The UK government has unveiled a $1.47 billion plan to become less dependent on foreign artificial intelligence equipment.
Under the measures announced on Monday, Britain will spend more than $1 billion on a national AI supercomputer. It will be equipped with $530 million in hardware, including $200 million for specialized inference chips to process AI tasks. Priority will be given to emerging UK companies in the procurement process; the government has identified Olix and Fractile – two British start-ups developing recent styles of inference chips – as potential beneficiaries. British researchers and start-ups are expected to be able to utilize the supercomputer from 2030.
The recent measures are part a broader effort by the UK government to minimize dependence on foreign powers for access to AI products and services – a move that is more urgent given the apparent deterioration in relations between the US and its European counterparts. The European Union presented a a similar proposal for “technological sovereignty.” last week. This year, European leaders found themselves in confrontation with the Trump administration on a variety of issues, including: Greenland’s sovereignty Down tariff policy Down immigration, leading to speculation about A deterioration of the NATO alliance. In this context, dependence on American technology may constitute a burden that the United States can utilize as leverage against European countries.
“The geopolitical understanding of the last 40 years has been broken – and many say it is gone for good,” British Technology Secretary Liz Kendall said in an April speech to the Royal United Services Institute, a defense and security think tank. “For the UK, AI sovereignty is about reducing over-dependencies and increasing resilience.”
“There are those who say this race is already lost – that it is too late to challenge the dominance of the United States or China in AI chips – but I do not accept such defeatism,” she added.
Last November, the UK began to establish “AI development zones“, regions across the country with fewer administrative and regulatory barriers to building data centers. The $675 million SovAI venture capital fund was launched in April to invest in homegrown AI startups in fields ranging from model development to agent-based AI to drug discovery. The plan for supercomputing hardware is the latest piece of this expanding mosaic.
Although the UK is home to well-known companies such as ARM, whose chip architectures are ubiquitous around the world, semiconductor design and manufacturing is otherwise dominated by American and Asian companies. By acting as a vast customer of domestic chip startups, the UK government seeks to both support their growth and encourage them to stay in the country long-term.
“Historically, the UK government has just been inscrutable… being willing to support UK businesses with innovative technologies through tough deals is a really important milestone,” says Ed Bussey, CEO of Oxford Science Enterprises, a venture capital firm that participated in Fractile’s seed round in 2024. “If we can build a revenue supply pipeline for these companies, it will help anchor them here.”
The changes taking place in the design of AI data centers – away from homogeneous fleets of chips towards a set of specialist equipment for different purposes – represent an opportunity for the UK to carve out a strategically crucial niche.
“You can’t do everything yourself, so you really have to be bold in what areas you want to specialize in,” says Keegan McBride, director of science and technology at the Tony Blair Institute, a think tank founded by the former British prime minister. “Britain is playing a very smart game… If it plays it right, it has a huge opportunity. If other companies start relying on British chips, you will have an advantage.”