Jensen Huang Demands Rigorous Containment for Agentic AI

Nvidia CEO Jensen Huang warned on November 12, 2026, that agentic AI systems must be placed in rigorous containment and monitoring systems to mitigate high risks. He asserted that developers should not test or deploy AI agents if they lack the ability to monitor them.

Huang specifically directed this warning toward AI developers including OpenAI and Anthropic, stating that large-scale experiments should stop if containment and monitoring capabilities are not in place. He argued that experiments must be contained rather than shutting down AI labs entirely.

Addressing critics who claim such containment is impossible, Huang dismissed those assertions as "made up nonsense." He maintained that the development of agent ecosystems requires strict oversight to ensure safety.