Nvidia CEO Jensen Huang has rejected predictions that artificial intelligence could destroy humanity within the next several years, saying there is a “0% chance” of such an outcome by 2030, according to The Guardian.
Huang described extinction warnings as unsupported “doomsday narratives” that could unnecessarily frighten the public. His position places Nvidia on the opposite side of a growing debate involving executives and researchers at several leading AI laboratories.
Anthropic CEO Dario Amodei, OpenAI CEO Sam Altman and Elon Musk have supported calls for greater caution or slower development following warnings about cybersecurity threats, biological risks and the potential loss of human control over advanced systems.
Huang argued that existing laws covering unauthorized computer access, product liability, contracts and damages should be applied before governments create entirely new regulatory systems.
The debate has significant commercial implications for Nvidia, which supplies the graphics processors and computing systems used to train and operate many leading AI models. Slower development or tighter regulations could reduce the pace of data-center investment and affect demand for Nvidia’s products.
At the same time, major safety failures could undermine confidence in AI and create even stronger regulatory pressure. Nvidia therefore has an interest in supporting responsible deployment while resisting policies that could limit infrastructure growth.
Huang’s remarks also reflect disagreements over whether AI risks should be treated as extensions of familiar technology problems or as challenges requiring entirely new oversight institutions.
Governments are already evaluating rules covering model evaluations, cybersecurity, transparency and the use of AI in critical infrastructure. The United States and China have also discussed establishing an international notification system for potentially dangerous incidents.
The central question is whether existing legal frameworks can address the speed and scale of advanced AI. Huang believes they can, while critics argue that waiting for harm to occur may be insufficient when increasingly autonomous systems are involved.
