Semafor14 min
UK AI Minister Kanishka Narayan at Silicon Valley & The World | Semafor
Kanishka Narayan, Julie Pace, Julia
Interview
- The UK is currently the fastest-growing economy in the G7 (H1 this year), tracking closely with the US, driven by a boom in tech and AI sectors.
- European venture capital concentration has shifted dramatically toward the UK, which now receives 50% of European capital compared to 30–35% historically.
- Despite strong growth, the absolute scale of US and Chinese AI ecosystems remains "radically in a different league" from the UK.
- The UK strategy prioritizes maximizing value creation through specific sector bets rather than matching the sheer scale of US/China frontier models immediately.
- Targeted investment areas include:
- Next-generation inference chips via companies like Fractal and Olyx.
- Non-language frontier models for drug discovery (Isomorph), materials discovery (Caspia), and computer vision (Wave, DeepMind).
- AI adoption in life sciences, services, cybersecurity, and trust/safety infrastructure.
- UK officials do not view the necessity of building a domestic, large-scale frontier model as critical, provided the UK maintains trusted interdependence and access to global models.
- The AI Security Institute is positioned as the only lab outside the US with pre-deployment access to frontier models for risk assessment.
- A recent temporary loss of access to US models for the AI Security Institute was attributed to logistical and operational diligence rather than fundamental policy shifts.
- The UK government is investing £100 million in biosecurity alongside £100 million from Anthropic and OpenAI to create a joint biosecurity lab.
- UK AI security strategy focuses on "agency and control" (sovereignty) achieved through:
- Immediate access to GPUs, including the largest deployment in Europe last year.
- Diversifying compute supply across multiple vendors (e.g., Cerebras, SambaNova).
- Building domestic capacity in next-generation inference and photonics.
- Prime Minister Keir Starmer acknowledges that the current post-Brexit trade deal has "let the country down," citing friction for farmers and manufacturing.
- The government aims to walk a dual path: respecting the democratic judgment of Brexit while seeking deeper economic integration with the EU to restore market access and prosperity.
- The UK rejects a binary choice between innovation and security, opting instead for a "British model" that combines targeted regulation with significant state capacity building.
- Independent model testing is viewed as a public trust imperative, though the mechanism for mandatory testing (statutory vs. voluntary) remains under consideration.
- UK policy aims to prevent US-style public distrust by ensuring private data (especially healthcare) remains within sovereign UK infrastructure.
- AI infrastructure development is tied to local community engagement to ensure local jobs, supply chains, and a sense of voice in the build-out.
- The US midterm election results confirming public skepticism toward AI speed and scale are viewed as a "confirmatory moment" validating the UK's need to solve trust and technology simultaneously.