G20 tech leaders debate light touch AI governance and power needs

This G20 side meeting is revealing because it shows how tech elites are trying to shape the global narrative on AI regulation and infrastructure. Musk and Sacks are arguing for a ‘Carolina Principles’ style regime that avoids FDA like pre-approval of AI systems, favors local decision-making on data centers and treats AI as a new industrial revolution requiring trillions in compute and power investment. That is a clear pushback against the more restrictive instincts of the EU AI Act and some civil society groups.

Their GDP and robot projections are aggressive, but the direction of travel is clear: they want governments to see AI and humanoid robotics as central to economic competitiveness. That framing makes it easier to justify subsidies for data centers, new transmission lines and perhaps even fusion research, while casting heavier regulation as a drag on national prosperity.

From a race-to-AGI standpoint, this style of lobbying matters because it influences whether coordination on slowing frontier development is politically feasible. If G20 economies internalize the idea that being first in AI means a 20–30 percent GDP bump, they will be less inclined to accept hard compute caps or strict licensing. The debate captured in this article is an early glimpse of how that tension will play out at the highest political levels.