From London to Riyadh, states are no longer leaving artificial intelligence entirely to the private sector. Sovereign funds focused on AI infrastructure are becoming the new black and UK is now on trend.
Something has shifted in how governments think about artificial intelligence. It is no longer treated as a technology sector to regulate at arm’s length. It is increasingly seen as strategic infrastructure, as fundamental to national power as roads, energy grids, or defence budgets. Britain made that philosophy concrete this week, and it is far from alone.
Britain Bets on Its Own Founders
The UK’s Sovereign AI Fund unit is a £500 million first-of-its-kind national effort to back Britain’s smartest founders and keep the future of AI built on British shores. What makes it genuinely distinctive is not just the money, but the operating model. Unlike traditional government programmes, Sovereign AI is built to work at the pace of the AI industry, just like a top-tier VC firm.
The first cohort of backed companies gives a clear sense of the Unit’s priorities. The first equity investment went to AI infrastructure startup Callosum, while six further startups received access to some of the UK’s foremost supercomputing capacity. The fields represented span drug discovery targeting conditions like Alzheimer’s and Parkinson’s, cheaper supercomputing, and next-generation chip development — areas where British research talent is strong but commercial scaling has historically faltered.
“Sovereign AI is unlike anything government has ever done before. Its unique approach will help break down the barriers that have too often held back British enterprise and innovation. This is how we ensure Britain’s economic prosperity and national security in the modern age.”
UK Technology Secretary Liz Kendall
Beyond capital, the package is designed to remove the friction that typically kills promising startups before they can grow. Every company receiving investment gets visa decisions within a working day, plus access to an initial ten cost-free visas for the world’s top R&D talent, as well as fully funded access to the UK’s largest AI supercomputers, with up to one million GPU hours available per startup.
The United States: A Dual Approach
Washington has pursued a different but equally deliberate strategy — one that combines state capital with the mobilisation of private giants. On the sovereign side, the Trump administration’s newly created sovereign wealth fund has already taken a 10% stake in Intel and signalled plans to invest in chip startup XLight, receiving up to $150 million to develop advanced semiconductor lasers.
Running parallel is Stargate — a different animal entirely. The joint venture, created by OpenAI, SoftBank, Oracle, and MGX, plans to invest up to $500 billion in AI infrastructure in the United States by 2029. It is private capital with presidential blessing rather than public money, and its first data center in Abilene, Texas is already operational. The model is essentially government as enabler: fast-tracking permits, offering political support, and signalling national priority, while leaving the financial risk with the private sector.
China: State-Backed and Moving Fast
China’s AI progress has been fuelled by a combination of state direction, abundant engineering talent, and strategic patience. The most striking recent proof came from DeepSeek — a Chinese AI model that stunned Western observers at the start of 2025 by matching frontier American models at a fraction of the reported training cost, suggesting that export controls on chips may be less of a ceiling than previously assumed.
China has also launched an $8.2 billion early-stage AI fund, with one of its main focuses being “embodied AI.” And the strategic compute picture may be more formidable than it appears. Nvidia CEO Jensen Huang warned this week that China already has the computing power and data center capacity necessary to train an AI model at the same level as Anthropic’s Claude Mythos — a model with identified cyberoffensive capabilities. Huang said China manufactures 60% of the world’s mainstream chips, has some of the best computer scientists, accounts for 50% of the world’s AI researchers, and has an abundance of energy. The chip export controls, in other words, may not be the constraint many assumed.
Saudi Arabia: The Biggest Bet of All
If Britain is thinking in hundreds of millions and the US in hundreds of billions, Saudi Arabia is operating in a different register altogether. Project Transcendence, backed by the Public Investment Fund, has committed $100 billion to domestic compute infrastructure and sovereign AI capabilities, making it the largest single government AI pledge outside the US and China. The ambition is not merely to use AI, but to become a producer and exporter of it, reducing the kingdom’s long-term dependence on fossil fuel revenues by positioning itself at the centre of the next economy. At the same time, the project so far is in planning stage, and with worsened geopolitical situation in the region, this AI initiative may be put on hold for a while.
The broader pattern is unmistakable: AI sovereignty has become a matter of national policy, and the race to build it is accelerating on every continent.
