructure governance raises a different set of questions, such as who has access to the computing power needed to develop advanced AI? Who controls the cloud infrastructure that underpins AI services? Where are critical datasets stored? Which countries have the physical, financial, and institutional capacity to participate meaningfully in the AI economy?

These questions ultimately shape who can innovate, compete, and benefit from AI.

As a result, debates that once focused primarily on algorithms increasingly encompass semiconductors, cloud infrastructure, digital public infrastructure, data centres, energy systems, connectivity, and workforce development.

From governing AI to enabling AI

The next phase of AI governance is likely to be defined less by entirely new regulatory principles than by long-term investment decisions.

Building trustworthy AI ecosystems will require governments to develop resilient digital infrastructure, strengthen education and digital skills, encourage research, modernise energy systems, expand computing capacity, and foster international cooperation on shared digital resources.

The countries that succeed may not simply be those that write the most comprehensive AI regulations. They may instead be those that create the conditions that allow responsible AI to flourish in the first place.

The global conversation about AI has not moved beyond governance. It has moved deeper into its foundations.

In this sense, the future of AI governance may increasingly depend not only on how societies regulate AI, but also on how they build the infrastructure that makes AI possible.

Would you like to learn more about AI, tech and digital diplomacy? If so, ask our Diplo chatbot!