In a decisive move to anchor the United Kingdom’s position in the global technology race, Technology Secretary Liz Kendall has officially launched the government’s £500 million Sovereign AI Fund. Speaking from the headquarters of autonomous vehicle pioneer Wayve, Kendall urged the British public to look past fears of job displacement and cybersecurity threats to embrace a future where AI serves as a primary engine for national growth. The fund represents a strategic shift in British industrial policy, transitioning the state from a mere regulator to a direct venture stakeholder in the nation’s most promising startups.
The fund’s inaugural investment has been directed toward Callosum, a London-based firm specializing in the optimization of heterogeneous chip architectures for AI model training. By improving the efficiency with which different semiconductors communicate, Callosum addresses a critical bottleneck in the current AI infrastructure. This first deployment of capital underscores a broader mandate: ensuring that the foundational layers of the artificial intelligence revolution remain rooted in British soil rather than migrating to the well-funded ecosystems of Silicon Valley.
To balance the high costs of innovation, the government is introducing a 'compute-for-equity' model. Six selected startups, including the Alzheimer’s-focused Prima Mente and the DeepMind-alumni-led Cursive, will gain access to the UK’s elite supercomputing network. In return, the state secures preferential investment rights, effectively turning taxpayer-funded infrastructure into a long-term sovereign asset. This approach aims to provide local entrepreneurs with the massive computational power required to compete globally, a resource traditionally dominated by American tech giants.
Chancellor Rachel Reeves highlighted that this sovereign intervention is designed to break the cycle of British startups scaling up abroad. By providing a combination of capital, compute power, and political backing, the government intends to ensure that the next generation of 'world models'—AI systems capable of interacting with physical reality—are developed and commercialized within the UK. Despite acknowledging that some roles will inevitably be automated, officials maintain that the economic dividends of a robust domestic AI sector far outweigh the transitional risks to the labor market.