NVIDIA CHIEF URGES UK USERS TO ADOPT OPEN AI MODELS FOR CRITICAL WORK

Nvidia chief executive Jensen Huang urged British businesses, government agencies and individuals to use open artificial intelligence models for critical work during a London visit on Wednesday. He said proprietary systems from Anthropic and OpenAI could leave users without sufficient control over sensitive tasks. “You simply cannot afford to outsource that. In that case, you need to have open models,” Huang said. He cited trust, regulation and sovereignty concerns, while saying both types of model had uses.

Advocates of open models, including systems from Meta and Mistral, argue that running and adapting them locally gives users greater privacy and control. Large AI laboratories argue that open models lack the safety safeguards their products provide. Nvidia has also pledged $250m in computing resources for British research into autonomous AI agents, comprising $50m annually over five years. The company said UK-based data centres would provide resources for Oxford, University College London and Imperial College London, rather than cash grants.

Forrester principal analyst Indranil Bandyopadhyay warned that reliance on Nvidia’s technology could influence the safety standards researchers develop. “Researchers working mainly on one vendor’s tooling may produce benchmarks that reflect that stack”, he said, urging the AI Security Institute and DSIT to require open publication and align research with national evaluation priorities. IG chief market analyst Chris Beauchamp described Nvidia’s announcements as a “co-ordinated bet on the UK as a firm AI hub”. Nvidia also announced plans to take minority stakes in five British startups — Boltz, Foresight, Hiverge, Prima Mente and Riverlane — following investments in four startups in France.