CNBC's The China Connection newsletter: Money or power? The key to winning the AI race

The United States is set to inform several countries that they must select between Washington's and Beijing's competing AI frameworks, according to Reuters. A US official and an internal State Department draft revealed that Washington aims to exclude countries joining China's AI initiative from a US-led coalition if they participate in Beijing's rival framework.

This comes as US and Chinese tensions escalate over AI, semiconductors, and critical minerals needed for advanced AI technology. The US launched Pax Silica last year to bolster supply chains for AI models, semiconductor chips, and critical minerals, aiming to lessen dependence on China and other competitors. Around 24 countries, including Japan, Australia, and South Korea, have joined Pax Silica, while Kazakhstan, a potential source of critical minerals, has joined both Pax Silica and China's rival AI initiative.

The draft letter, addressed to 35 signatories of a US AI Opportunity Statement, warns that countries cannot simultaneously participate in both frameworks. It urges countries to make a deliberate choice and emphasizes that membership in Pax Silica cannot coexist with duplicative initiatives with conflicting objectives. Kazakhstan, the only known country participating in both initiatives, has raised concerns for US officials due to its significant reserves of critical minerals.

The US push follows China's launch of the World Artificial Intelligence Cooperation Organization, a rival framework aimed at promoting Chinese AI technology. China's embassy in Washington criticized the US approach, stating that such actions will stifle global AI advancements and serve no one's interests. The US-China AI competition extends beyond software and AI models to the supply chains supporting the technology, with China currently dominating critical-mineral supply chains.

The rivalry intensifies as AI capabilities advance rapidly, and China considers restricting overseas access to some of its leading AI models.