China’s bid to lead the global AI order

Artificial intelligence is entering a new phase of geopolitical competition. The contest is no longer confined to who possesses the most advanced chips, the greatest computing capacity or the most powerful models. It now extends to the institutions, standards and partnerships that will determine how the technology is governed and who can access it.

Chinese President Xi Jinping used the World Artificial Intelligence Conference, held in Shanghai from July 17 to 20, to argue that China should help lead that next phase. Speaking before government officials, international organizations and technology companies including Alibaba and Huawei, Xi promoted open-source AI, greater international cooperation and wider access for developing countries. One day earlier, representatives from 29 countries had signed an agreement establishing the World Artificial Intelligence Cooperation Organization, or WAICO, which Beijing says will coordinate AI governance, standards and international cooperation. The announcement marked China's most ambitious effort yet to position itself as a leading architect of an alternative international AI order.

From principles to an institution

China's campaign to influence global AI governance did not begin in Shanghai. Over several years, Beijing has advanced an alternative vision through policy frameworks, diplomacy and capacity-building initiatives.

That effort began with a position paper on ethical AI governance in November 2022 before taking a more ambitious turn with the Global AI Governance Initiative in October 2023. The initiative called for international rules based on broad consensus, greater representation for developing countries and discussions within the United Nations over the creation of an international AI governance institution.

Beijing built on that foundation in 2024 and 2025 with the Shanghai Declaration on Global AI Governance, the AI Capacity-Building Action Plan for Good and for All, the Global AI Governance Action Plan and the AI Plus International Cooperation Initiative. WAICO represents the next step in that progression, establishing an institution to advance the principles Beijing has spent several years promoting. Beijing says WAICO will complement the United Nations framework while providing China with a dedicated institution to advance its vision of global AI governance.

An appeal built around access

Xi framed China's approach around access rather than restriction. He urged governments to promote open-source development and warned that unequal access to artificial intelligence could create "new historical injustice." The message presents AI governance not only as a security issue, but as a development challenge.

China backed that message with concrete commitments. Over the next five years, Xi said Beijing would provide developing countries with 5,000 places in AI training and seminar programs. It would establish application cooperation centers with ASEAN, the Arab League, the African Union, the Community of Latin American and Caribbean States, the Shanghai Cooperation Organization and BRICS, while helping 30 countries deploy MAZU, an AI-powered meteorological warning system.

For governments lacking frontier AI capabilities, those commitments position China as a provider of technology, expertise and infrastructure.

Competing visions with shared contradictions

China's message also reflects a changing balance in the AI race. According to the Stanford AI Index Report 2026, U.S.-based institutions produced 59 notable AI models in 2025 compared with 35 from China. Although the United States remains the leading developer of frontier models, the performance gap has narrowed substantially, with Anthropic's leading model outperforming China's strongest model by just 2.7% in March 2026. As technological differences narrow, competition is shifting toward governance, access and international partnerships.

Xi did not mention the United States directly, but the contrast was unmistakable. He criticized the overextension of national security concepts and argued that advanced technology should not be controlled by a limited number of countries. The remarks stood in contrast to a U.S. strategy that has relied heavily on export controls, trusted technology partnerships and restrictions on advanced semiconductor exports to China.

U.S. restrictions have extended beyond hardware. Earlier this year, the Commerce Department temporarily limited overseas access to Anthropic's latest Claude models before later easing the restrictions, illustrating Washington's growing willingness to treat frontier AI models themselves as strategically sensitive technologies.

The divide, however, should not be reduced to a simple contest between an open China and a restrictive United States. Washington argues its export controls target narrowly defined security risks while American companies continue to release widely used models and developer tools. China, meanwhile, maintains extensive domestic controls over information and online content, while Reuters reported authorities were considering limits on overseas access to some leading Chinese models despite Xi's promotion of open-source AI.

Nor has China's vision been universally embraced. A July 2026 analysis by the Mercator Institute for China Studies argued that Beijing is seeking a larger role in defining international governance while pursuing an order distinct from the liberal system led by the United States. For many governments, concerns surrounding data governance, political influence and technological dependence remain an important counterweight to China's offer of greater AI access.

The race to define the system

China’s proposal also arrives as international AI governance remains fragmented. The 2026 Global Index on Responsible AI found that 126 countries had adopted at least one AI policy or initiative, yet 78% of governance frameworks across the Global South remain nonbinding, highlighting the absence of a coherent global system.

That weakness creates an opening for countries capable of offering institutions, financing, training and technology at the same time. Beijing has spent several years preparing to occupy that space. The Shanghai announcement represents its most formal step yet, connecting Chinese technological cooperation with a permanent international body and a network of partnerships across the Global South.

The AI race will still be influenced by chips, computing power and scientific breakthroughs. But as frontier models converge in capability, another competition is emerging. The country that persuades others to adopt its standards, institutions and vision for governing artificial intelligence may ultimately wield influence extending far beyond the technology itself.