In 2020, the government of Uganda signed a deal with Huawei to deploy facial recognition cameras across Kampala. Within months, Ugandan security services used the system to track opposition leader Bobi Wine and his supporters ahead of a contested election. The technology did not come from China’s military. It came from a commercial contract for “public safety infrastructure.” The effect was the same.
This is the new geography of power. Not borders drawn by armies, but dependencies created by technology — fiber cables, AI systems, cloud infrastructure, and data agreements that determine which countries can see, which can act, and which can be seen acting. The world map is being redrawn, and the instrument is artificial intelligence.
The Infrastructure of Influence
Control over physical infrastructure has always translated into geopolitical leverage. The British Empire’s dominance of undersea telegraph cables in the 19th century gave London intelligence advantages over every rival. American control of the SWIFT financial messaging system has allowed Washington to impose sanctions with global reach. The pattern repeats: whoever controls the infrastructure of communication and exchange controls the terms on which everyone else operates.
AI infrastructure is the 21st century version of this dynamic. The countries and companies that build and operate the AI systems other nations depend on — cloud computing platforms, AI development tools, semiconductor supply chains, submarine cable networks — acquire leverage over those nations that is structural rather than coercive. It does not require threats. It operates through dependency.
Amazon Web Services, Microsoft Azure, and Google Cloud collectively host a significant portion of the world’s digital infrastructure, including government systems in dozens of countries. When those governments need to develop AI capabilities, they build on American platforms, using American tools, generating data that flows through American systems. The relationship is commercially beneficial to both parties — and strategically advantageous to one.
China’s Infrastructure Diplomacy
China recognized the leverage embedded in infrastructure earlier than most Western observers acknowledged. The Belt and Road Initiative — formally an infrastructure investment program — included from its inception a digital component: undersea cables, terrestrial fiber networks, data centers, and telecommunications infrastructure built by Chinese companies in recipient countries.
The strategic logic was straightforward. A country whose telecommunications backbone runs on Huawei equipment, whose government cloud services run on Alibaba infrastructure, and whose urban surveillance systems run on Hikvision cameras has created dependencies that bind it to Chinese technology ecosystems. Those dependencies are not coercive — they can be ended, at significant cost — but they create switching costs and relationship structures that influence foreign policy decisions at the margin.
By 2023, Chinese companies had laid or were laying undersea cable connecting Africa to Asia and Europe along routes that bypassed American-allied infrastructure. Chinese data centers were operating in at least 65 countries. Huawei telecommunications equipment was present in the networks of more than 170 countries. The digital Belt and Road had created a parallel internet infrastructure — not fully separate from the Western-aligned internet, but increasingly capable of operating independently of it.
The Alliance Realignment
Traditional alliance structures were built around military security guarantees and economic partnerships. The AI era is adding a third dimension: technology alignment — shared standards, interoperable systems, and compatible data governance frameworks that create deep integration between countries using the same technology ecosystem.
The United States has pursued technology alignment through the Quad — the security partnership with Japan, Australia, and India — and through bilateral agreements on semiconductor supply chains, AI research cooperation, and data sharing. The CHIPS and Science Act, passed in 2022, included provisions designed to ensure that allied countries receiving US semiconductor investment align their export control policies with American restrictions on China.
This approach has been partially successful. Japan and the Netherlands — the two countries whose companies produce the equipment necessary to manufacture advanced semiconductors — have aligned their export controls with American restrictions. South Korea and Taiwan, deeply integrated into US-aligned technology supply chains, have maintained their positions despite significant Chinese economic pressure.
But the alignment is not universal. India, despite its Quad membership, has maintained relationships with Chinese technology companies and declined to fully align with American restrictions. Saudi Arabia, a longstanding American security partner, has signed major AI cooperation agreements with Chinese companies. Brazil, the largest economy in Latin America, uses Chinese telecommunications infrastructure extensively while maintaining political independence from both powers.
The emerging pattern is not a clean division into two technology blocs. It is a messy, interest-driven, country-by-country negotiation in which most nations are attempting to extract benefits from both sides while committing fully to neither.
The Data Sovereignty Movement
The recognition that data is a strategic asset has produced a growing movement among governments to assert control over data generated within their borders — a phenomenon called data sovereignty.
The European Union’s General Data Protection Regulation established the principle that European citizens’ data cannot be transferred to third countries without adequate privacy protections — effectively requiring American tech companies to treat European data differently from data generated elsewhere. The EU’s proposed Data Act and AI Act extend this framework, asserting European regulatory jurisdiction over AI systems that operate within European markets regardless of where those systems were built.
India has proposed data localization requirements that would require companies operating in India to store Indian users’ data within Indian borders — reducing the flow of Indian data to foreign AI training systems. Indonesia, Vietnam, and several African countries have pursued similar frameworks. The common thread is the recognition that data flows are value flows — that allowing foreign companies to collect and process national data transfers economic and strategic value outward.
The data sovereignty movement creates friction in the global AI ecosystem. AI systems trained on globally diverse data perform better than systems trained on nationally restricted data. Fragmentation of the global data environment may slow AI development overall while creating national AI ecosystems that serve political interests rather than technical ones. But from the perspective of governments that have watched their citizens’ data fuel the development of foreign AI systems with no return, the political logic is compelling.
The Standards War
Technical standards — the specifications that determine how systems communicate, how data is formatted, how algorithms are evaluated — are a form of geopolitical power that receives little public attention and produces enormous consequences.
The country or bloc that sets technical standards for AI systems determines the terms on which AI operates globally. Standards embedded in international bodies like the International Telecommunication Union, the International Organization for Standardization, and the Institute of Electrical and Electronics Engineers shape which technologies are interoperable, which are certified as safe, and which are effectively excluded from global markets.
China has pursued an aggressive standards strategy, increasing its participation in international technical bodies and proposing standards that favor Chinese technical approaches. The United States, historically dominant in technical standard-setting, has recognized the strategic importance of this arena and is attempting to coordinate allied positions in international standards bodies to maintain Western influence over AI governance frameworks.
The standards war is unglamorous. It happens in conference rooms in Geneva and online working groups, not in diplomatic summits or military exercises. But its outcomes will determine whether the AI systems that govern the world in 2040 operate according to frameworks that reflect democratic values or authoritarian ones — whether they are designed for transparency and accountability or for surveillance and control.
The Countries Being Left Behind
The geopolitical competition over AI infrastructure is primarily a contest between wealthy, technologically advanced nations. For the majority of the world’s countries — in sub-Saharan Africa, South and Southeast Asia, and Latin America — the more immediate reality is not which great power they align with but whether they have any meaningful capacity to participate in the AI economy at all.
The infrastructure gap is severe. Cloud computing, reliable high-speed internet, and the electricity infrastructure that AI systems require are unevenly distributed in ways that track closely with existing global inequality. A country without reliable electricity cannot run data centers. A country without high-speed internet cannot access cloud AI services. A country without universities producing AI researchers cannot develop indigenous AI capability.
The geopolitical competition between the United States and China is producing some infrastructure investment in developing countries — both powers are offering connectivity and AI development assistance as tools of influence. But the terms of that assistance — the ownership structures, the data agreements, the technology dependencies it creates — are determined by the interests of the powerful, not the needs of the recipients.
The risk is that the AI era replicates the pattern of previous technological revolutions: the countries that control the technology capture the value, and the countries that provide the raw materials — in this case, data and labor — receive a fraction of what their contribution is worth.
The Map That Is Being Drawn
The world map being drawn by AI is not a map of territories. It is a map of dependencies — of which countries can act autonomously and which require permission, of which populations are monitored and which do the monitoring, of which governments make decisions and which have decisions made for them by systems they did not build and cannot fully control.
This map will not be finalized in a treaty conference. It is being drawn now, in procurement decisions and infrastructure contracts and research partnerships and data agreements, by actors who are often not thinking about the map they are drawing.
The countries that understand what is being decided, and act accordingly, will have more agency in the world that results. The countries that do not will find themselves on the wrong side of a geography they did not know was being created.
If this analysis interests you, read next: AI and the New Cold War: Who Controls the Future