UK launches $675 million sovereign AI fund to back homegrown startups

The UK government is launching a $675 million sovereign artificial intelligence fund. It is the largest government initiative in the history of Britain's AI sector, aimed at supporting domestic startups and cutting the country's critical dependence on foreign technology platforms.

There is serious geopolitical logic behind the decision. The critical layers of AI infrastructure today, including foundation language models, compute and cloud platforms, are concentrated mostly in the hands of American corporations: OpenAI, Google, Microsoft and Meta*. European governments largely find themselves in the role of consumers of other countries' technology. The UK intends to change that by betting on its own developments and preserving its technological independence.

The fund will focus primarily on early-stage companies, from seed to Series A, in AI infrastructure, industrial AI, healthcare and national security. The fundamental condition for recipients: their developments must remain under British jurisdiction. Data, model weights and intellectual property must not leave the country. In effect, this requirement sets a new standard for a sovereign contract between a state and an AI company.

The initiative fits into London's consistent, years-long strategy. In 2023, the UK hosted the first global AI Safety Summit at Bletchley Park, bringing together representatives of 28 countries and the world's leading technology companies. The AI Safety Institute, created as a result, took on the evaluation of risks from frontier models. In parallel, a national AI supercomputer and an expanded budget for the Alan Turing Institute were announced. The sovereign fund is a commercial layer on top of this whole architecture: venture money for private companies from which British AI champions are supposed to grow.

For the entrepreneurial community, this is an important institutional signal. The UK has historically led Europe in venture investment in AI: in 2024, British companies raised more than $6 billion in total from private investors. Direct government participation, however, long remained minimal. The sovereign fund shifts the balance: the state acts as a co-investor that reduces risk for private capital and opens room for long-term projects with a horizon of seven to ten years, those that struggle to attract classic venture capital with its three-year exit.

Critics point to the limited scale. $675 million is less than Microsoft put into OpenAI in a single round. Next to the multibillion-dollar AI funds of the US, China and Saudi Arabia, the sum looks modest. There is also a talent problem: keeping AI talent in London when six-figure offers arrive from San Francisco cannot be done with government money alone. It takes a full set of conditions: competitive salaries, a flexible visa regime and a strong academic environment.

Still, the direction of travel is clear. More and more states see artificial intelligence not as a sector-specific technology but as an element of critical national infrastructure, on a par with energy and defense. The UK, which has bet on regulatory leadership in AI safety, is now backing it with direct investment. The question is whether the country can secure a strong enough position before the market for foundational AI technologies is finally divided among two or three global players.

*Meta is designated an extremist organization and banned in Russia.