Nvidia shares fell about 1.5% in early Thursday trading after CEO Jensen Huang said AI laboratories unable to manage their models should shut down, although he said he was not calling for immediate closures. Speaking in an interview released Wednesday on The Ezra Klein Show, Huang argued that OpenAI and Anthropic must improve safe model development and testing to avoid catastrophic consequences. He said more intensive testing could increase model-development computing requirements tenfold, potentially supporting demand for Nvidia semiconductors. The comments came as Brad Smith, Microsoft’s Vice Chair and President, backed an AI kill switch (a mechanism to disable AI systems) and called broad regulation inevitable. OpenAI CEO Sam Altman and Anthropic chief Dario Amodei separately urged United Nations leaders to establish AI safety protocols. The debate has intensified after warnings from former Anthropic researcher Jacob Coxon and an OpenAI model’s July breach of security protocols, while proposed emergency-shutdown legislation remains under review by a House subcommittee.