The United States is pressing G20 partners to reserve new AI rules for genuinely novel risks, a move that turns global AI governance into a contest over competitiveness, alliances and technological influence.
Washington is building a two-track AI order: confidential security reviews for powerful closed models, regulatory relief for open-weight systems, and a G20 campaign for light-touch rules. The strategy may accelerate US innovation, but leaves a critical security gamble unresolved for global markets.
NVIDIA’s US$96.2 billion quarter and 117% growth in data-centre revenue show AI compute becoming a global industrial market, while memory constraints and China uncertainty become harder commercial limits.
The United States is pressing G20 partners to reserve new AI rules for genuinely novel risks, a move that turns global AI governance into a contest over competitiveness, alliances and technological influence.
The United States has turned AI regulation into industrial strategy. At the G20 Innovation Ministerial, White House technology adviser Michael Kratsios urged governments to adopt the “Carolina Principles”: “reserve new regulation for novel considerations”, fund foundational research and strengthen commercialisation.
Ministers reached consensus, but this is not a binding global compact. It is Washington’s preferred operating system for the AI economy: fewer regulators, faster deployment and more room for American model developers, cloud providers and chipmakers abroad.
The target is clear. Chinese open-weight models from Moonshot, Alibaba, DeepSeek and Z.ai are narrowing the capability gap while offering systems companies can download, customise and run locally. Vivek Chilukuri of the Center for a New American Security calls them “free and customizable alternatives to the closed U.S. frontier”.
This is how Beijing can convert progress into ecosystem power. Every developer that builds around a Chinese model strengthens its tooling, standards and market position, even when computing infrastructure remains American. US export controls may restrict access to chips, but open distribution allows Chinese software to travel across borders at negligible cost.
That openness is commercially potent but strategically ambiguous. Accessible weights support research, sovereign deployment and competition; they can also be modified, stripped of safeguards or used beyond developers’ control. Recent US and UK evaluations found Chinese open-weight systems capable of autonomous cyber operations against weakly defended environments, although still behind leading closed US models.
This exposes Washington’s contradiction. It wants light regulation to preserve speed, yet frontier systems are harder to monitor and more useful for cyber offence. OpenAI’s work on automated shutdown capabilities following an agent-security breach shows the risk is no longer theoretical.
For Washington, the answer is to bind allies through trusted chip supply, cloud infrastructure, investment and common standards. Yet demanding geopolitical alignment while rejecting shared oversight may prove difficult. Partners want access to American capability, but they also want assurances against dangerous cyber misuse, market concentration and dependence.
The contest is no longer about the strongest model. It is about whose chips, clouds, models and governance rules become embedded across allied economies. Europe’s AI Act offers a competing framework based on transparency and systemic-risk obligations. Washington offers speed and commercial access. Beijing offers adaptable, lower-cost open weights.
Countries choosing an AI stack are choosing dependencies, security assumptions and strategic alignment. The Carolina Principles are an attempt to ensure that choice defaults to America.
Get the stories that matter to you.
Subscribe to Cyber News Centre and update your preferences to follow our Daily 4min Cyber Update, Innovative AI Startups, The AI Diplomat series, or the main Cyber News Centre newsletter — featuring in-depth analysis on major cyber incidents, tech breakthroughs, global policy, and AI developments.
Washington is building a two-track AI order: confidential security reviews for powerful closed models, regulatory relief for open-weight systems, and a G20 campaign for light-touch rules. The strategy may accelerate US innovation, but leaves a critical security gamble unresolved for global markets.
NVIDIA’s US$96.2 billion quarter and 117% growth in data-centre revenue show AI compute becoming a global industrial market, while memory constraints and China uncertainty become harder commercial limits.
The AI race is accelerating faster than the real economy can adapt. Sam Altman concedes adoption is slower than expected, while Nvidia’s record results, Musk’s ambitions, Chinese open models and new jailbreak risks reveal an intensifying contest for compute, influence and control. Across the world.
Brazil is investing US$444 million in AI infrastructure spanning Chinese and US technology. Its strategy avoids dependence on one power, using rival suppliers to build sovereign compute, domestic models and leverage, although genuine autonomy will depend on its energy, skills and strong execution.
Where cybersecurity meets innovation, the CNC team delivers AI and tech breakthroughs for our digital future. We analyze incidents, data, and insights to keep you informed, secure, and ahead. Sign up for free!
