China’s WAICO Offers Global South AI Access, Challenging Western Dominance

China has formally entered the contest to shape global AI governance with the launch of the World Artificial Intelligence Cooperation Organisation (WAICO), a 29-nation body headquartered in Shanghai that positions Beijing as a champion of the Global South. Announced by President Xi Jinping at the ninth World Artificial Intelligence Conference (WAIC) in Shanghai on July 17, 2026, WAICO frames AI as a “public good” and offers developing countries training, hardware, and open-weight models as an alternative to what it describes as the exclusionary, profit-driven approach of the US and its allies.

For executives, founders, and policymakers worldwide, the initiative signals a bifurcating AI ecosystem. Nations from Brazil to Kazakhstan now have a state-backed path to AI infrastructure that bypasses US export controls, but the move also raises questions about long-term dependency, strategic restrictions, and whether China’s vision of “equitable” access aligns with the diverse needs of the Global South.

A new bloc for AI governance and capacity-building

WAICO’s 29 founding members—including Brazil, Russia, Pakistan, Indonesia, Kazakhstan, and Laos—signed the agreement on July 16, 2026, establishing a non-governmental organisation that China presents as a counterweight to Western-dominated AI frameworks. Xi’s keynote speech framed the body as a “symphony of international cooperation” rather than a “solo performance,” a veiled critique of US and EU policies that prioritise frontier model controls, national security, and corporate profitability.

Concrete commitments underpin the rhetoric. Xi pledged that China will provide 5,000 training and seminar opportunities over the next five years for professionals from developing nations, alongside new AI application cooperation centres with ASEAN and BRICS. The capacity-building promise is central to WAICO’s appeal, offering a tangible on-ramp for countries often sidelined in global tech discussions. For instance, the centres are expected to focus on adapting AI tools for local needs in agriculture, healthcare, and public administration—sectors where Western providers have historically shown less engagement.

“China’s focus on public-good applications, rather than subscription chatbots, strengthens its global appeal in the Global South,” said Gerui Wang, a scholar at Lingnan University, in a recent South China Morning Post opinion piece. This emphasis on infrastructure and public services—rather than enterprise software—resonates with governments seeking to embed AI in healthcare, agriculture, and education, where cost and accessibility are critical barriers.

Hardware and open models as geopolitical leverage

WAICO’s launch coincided with two major technological announcements that underscore China’s bid for AI autonomy. At the conference, Huawei unveiled the Atlas 950 SuperPoD, which the company claims delivers 6.7 times the computing power of Nvidia’s NVL144 system. For nations facing US export restrictions, the Atlas 950 offers a domestic alternative for training and inference. Russia’s Sberbank, for example, has already turned to Huawei hardware amid persistent Nvidia shortages, highlighting how sanctions and supply chain disruptions are driving demand for Chinese alternatives.

Meanwhile, Beijing-based Moonshot AI released Kimi K3, a 2.8-trillion-parameter open-weight model that independent benchmarks score at 57.1 on the Artificial Analysis Intelligence Index v4.1, slightly behind Anthropic’s Claude Fable 5 (59.9) but reportedly stronger in specific coding tasks. Moonshot has committed to fully open-sourcing Kimi K3’s weights by late July 2026, a move that aligns with WAICO’s narrative of accessibility. For developers in the Global South, open-weight models lower the barrier to entry, enabling fine-tuning for local languages and use cases without reliance on proprietary US systems. The model’s performance in coding arenas could be particularly valuable for countries looking to build local tech industries.

Yet the openness has limits. A Reuters report on July 7, 2026, revealed that Beijing has held meetings with domestic tech firms about potentially restricting overseas access to its most advanced models, suggesting that strategic controls may accompany the “public good” rhetoric. George Chen, a senior analyst at The Asia Group, noted that China is leveraging uncertainty around US technology policy—such as fluctuating export controls on frontier models—to present itself as a steadier partner for Belt and Road nations. “Beijing is using AI diplomacy to rally its allies,” Chen said. “The message is clear: if you want stability and access, China is here.”

Competing visions of AI as a public good

WAICO’s founding reflects a broader philosophical divide in global AI governance. While the US and EU focus on national security, frontier model oversight, and corporate accountability, China’s pitch centres on affordability, technological autonomy, and open-source access. Xi’s vision frames AI as a tool to bridge digital divides and prevent “new historical injustice,” a narrative that resonates with many developing nations frustrated by the concentration of AI power in the West.

However, the Ada Lovelace Institute’s 2025 report, Making Good, highlights that definitions of “public good” are not monolithic. Diverse communities often define it holistically—prioritising fairness, life improvement, and local relevance—rather than just economic growth, a nuance that China’s state-centric model may overlook. For example, while China’s training programs and open models may provide technical access, critics argue that top-down governance could limit local adaptation and community-driven innovation.

Critics also warn that joining WAICO could complicate nations' access to US-made models and proprietary technology. The risk of dependency is real: countries adopting Chinese hardware and models may find themselves locked into Beijing’s standards for infrastructure, chips, and governance frameworks. Yet for many in the Global South, the alternative—relying on US technology subject to sudden export controls—is equally unappealing. The choice between the two systems increasingly forces nations to weigh immediate accessibility against long-term strategic alignment.

The fragmentation of the global AI landscape

WAICO’s launch is the most explicit challenge yet to the US-led AI order, but its long-term impact remains uncertain. The organisation’s success will depend on whether its promises—training, hardware, and open models—translate into tangible benefits for member states without hidden strings. For now, the initiative offers the Global South a seat at the table, but the menu may be limited by Beijing’s own strategic interests, including potential restrictions on its most advanced technologies.

For professionals, the immediate takeaway is the acceleration of a bifurcated AI market. Developing nations now have a viable alternative to US technology stacks, but the choice between Washington and Beijing is increasingly a choice between competing visions of governance, access, and control. As open-weight models like Kimi K3 and hardware like the Atlas 950 SuperPoD proliferate, the global AI landscape will likely fragment further, with standards, regulations, and supply chains diverging along geopolitical lines.

The question for executives and policymakers is whether this fragmentation leads to greater equity—or simply a new form of dependency. As Gerui Wang noted, China’s emphasis on public-good applications could democratise AI, but only if the Global South retains the agency to define what “public good” means for its own communities. The coming years will reveal whether WAICO can deliver on its promises or if its members will find themselves navigating a new set of constraints in an increasingly divided AI world.