NVIDIA’s Jensen Huang to G20: Regulate AI’s real harms, not hypothetical fears

At a G20 innovation gathering in Chapel Hill, North Carolina on September 2, NVIDIA CEO Jensen Huang urged governments not to build sweeping AI rules around “theoretical harms.” Speaking as the head of the company whose GPUs power much of today’s AI boom, Huang called for regulators to focus on problems that are already manifesting in the real world.

According to his remarks, issues like misinformation, privacy violations and misuse of automated decision systems deserve targeted regulation and enforcement. But he warned that creating new bureaucracies or overly broad pre‑approval regimes based on worst‑case future scenarios could slow deployment of useful systems and undercut countries’ competitiveness in AI hardware and software.

Huang’s comments align with the current US push at the G20 to promote principle‑based, flexible governance instead of rigid ex‑ante licensing. That stance contrasts with the EU’s detailed AI Act and China’s tight content and model controls, deepening a global debate over how far prior regulation should go. For infrastructure providers such as NVIDIA, the outcome will significantly shape where AI data centers are built and how quickly new models can reach users.