AI Nationalism and the Fracturing Digital Order

The single most important fact about AI geopolitics right now is this: the United States has stated, in official policy documents, that its objective is “unquestioned and unchallenged global technological dominance.” Not competitiveness. Not leadership. Dominance. That word choice matters, because it tells you everything about the strategic posture driving semiconductor export controls, alliance formation, and the weaponisation of cloud infrastructure.

I have been tracking the emerging techno-economic blocs forming around AI for a while now, through my work on Orchestrix and related research. What follows is a strategic assessment of AI nationalism — the forces fracturing the global digital order into competing, and increasingly incompatible, technological ecosystems.

Three visions, one tech stack

The US, China, and the EU each bring fundamentally different theories of the case.

The American doctrine fuses private-sector innovation with defensive economic statecraft. The White House AI Action Plan calls for exporting “secure, full-stack AI export packages” to allies — hardware, models, software, standards — creating durable technological dependency. Offensively, it turbocharges domestic innovation. Defensively, it uses semiconductor export controls to actively degrade China’s capacity to compete. This is not a trade dispute. It is economic warfare with a coherent strategic logic: weaken the rival’s ecosystem while making your own the only viable high-end alternative.

China’s strategy treats AI as sovereign geopolitical infrastructure. The 2017 New Generation AI Development Plan sets a phased roadmap to become the world’s primary AI innovation centre by 2030, with a core industry worth over 1 trillion yuan. The Military-Civil Fusion doctrine eliminates barriers between civilian tech and military applications. And the Digital Silk Road exports Chinese technology, governance models, and technical standards to the Global South — not just selling hardware, but building an alternative ecosystem insulated from US influence.

The EU occupies a third position as regulatory superpower. The AI Act is the world’s first comprehensive legal framework for AI, built on risk-based categorisation. But the EU faces what I call the Regulator’s Dilemma: the compliance costs it imposes disproportionately benefit the large US and Chinese firms that can absorb them, potentially entrenching the very incumbents it seeks to constrain. The Brussels Effect may successfully export European values while undermining European competitiveness.

The chip wars are the sharpest front

In October 2022, the US implemented sweeping export controls designed to sever China’s access to high-end semiconductor technology — not just chips, but the manufacturing equipment needed to produce anything below 14nm. This was a decisive shift from denying military applications to impeding an entire sector of a rival’s economy.

To sustain this, the US formed the Chip 4 alliance: the US, Japan, Taiwan, and South Korea. The strategic logic combines US design leadership, Dutch lithography (ASML), Japanese materials, and Taiwanese/Korean manufacturing. But the alliance contains fault lines. South Korean firms have massive investments in China and depend on the Chinese market for the majority of their revenue. They are caught between their security ally and their largest commercial market.

The result is a sovereignty paradox: pursuing independence from a rival forces deeper interdependence with allies. Decoupling from China means recoupling with the Chip 4 bloc, with all the internal friction that entails.

China’s response has been to accelerate domestic self-reliance. Huawei’s development of a 7nm chip using SMIC’s existing technology — under crippling sanctions — demonstrates that restrictions on one layer of the tech stack can spur breakthroughs in another. DeepSeek-V3 achieved frontier-comparable performance at a fraction of the training cost, innovating around hardware constraints. Necessity, it turns out, really is the mother of invention.

Cloud, data, and the new sovereignty

The battle extends beyond hardware. Cloud infrastructure — once logistical backrooms — has become a strategic battleground. The US dominates global compute through AWS, Azure, and Google Cloud, and is beginning to treat cloud access as a tool of statecraft: a privilege to be granted or withheld.

This has fuelled “data sovereignty” movements worldwide. Data localisation laws proliferate. The US CLOUD Act, which gives law enforcement access to data held by American firms regardless of location, has intensified these concerns even among allies. The result is a new market for what amounts to “Sovereignty as a Service” — cloud providers selling the feeling of sovereign control while deepening actual dependency on foreign-owned infrastructure.

The open-source dilemma

The most interesting internal tension in the US strategy is the debate over open-source AI. Meta argues that releasing models like Llama entrenches the American ecosystem as the global standard. The national security community counters that open-sourcing state-of-the-art model weights hands competitors billions of dollars in free R&D, neutralising the semiconductor export controls.

This is not a minor policy disagreement. It is a clash between two core pillars of the American strategy: champion a globalised innovation ecosystem, or pursue technological containment. You cannot do both simultaneously. The resolution will determine whether the AI landscape becomes more multipolar or more tightly controlled.

The Global South is the decisive arena

The nations of the Global South — the majority of the world’s population — are where the competition will be won or lost. China has been systematic here, using the Digital Silk Road to provide affordable, often turnkey AI solutions: smart city tech, surveillance systems, digital government platforms, bundled with financing. The US approach has been perceived as less coherent, more focused on restriction than on offering alternatives.

But Global South nations are not passive. They are forming alliances — the African Union, ASEAN, Mercosur — to pool diplomatic capital, co-develop standards, and negotiate as blocs. They are building regional compute hubs and asserting data sovereignty. They refuse to accept digital colonialism from either side.

What this means

The fiction of a single, unified global technology market is over. Strategy now requires accounting for competing blocs, divergent standards, and weaponised supply chains. The location of a data centre, the nationality of a cloud provider, the provenance of a semiconductor — these are no longer logistical details. They are core strategic concerns.

For those of us building in the AI space, the implications are concrete. The tools we use, the platforms we deploy on, and the standards we adopt are not neutral technical choices. They are geopolitical alignments, whether we acknowledge it or not. The fracture is here. The question is how we navigate it.

This post is adapted from research conducted as part of the Orchestrix strategic AI analysis project.