US vs China AI race enters new phase: Why battle over AI rules may matter more than models

Gates told Reuters that he is seeking a meeting with Chinese President Xi Jinping later this year to discuss international rules for potentially dangerous AI models. His concerns extend beyond the race to build increasingly capable systems. He wants governments to establish safeguards for risks that could cross national borders.

Gates has also warned about AI’s potential to design molecules that could facilitate biological attacks. He has proposed an international body for AI safety that could draw lessons from nuclear inspections, aviation safety and international environmental agreements, according to Reuters.

The push comes as the US-China AI competition expands beyond who can build the most powerful models. Both countries are trying to shape the rules, technology supply chains, standards and international partnerships that will influence how AI develops globally.

How are the US and China building competing AI ecosystems?

China has backed the creation of an international organisation focused on AI cooperation and governance. 

On July 16, representatives from 29 countries signed an agreement establishing the World Artificial Intelligence Cooperation Organization (WAICO), headquartered in Shanghai. The organisation says it aims to promote international cooperation and global AI governance and support AI that is “beneficial, safe and fair”.

China said the organisation would operate in line with the principles of the UN Charter. Its proposed areas of cooperation include AI development strategies, governance rules and technical standards.

The founding group includes Russia, Pakistan, Indonesia, Kazakhstan, Brazil, Venezuela, Ethiopia and Cameroon.

WAICO gives Beijing a platform to promote its approach to international AI cooperation, particularly among developing economies.

The US has taken a different approach through Pax Silica, an initiative focused on the technology and supply chains needed for advanced AI. The US State Department describes it as covering critical minerals, energy, advanced manufacturing, semiconductors and AI infrastructure.

The two initiatives are therefore not direct equivalents. WAICO is focused on international AI cooperation and governance, while Pax Silica is centred more heavily on technology, infrastructure and supply chains.

Are countries being pushed to choose between US and China?

Reuters reported on August 14 that Washington was preparing to tell dozens of countries they could not simultaneously participate in the US-led Pax Silica framework and China’s competing AI framework.

The US position reflects concerns in Washington that countries participating in its technology and supply-chain network should remain trusted partners as competition with China intensifies.

Pax Silica includes countries such as Japan, Australia, South Korea and Kazakhstan. Kazakhstan has also joined WAICO, illustrating the challenge for countries seeking economic and technological ties with both powers.

Southeast Asia offers another example. The South China Morning Post reported in August that both Washington and Beijing were seeking greater support from Southeast Asian countries for their respective AI frameworks.

For these countries, the choice could become increasingly difficult if the two systems become more exclusive. Governments want access to investment, technology and markets from both powers, while also retaining flexibility over their own AI policies.

The competition is therefore moving beyond AI models to infrastructure, supply chains, standards and governance.

Where does India fit into AI divide?

India has already joined the US-led Pax Silica framework. It became a member in February 2026, with the Indian government describing the initiative as a strategic partnership covering critical minerals, semiconductors, AI systems and the infrastructure required to deploy advanced technologies.

The move fits India’s efforts to expand domestic semiconductor production, computing capacity and advanced technology capabilities.

The Indian Council of World Affairs wrote on August 13 that participation in Pax Silica could provide India with access to computing resources, supply-chain development, investment and greater geopolitical leverage.

But India’s position is more nuanced than simply choosing between Washington and Beijing. India has economic and strategic relationships with both countries and has an interest in influencing global AI rules rather than simply adopting standards developed elsewhere.

That makes the evolution of WAICO and Pax Silica important for India. If the global AI system remains relatively open, India could potentially work with different countries on different parts of the technology ecosystem.

If competing US- and China-led systems develop into more exclusive blocs, however, India’s room for manoeuvre could narrow.

Why could AI rules become as important as AI models?

The race to build more capable AI models gets most of the attention, but governments are also deciding how those systems should be developed and used.

That includes questions around AI safety, biological and cybersecurity risks, technical standards, access to computing infrastructure and critical components.

Gates’ proposal for international cooperation on dangerous AI shows one side of this debate. His argument is that some risks created by advanced AI could cross borders and therefore require mechanisms that individual countries cannot manage alone.

At the same time, the US sees semiconductors, computing infrastructure and critical minerals as strategic assets, while China is seeking a larger role in international technology governance and partnerships with developing countries.

This makes AI governance increasingly difficult to separate from trade, national security and technology supply chains.

The creation of WAICO and the expansion of Pax Silica show how the AI competition is developing into a broader contest over the infrastructure and rules surrounding the technology.

The US and China may have a shared interest in preventing dangerous uses of advanced AI. But they are also competing for technological leadership and influence over the countries that will build, regulate and use these systems.