Researchers Critique NVIDIA Open AI Support as Regulatory Capture

AI safety experts including Dylan Hadfield-Menell criticize NVIDIA for minimal safety spending despite its leading position in AI hardware. They describe the company's open-weights advocacy and recent letter as regulatory capture and PR moves aimed at profit maximization rather than genuine risk management. Replies call for NVIDIA to instead fund large defensive inference clusters, safe pre-training research, and broader open-source safety ecosystems. Observers note a perceived double standard where proprietary model firms face capture accusations but hardware sellers receive praise for similar coordination. The exchanges highlight skepticism toward corporate motives in open AI development.