Co-hosted by US Commerce Secretary Howard Lutnick and White House Office of Science and Technology Policy Director Michael Kratsios, the summit brought together representatives from major economies including Canada, Germany, Brazil and China to tackle AI, infrastructure and governance questions at a pivotal moment for the sector.

The Trump administration pushed for adoption of the "Carolina Principles", under which participating nations commit to avoiding unnecessary new AI regulations, reserving new rules only for "novel" situations that existing laws cannot address. Lutnick argued that current legal frameworks are already sufficient to manage emerging AI risks, and that additional constraints would only slow progress.

Day one featured video addresses from three of the industry's most influential leaders. Google DeepMind CEO Demis Hassabis urged G20 officials to establish standardised safety evaluations for advanced AI systems, while Meta CEO Mark Zuckerberg cautioned against restricting open-weight models as Chinese-developed open-weight systems gain traction among US enterprises, raising domestic security concerns.

SpaceX CEO Elon Musk criticised the European Union's regulatory approach as inhibiting progress, and urged non-Chinese leaders to rapidly scale alternative energy generation to meet the surging power demands of next-generation data centres.

NVIDIA CEO Jensen Huang said favourable US regulations create strong conditions for AI infrastructure expansion, though he noted that oft-cited employment gains from data centres tend to be inflated — once construction ends, permanent staffing needs are low.

Palantir CEO Alex Karp struck a more cautious tone, warning the AI transition carries real economic risk and "huge dangers," pointing to concerns over wealth inequality. As reported by CNBC, Karp said the growing wealth gap "leaves a bad taste in people's mouth."

Key Takeaways

• →The "Carolina Principles" commit G20 nations to avoiding unnecessary AI regulation, reserving new rules for novel risks existing laws can't address.
• →Open-weight AI models sparked debate, balancing open-source innovation against national security risks tied to foreign-developed systems.
• →Leaders stressed that scaling data centres requires urgent investment in clean energy generation and grid capacity.
• →Executives including Palantir's Alex Karp flagged wealth concentration and workforce disruption as real risks of rapid AI adoption.
• →Global compute capacity was framed as essential to national economic competitiveness heading into the next decade.
• →A non-binding pledge from tech firms to absorb data centre energy costs left local communities with protests and no enforcement mechanism.