Britain's AI minister has said the country does not need to build its own cutting-edge AI model to keep up with the United States and China. Speaking during a visit to Silicon Valley, Kanishka Narayan argued that reliable access to models built by allies matters more than copying them, and that the government will put its money elsewhere.
What the minister said
City AM reported on 18 November that Narayan was asked directly whether Britain needed a home-grown frontier model to compete with Washington and Beijing. His answer: "I don't see it as a matter of necessity as long as we have trusted interdependence and access."
He said the government would back British companies that try to build such a model and do well, but that he was not planning to pour large sums of public capital into that goal. His focus, he said, was on building things other countries come to depend on, so that Britain has leverage of its own.
The areas he named are chip design, compound semiconductors and scientific research. In plain terms, the plan is to own a few important pieces of the AI supply chain rather than the whole thing.
Why the AI Security Institute matters to this argument
Narayan pointed to the UK's AI Security Institute (AISI) as evidence that Britain can keep influence without making its own frontier system. City AM noted that AISI has secured access to advanced models before they are released to the public.
In an interview with Axios before the trip, published on 16 November, he described the institute as central to a distinctly British approach built on government expertise, targeted regulation and international cooperation. He also said he wants AISI's model-testing work to be seen as a resource that benefits Americans as well as Britons.
The wider pitch to the US
Axios reported that the visit included meetings with US government counterparts, AI start-ups and the leading AI labs, plus a talk at Stanford. Narayan told the outlet that Britain wants a larger share of AI hardware and manufacturing, both for British jobs and to make US and UK supply chains more resilient.
He also said he wants Britain to help develop powerful open-weight AI models with Western allies, while improving the science of keeping such systems secure. Looking ahead, the UK plans to use its 2027 G20 presidency to push for international AI standards, including making companies more responsible for risks they find in their own systems.
Not everyone agrees
The comments land while ministers are under pressure to reduce Britain's reliance on American tech firms. City AM recalled that last month the trade body UKAI, which represents more than 270 businesses, urged the government to set out a full sovereign AI strategy and warned that Britain risked becoming dependent on foreign technology providers. Its chief executive argued that existing commitments, including the £500m Sovereign AI Fund, lacked a coherent plan for building capability at home.
The minister's position rests on one big condition: that access to allied models stays reliable. He spoke of wanting a form of insurance on that access, but neither report sets out what that insurance would look like in practice.
The UK angle
For UK businesses, the practical message is that the government is not about to launch a state-backed chatbot or model to rival the US labs. Most British firms will keep building on models made elsewhere, so questions about where data is processed, how models are tested and what happens if access is withdrawn stay firmly on the buyer's checklist.
It also sits alongside a separate debate in Parliament. This week a Commons committee called for a statutory AI regulator with power to halt unsafe models, a story we covered in UK MPs Call for a Statutory AI Watchdog. Together, the two stories show the UK trying to shape AI it largely does not build: through testing, standards and, possibly, law.
FAQs
Is the UK government building its own AI model?
No. The AI minister said a British frontier model is not a necessity as long as the UK has trusted access to allies' models, and that he does not plan to put large amounts of public capital into one. He said the government would still support British companies that try.
Where will UK public AI investment go instead?
The minister named chip design, compound semiconductors and scientific research as areas where Britain could build capabilities that other countries rely on.
Sources
- City AM: UK does not need its own frontier AI model, says AI minister (18 Nov 2026)
- Axios: Britain brings its AI pitch to Silicon Valley (16 Nov 2026)
- UK AI Security Institute: official site
Follow AI Decoded on Substack
Weekly plain-English AI news and verified UK events.
Subscribe