Geopolitics in a new age of technological dependence

Control over advanced technologies is no longer exercised solely through what states can build or deploy. It is increasingly determined by who governs the systems that make those technologies usable in the first place. From semiconductors to cloud infrastructure, power now operates through access and the ability to permit, restrict, and condition participation in the technological ecosystem itself.

Noah Tan, a junior fellow for the Technology and International Affairs Program at the Carnegie Endowment for International Peace, said this shift reflects a deeper reorganization of technological power. Speaking with The Beiruter, he described a landscape in which influence no longer rests primarily in innovation alone, but in control over the infrastructure, supply chains, and systems that determine how and by whom technologies are accessed.

Infrastructure as the new center of leverage

Much of the public conversation around artificial intelligence remains focused on chips and model development. Tan argues that this emphasis is increasingly incomplete.

When we think about the AI ecosystem, we must go beyond the design of the chip. Not only are we in a compute-constrained world, but we are also in a power-constrained world.

In practical terms, this reframing directs attention toward the physical and economic foundations of AI. Data centers operate at a scale that demands vast and stable energy supply, access to land, and the ability to deploy infrastructure quickly. A 2024 report by the International Energy Agency estimates that global data center electricity consumption could double by the end of the decade, underscoring the growing importance of energy as a limiting factor in AI expansion.

Tan emphasizes that these constraints elevate certain geographies within the global system. Countries capable of offering low-cost energy, favorable regulatory environments, and rapid permitting processes are not peripheral actors. They are emerging as key nodes within the AI supply chain. As Tan notes:

If you are a country that is endowed with all of this—cheap energy, cheap land, quick permitting, sovereign wealth funds, and a lack of ESG or regulatory constraint—that is leverage.

This helps to explain the concentration of data center investment in regions such as the Gulf, where these conditions converge, he explains. In this context, control is no longer confined to technological design. It extends to the ability to host, power, and scale the systems that sustain it.

Access as a mechanism of control

As infrastructure becomes more central, access to it is no longer taken for granted but actively controlled. Compute, datasets, and distribution channels are not neutral inputs. Who can access them, and on what terms, Tan explains, is determined by what he characterizes as an interaction between leading private firms and the governments that regulate them.

Large cloud providers and chip manufacturers determine pricing and availability, while states impose restrictions through export controls and security frameworks. A 2025 analysis by the Center for Strategic and International Studies highlights how U.S. export controls on advanced semiconductors have extended beyond limiting access to China, reconfiguring global supply chains and influencing third-country behavior.

Within this system, Tan argues, access is increasingly conditional.

Access to compute is also becoming both a weapon and a bargaining chip as a tool of alignment. You get the compute if you align with the security standards of the provider state.

This dynamic reflects a broader shift in how power is exercised. Rather than denying access outright, states and firms are embedding political and security conditions into the infrastructure itself. Agreements over cloud services or chip supply now carry implications that extend well beyond the technical domain.

The result is a system in which access becomes a site of negotiation. Influence lies not only in the ability to produce advanced technologies, but in the authority to determine who can use them and under what terms.

The expansion of private authority

A defining feature of this landscape is the growing role of private actors. Tan situates this shift within a broader transformation of authority. Historically, states exercised power through control over territory, populations, and physical goods. Today, Tan notes, that framework is being extended into domains governed by infrastructure and data.

Private actors have increasingly been able to convert this technological dominance and market share into actionable political power that would otherwise be outside their jurisdiction.

This transformation is reinforced by the growing reliance of governments on privately operated systems. A 2024 report by the International Institute for Strategic Studies notes that critical national functions, from public administration to military communications, are increasingly dependent on commercial cloud providers. In this arrangement, the boundary between public authority and private infrastructure becomes less distinct.

When states entrust core functions to external platforms, they are not simply outsourcing services. They are integrating elements of governance into systems that operate beyond traditional jurisdictional boundaries. This, Tan argues, complicates the exercise of sovereignty and redistributes power across a wider set of actors.

Strategic interdependence and the limits of autonomy

Despite intensifying competition, the trajectory of the system does not point toward complete technological independence. Tan argues that the emerging landscape is better understood as a highly interconnected order, in which different actors control different segments of the same system.

We are moving toward a world of strategic interdependence, where no state— perhaps not even the United States or China—can be truly sovereign when it comes to technology.

Recent research supports this assessment. A 2025 study in the Journal of International Economics finds that disruptions in semiconductor supply chains produce ripple effects across multiple regions, reflecting the extent to which production, design, and assembly are distributed globally. Similarly, analyses by the International Monetary Fund have shown that technological fragmentation increases costs and inefficiencies without fully eliminating cross-border dependencies.

Within this system, Tan explains, control is distributed across multiple nodes, with semiconductor manufacturing, data generation, and infrastructure often located in different regions yet functioning as an integrated whole.

This does not eliminate competition but reframes it. While major powers such as the United States and China continue to pursue full-stack autonomy—investing in domestic chip manufacturing, cloud infrastructure, and end-to-end control—this objective remains difficult to fully realize. For middle power states, Tan argues, the focus shifts away from achieving full autonomy and toward securing advantageous positions within a network of interdependence.

Leverage in a constrained landscape

For countries that do not control the full technological value chain, this environment presents a complex strategic challenge. Attempts to replicate entire ecosystems domestically are often constrained by cost, capability, and time. At the same time, reliance on external systems introduces vulnerabilities.

Tan argues that the starting point for these countries should be a reassessment of their position within the system.

“The first thing that these countries should recognize is that they are not powerless. They simply need to find their leverage points,” he said.

These leverage points vary across contexts, Tan explains. Some countries possess favorable conditions for data center development, including energy availability and regulatory flexibility, while others occupy critical positions within semiconductor supply chains, producing specialized components that are difficult to substitute. In some cases, he adds, value lies in data itself, particularly when it is generated through real-world processes that cannot be easily replicated elsewhere.

Tan points to the example of wartime data collection as a source of strategic value, where real-time information can shape both technological development and operational decision-making. This illustrates how leverage can emerge from participation in the system rather than control over it.

The strategic objective, in this context, is not to eliminate dependence but to manage it. Countries can position themselves as indispensable within specific segments of the supply chain, using that position to negotiate access, shape standards, and advance their own technological interests.

This approach reflects a broader shift in how power is understood. Control is no longer derived from dominance across the entire system, but from the ability to influence key points within it.