The President’s state visit to China has, beneath the diplomatic pleasantries, laid bare a stark reality: Beijing’s technological ambitions are accelerating at a pace that demands an urgent recalibration of our own national strategy. From quantum computing labs in Hefei to semiconductor fabs in Shanghai, the message is clear: China is not just catching up; it is poised to leapfrog in key domains. For Britain, this is not a time for geopolitical hand-wringing but for decisive action on sovereign AI and semiconductor capabilities.
During the visit, Chinese officials showcased next-generation AI models trained on domestic hardware, circumventing export controls with impressive efficiency. Their semiconductor ecosystem, while not yet at parity with Taiwan or South Korea, is developing alternative architectures and advanced packaging techniques that could bypass traditional bottlenecks. The message is unmistakable: the era of technological dependency is ending for them, and if we are not careful, we will find ourselves dependent on them.
Here in Britain, we must confront an uncomfortable truth. Our current semiconductor strategy, while a step in the right direction with the National Semiconductor Strategy, remains too fragmented and underfunded. We lack a dedicated AI compute infrastructure that is sovereign and secure. The private sector alone cannot build this; it requires a coordinated national effort akin to the Manhattan Project or the Apollo programme. We need a British Institute for AI and Semiconductor Sovereignty, a central body that coordinates research, funding, and deployment across universities, startups, and established firms.
Consider the user experience of our society. Every citizen interacts with technology that relies on chips designed abroad and AI models trained on foreign data. This is a vulnerability, both economically and in terms of privacy. Digital sovereignty means owning the means of production for the foundational technologies of the future. It means building our own large language models, trained on British values and data, compliant with our privacy laws. It means a semiconductor supply chain that does not collapse with a geopolitical tremor in the South China Sea.
The stakes are high. China’s ambition is not limited to hardware; it is embedding AI into governance, surveillance, and social credit systems. We must ensure our technology reflects our own ethical standards, building transparent, accountable systems that enhance democracy rather than undermine it. The Black Mirror consequences of ceding these technologies to foreign powers are all too real: algorithmic bias, loss of autonomy, and the erosion of trust.
So what must be done? First, establish a sovereign cloud for AI training, using British-designed chips where possible. Second, create a national semiconductor foundry for critical applications, perhaps through a public-private partnership with companies like Pragmatic Semiconductor. Third, invest in quantum computing research, as this will be the next battleground. Fourth, rewrite our digital regulation to encourage innovation while safeguarding privacy, making it easier for startups to build on British infrastructure. Fifth, and most importantly, treat this as a matter of national security, not just industrial policy.
The time for incrementalism is over. China has shown it can move fast. Britain must accelerate its own sovereign AI and semiconductor strategy, or risk being left behind in the race for the future. The user experience of our society depends on it.
