China Convenes 29 Nations for AI Pact, Xi Sets Out His Terms

The World AI Conference in Shanghai opened on 18 July with a move that signals Beijing’s ambition to build a parallel global architecture for artificial intelligence. Before the event formally began, representatives of 29 countries – including Russia – signed an agreement to create the World AI Cooperation Organization (WAICO). The new body is designed to “ensure fair mechanisms for cooperation among states to build their digital capacity and develop their own technologies and solutions,” and will treat the United Nations as the primary platform for global AI discussion.

In his keynote, President Xi Jinping framed the moment as a historic inflection point. He traced the concept of AI back to the 1956 Dartmouth workshop in the United States, then laid out a four-part prescription for the technology’s peaceful development: uphold openness and mutually beneficial collaboration; guarantee AI controllability through legal frameworks and human oversight while pushing back against overly broad definitions of national security; promote inclusivity and dialogue among civilisations; and strengthen global governance centred on the UN.

The political messaging was immediately reinforced by a commercial thunderbolt. Semiconductor unit T-Head of Alibaba announced it was open-sourcing the SAIL software stack, the foundational architecture for its Zhenwu series of AI chips. The move was widely read as a direct assault on Nvidia’s dominant ecosystem, offering chip buyers – especially in the Global South – a fully open alternative.

Russia was visible on the sidelines. Senior Sber Vice President Andrey Belevtsev guided the Russian delegation around the exhibition, which featured the GigaChat AI assistant, the Kandinsky image and 3D generation models, and an anthropomorphic robot named Green. Notably, the signatory list lacked India, Beijing’s largest Asian rival and a major AI power in its own right.

From Chips to Governance: How Beijing Is Rewiring the Global AI Order

The WAICO Gamble: Building a Parallel AI Governance Track

By anchoring WAICO in the UN and gathering 29 countries, Beijing is constructing an alternative to the US-led, club-of-democracies approach to AI governance. Xi’s speech deliberately hit every note that resonates with the Global South: openness, rejection of technology monopolies, and a call to avoid weaponising national security. Yet the tension is visible. Chinese regulations still require generative AI to “reflect the core socialist values,” and Xi himself acknowledged that China equally prioritises development and security. Whether WAICO can offer meaningful openness while Beijing maintains strict domestic controls will shape the new body’s credibility and reach.

Alibaba’s Open-Source Chip Move: A Direct Challenge to Nvidia

The SAIL stack release is more than a technical gesture. Alibaba is giving system builders a permission-free, modifiable software layer for its Zhenwu AI chips, lowering both cost and lock-in. For enterprises and governments that are wary of dependence on US-designed hardware – and the export controls that accompany it – this could accelerate a shift toward Chinese silicon. The immediate target is the inference and training market that Nvidia currently dominates. The open-source model mirrors strategies that worked in software; whether it can succeed in the capital-intensive chip arena depends on the performance and manufacturing scale T-Head can deliver.

Russia’s Balancing Act: Sovereignty vs. Reliance

Russia’s prominent presence – from the WAICO signature to the Sber exhibition – illustrates a delicate dual-track strategy. Moscow is deepening practical collaboration with Chinese AI companies while insisting, through officials and experts, that it fully understands the risks of ecosystem dependency and is therefore preserving its own sovereign model programmes. The diplomacy serves to signal alignment with Beijing’s governance push while maintaining hedged technological housekeeping.

The Missing Elephant: India’s Absence and the Limits of China’s Appeal

That New Delhi did not join WAICO, even as a founding member, is telling. India boasts a fast-growing domestic AI sector, its own ambitions for digital public infrastructure, and a strategic relationship with the West that tempers its willingness to sign onto a China-led multilateral initiative. Its absence suggests that while WAICO may consolidate Beijing’s influence among states that lack a major independent AI base, it will struggle to become a truly global forum without buy-in from the large democracies of the Global South.

Immediate Implications for Industry, Investors and Policy

For semiconductor companies and AI platform builders, the Alibaba SAIL announcement is an actionable signal. Evaluate the performance benchmarks of the Zhenwu chips against existing Nvidia product lines, and engage early with cloud providers and government procurement agencies that are likely to favour open-source, non-US supply chains – especially in WAICO-member states. For Western AI firms, the existence of a UN-centred alternative governance track means that compliance with US or EU standards alone may not suffice to maintain market access in dozens of countries; begin scenario planning for fragmented regulatory regimes. For policymakers in the Global South, WAICO presents a concrete offer of capacity-building and technology access, but the package comes with alignment expectations that should be weighed against the leverage this gives Beijing over domestic digital infrastructure. For Russia’s tech leadership, the Sber showcase and WAICO signature reaffirm the near-term value of a China-facing strategy; the immediate priority is to lock in joint R&D projects that transfer hardware and software know-how while insulating sovereign AI programmes from exclusive dependence.

Risk & Opportunity Assessment

Commercial RiskHighWAICO and the open-source SAIL stack create a credible pathway for Chinese AI chips to capture market share from Nvidia, particularly in developing economies that join the new cooperation body.
Competitive RiskHighAlibaba’s SAIL release lowers the software barrier for adopting Zhenwu chips, potentially reshaping the competitive landscape for AI accelerators and enabling Chinese cloud providers to offer integrated open-source hardware stacks.
Regulatory RiskHighThe launch of a UN-centred AI governance body with 29 founding nations risks fragmenting global AI standards, increasing compliance costs for companies that must navigate both WAICO-aligned and Western regulatory frameworks.
Reputation RiskMediumXi’s calls for openness and resistance to overly broad national security measures coexist with China’s domestic AI censorship rules; if WAICO fails to deliver genuine openness, the credibility of Beijing’s governance model will suffer.
Technology DisruptionHighAn open-source AI chip software stack directly challenges the proprietary ecosystem model that underpins Nvidia’s dominance and could accelerate the diffusion of AI hardware design capabilities beyond a handful of suppliers.
Commercial OpportunityHighChinese AI companies and chipmakers stand to gain preferential access to infrastructure deals and government contracts in WAICO member states, leveraging Beijing’s ‘Digital Silk Road’ investments and open-source policy to lock in long-term customers.