Xi Jinping did not ask for permission. At the World AI Conference in Shanghai, he announced that China will offer 5,000 AI training slots to Global South countries over the next five years. A day earlier, 29 nations formally established the World Artificial Intelligence Cooperation Organization (WAICO), headquartered in Shanghai. The founding members include Russia, Brazil, South Africa, Pakistan, and Indonesia. Not a single Western country signed on. This is not a side event in the AI race. It is the clearest signal yet that China wants a separate rulebook for the most strategically important technology of the century.
The move is more than diplomatic theater. It is a structural challenge to the American-led AI order that has dominated since the launch of ChatGPT. The United States has used export controls, sanctions, and investment restrictions to keep advanced AI chips and models out of Chinese hands. China’s response is to build its own coalition of nations that either cannot access American technology or do not want to live under Washington’s rules. WAICO is the institutional wrapper around that strategy.
What WAICO Actually Does
On paper, the organization promotes cooperation, training, and shared governance. Xi’s 5,000 training slots are a practical down payment. China will bring government officials, engineers, and policymakers from the Global South to Shanghai to learn how to build, regulate, and deploy AI systems. The implicit offer is clear: we will teach you the technology, we will share the standards, and we will not attach human-rights or national-security lectures the way Washington does.
The headquarters in Shanghai matters. It gives China control over the agenda, the funding, and the staffing. International organizations often look neutral on paper but operate according to the country that hosts them. By anchoring WAICO in China, Beijing ensures that its priorities — data sovereignty, state-led development, and non-interference in domestic AI regulation — become the default norms of the bloc.
Xi also used the moment to push a familiar message: AI must remain under human control. He criticized the use of “overly broad national security” justifications in AI policy, a direct reference to US export controls on advanced semiconductors. It is a clever framing. China presents itself as the defender of open technological development while the United States is portrayed as the one hoarding chips and restricting access. For many developing countries, that argument lands.
The Economic Muscle Behind the Pitch
Xi did not come empty-handed. He said China’s “Smart Economy,” which spans AI and other digital technologies, is now worth over one trillion renminbi, roughly $140 billion. That is a fraction of the global AI market, but it is enough to underwrite scholarships, training programs, infrastructure projects, and joint research centers across dozens of countries. The model is classic development finance with a technological edge: China offers capacity-building today in exchange for market access, standards alignment, and diplomatic loyalty tomorrow.
For the Global South, the offer is tempting. Many of these countries are struggling to build domestic AI ecosystems. They lack compute, talent, and capital. The United States offers them cloud APIs, venture capital, and occasional partnership announcements. China offers them the possibility of becoming part of an alternative production network. That does not mean every member of WAICO will become a Chinese client state, but it does mean they will increasingly default to Chinese hardware, Chinese models, and Chinese governance frameworks.
A Global South Split, Not a Monolith
The membership list is the most interesting part of the announcement. Russia is a predictable partner, already aligned against the Western-led order. Brazil, South Africa, Pakistan, and Indonesia are more significant. These are large economies with their own technological ambitions and younger populations that will consume AI services for decades. If China can convince them that its stack is reliable, affordable, and politically non-judgmental, it gains enormous leverage.
But the Global South is not a single bloc. India, for example, stayed out. So did most of Southeast Asia outside Indonesia. African representation was limited. The Middle East was notably absent. WAICO is not yet a comprehensive alternative to the American-led order. It is a nucleus. The question is whether China can expand it before the United States and Europe offer a more credible alternative than export controls and moral lectures.
What This Means for the West
The American strategy has been to slow China down by cutting off its access to the most advanced chips. That has worked in some ways. Chinese labs have been forced to become more efficient and to build domestic alternatives. But it has also created a market for a parallel AI ecosystem. Countries that fear being cut off from American technology — or that resent the conditions attached to it — now have a place to go.
Europe faces a different dilemma. The EU AI Act is built on transparency, risk classification, and fundamental rights. Those principles are hard to export to a bloc that sees regulation as a tool of sovereignty rather than a constraint on state power. If WAICO adopts Chinese-style data governance and content controls, the transatlantic alliance will find it even harder to build a single global standard. The world could end up with two incompatible AI regimes: one built around American and European norms, the other built around Chinese ones.
For the private sector, the split is expensive. AI companies that want to operate globally may have to build two versions of their products, two compliance stacks, and two partner ecosystems. That is a nightmare for startups and a windfall for consultancies. It is also a gift to domestic champions in each bloc, who can use regulatory barriers to keep foreign competitors out.
🔥 Hot Takes
1. US export controls are the best recruiting tool China ever had. Washington convinced itself that starving China of chips would slow its AI development. Instead, it gave Beijing a perfect grievance to take on the road. Every country that has ever been denied technology by the West is now being told: China will share. That narrative is worth more than any training budget.
2. WAICO is not about ethics; it is about locking in dependencies. The “human control” language sounds reasonable, but the real product is a standards ecosystem. Once a country’s regulators, engineers, and infrastructure are trained on Chinese frameworks, switching costs explode. Beijing is not giving away AI education. It is building a customer base.
3. The West still thinks this is a technology race. China is playing a civilization game. While American politicians argue about which model benchmark is ahead this week, China is building institutions, training foreign officials, and writing the norms that will outlast any single model. The model leaderboard matters today. The rulebook matters for decades.
The Bottom Line
WAICO will not replace the Western-led AI order overnight. It does not have to. Its purpose is to create a credible alternative that can absorb countries that feel frozen out of the American ecosystem. China is betting that if enough of the world uses its models, its chips, and its governance templates, the United States will eventually have to negotiate rather than dictate.
The next phase of the AI race is not about who has the best model. It is about who writes the rules under which everyone else builds. China just opened a new venue for that fight. The West can either fill it with better offers, or it can watch the room fill up without it.