Sovereignty Hollowed: Britain’s AI Capture
Britain’s AI “sovereignty” is a mirage — US tech giants are wiring the country’s future to their own machines, bending policies, narrowing choices, and leaving citizens with the costs of capture.

Britain is being sold a dream of sovereignty. Ministers promise that billions from Microsoft, Nvidia, and OpenAI will transform the country into an “AI superpower.” Prime Minister Keir Starmer has declared that “Britain will be an AI superpower, with jobs and prosperity spread across the country.” Nvidia’s Jensen Huang echoed the line almost word for word, insisting that “the UK is going to be an AI superpower” as he unveiled a £500 million data centre investment. Corporate executives and government ministers alike now trade in the same vocabulary, reinforcing one another in a shared narrative of strength. But what is being built on British soil is not sovereignty. It is dependency, dressed up as triumph.
The term “sovereign compute” has become the buzzword of the moment. The idea is simple enough: Britain’s digital future will be secure because it hosts the world’s most advanced infrastructure at home. Yet the reality is more stark. The chips belong to Nvidia, the clouds to Microsoft, the models to OpenAI. Hosting is not the same as steering. Britain will own the land under the data centres, but not the systems that dictate how they operate or whose interests they serve.
This is not just a matter of corporate power. Microsoft, Nvidia, and OpenAI are not neutral actors. They are a key part of the industrial strategy, defence partnerships, and geopolitical ambitions of the United States. Their expansion into Britain extends American power, wiring the country’s digital nervous system directly into another state’s infrastructure. Britain may cut the ribbons, but the switches remain across the Atlantic.
Once captured, domestic governance does not vanish, it bends. Regulation becomes hesitant, softened by the fear of spooking investors whose presence has become indispensable. Energy priorities are reshaped to feed the insatiable demands of GPU farms, while communities live with strained grids and rising costs.
Recent signals from both government and Big Tech show how far this dependency already runs. AWS chief Matt Garman has openly urged Britain to embrace nuclear power to sustain the energy-hungry data centre boom, calling it a “great solution” for AI infrastructure. The government’s AI Growth Zones proposals include plans for dedicated nuclear capacity to feed these clusters. Starmer has backed the expansion of nuclear power across England and Wales, including the deployment of small modular reactors by 2032 and the lifting of long-standing geographic restrictions on new sites.
The justification is clear: powering AI. But this strategy comes at a cost. Nuclear power is the most capital-intensive and politically rigid form of energy. Once reactors are planned, approved, and built to service AI data centres, Britain will be locked into decades of policy commitments designed around the needs of foreign-owned infrastructure. Questions of safety, siting, and environmental impact become harder to contest when framed as essential to national technological ambitions. Sovereignty narrows still further: energy policy becomes subordinated to keeping corporate servers running, and democratic space for debate shrinks under the weight of sunk costs.
Universities tailor their teaching to Nvidia’s platforms, procurement systems assume Microsoft compatibility, and research funding flows toward corporate-defined frontiers. Industrial policy narrows, step by step, until it serves not the public interest but the priorities of firms tied to another state’s strategic design.
The consequences reach further still. For citizens, capture creeps into the fabric of everyday life. Welfare and healthcare systems increasingly rely on AI platforms that cannot be meaningfully scrutinised or held accountable locally. Public data moves seamlessly into opaque corporate pipelines. Choices shrink when every service, from banks to councils, is tethered to the same few infrastructures. Surveillance grows without equivalent oversight, leaving citizens less protected, not more.The consequences reach further still. For citizens, capture creeps into the fabric of everyday life. Welfare and healthcare systems increasingly rely on AI platforms that cannot be meaningfully scrutinised or held accountable locally. Public data moves seamlessly into opaque corporate pipelines. Choices shrink when every service, from banks to councils, is tethered to the same few infrastructures. Surveillance grows without equivalent oversight, leaving citizens less protected, not more.
What is described as sovereignty looks increasingly like its opposite. The rhetoric of national strength obscures the fact that Britain cannot legislate meaningfully against companies that have become essential to its own infrastructure. To host foreign-owned systems and call it sovereign is to confuse geography with power.
This moment is not without precedent. Colonial railways were laid across India and Africa under the banner of modernisation, but they were designed to extract resources outward, not to empower local communities. Britain now risks constructing the digital equivalent: infrastructure optimised to feed the pipelines of Microsoft, Nvidia, and OpenAI, embedded in the wider machinery of US power.
To call this capture is not exaggeration. It is to see clearly what is happening. Britain is swapping independence for the illusion of control: jobs and headlines in exchange for infrastructure that remains tied to another state’s strategic interests. Sovereignty is invoked to celebrate these projects, but in reality, it has been surrendered.
(Image generated by Sora)