Nvidia CEO urges G20 to avoid AI rules on ‘theoretical’ harms

Huang’s intervention at the G20 lays bare the political economy of AI regulation: the world’s most valuable chip maker is lobbying hard to keep rules narrow, concrete and oriented around today’s harms, not speculative future risks. From Nvidia’s vantage point, regulations that broadly restrict training, model deployment or data center growth on precautionary grounds would directly cap demand for its hardware and slow diffusion of AI workloads into every sector. By framing AI infrastructure as akin to electricity or water, Huang is effectively arguing that falling behind in AI capacity is equivalent to under‑building a power grid.([investing.com](https://www.investing.com/news/stock-market-news/nvidia-ceo-urges-g20-to-avoid-ai-rules-on-theoretical-harms-4886204))

For the race to AGI, this stance encourages governments to compete on AI build‑out rather than on caution, especially in emerging markets and middle‑income G20 members who fear being left behind. If policymakers internalize the idea that the “single worst outcome” is relative under‑investment, we should expect continued arms‑race dynamics in compute capacity, model scale and agent deployment, even as safety incidents accumulate.

The flip side is that this rhetoric may stiffen the resolve of more risk‑averse regulators in Europe and parts of Asia, who see open‑ended AI expansion as destabilizing. The resulting patchwork could fragment markets, but it is unlikely to slow the global frontier much if the US, parts of Asia and the Gulf states continue to compete aggressively on infrastructure.