UK Launches £500m Sovereign AI Fund with Initial Investment in Callosum

The UK government formally launched its £500 million Sovereign AI unit on April 16, 2026, establishing a state-backed venture capital organization to invest in domestic technology firms. The initiative began with two equity investments on launch day, including a shareholding in London-based chip startup Callosum. Technology Secretary Liz Kendall described the fund as one of the most critical steps for the country's future competitiveness.

The Details

The Sovereign AI unit, which operates within the Department for Science, Innovation and Technology (DSIT), is designed to provide support that extends beyond the capabilities of traditional commercial funders. As part of its first wave of activity, the government took an undisclosed stake in Callosum, a company founded in 2024 that develops technology to improve the efficiency of computer chips used to train and operate AI models. A second, currently unnamed business also received an equity investment during the launch.

Beyond direct equity, the fund incorporates the provision of computational power. Six UK companies have been granted access to government-funded supercomputers to develop AI models. In exchange for this resource, the government maintains a 'right of first refusal' to invest in these firms. This computational support is factored into the overall £500 million fund value.

The six startups selected for supercomputer access include Prima Mente, which focuses on biological foundation models for diseases such as Alzheimer's, and Cosine, an agentic AI firm. Other recipients include Cursive, a generative internet specialist founded by former Google DeepMind alumni, as well as Doubleword, Odyssey—which specializes in world models and simulations—and one additional unnamed firm.

The official launch took place at the offices of Wayve, a self-driving car startup. The choice of venue highlighted the scale of the UK's AI ambitions, as Wayve is currently valued at $8.6 billion (£6.4 billion).

Chancellor Rachel Reeves emphasized that the objective is to create 'national AI champions.' By providing targeted state support, the government aims to ensure that internationally competitive companies are able to start, scale, and remain headquartered in Britain rather than moving operations abroad.

Context

The Sovereign AI fund was first outlined in the UK government's Spending Review as a strategic response to the global AI race. The move comes amid a complex landscape of opportunity and risk. In January 2026, Secretary Kendall acknowledged that AI automation would likely lead to some job losses, though she maintained that entrepreneurs believe the technology will ultimately create new employment opportunities.

Safety and security concerns remain a primary point of discussion for the government. Kendall specifically referenced a recent discovery by US-based startup Anthropic, which revealed an AI model capable of posing significant cyber threats. Despite these risks, the UK administration has signaled that the potential benefits for national productivity and global standing outweigh the dangers, provided the technology is 'seized' effectively.

The UK's attractiveness as a tech hub was echoed by Callosum co-founder Danyal Akarca, who cited the nation's strong university talent and the presence of established private labs like DeepMind as primary reasons for basing his company in London.

What's Next

The government will now move to monitor the growth of the initial portfolio of companies and evaluate the impact of the supercomputer access program. Future investments will depend on the performance of the first round of startups and the government's ability to attract further high-growth AI firms to the UK ecosystem.

Observers will be watching for the announcement of the second equity investment made on launch day, as well as any updates on the 'right of first refusal' exercised by the Sovereign AI unit. The success of the fund will likely be measured by whether these firms can scale into global leaders while remaining based in Britain.

Additionally, the government is expected to maintain a balancing act between aggressive investment and the safety protocols required to mitigate the cyber risks highlighted by the Anthropic findings.