#1
Two visions: WAICO, Pax Silica and the future of AI supply chains in the Asia-Pacific
This article was co-authored with James Nicholls.
In July 2026, representatives from 29 countries met in Shanghai to sign an agreement establishing the World Artificial Intelligence Cooperation Organization (WAICO). WAICO is the first intergovernmental body established by treaty whose primary mandate is the governance of artificial intelligence (AI). Its founding members include China, Russia, Brazil, Indonesia, Malaysia, Pakistan and Laos, alongside other countries from Southeast Asia, Africa,
展开查看正文
This article was co-authored with James Nicholls.
In July 2026, representatives from 29 countries met in Shanghai to sign an agreement establishing the World Artificial Intelligence Cooperation Organization (WAICO). WAICO is the first intergovernmental body established by treaty whose primary mandate is the governance of artificial intelligence (AI). Its founding members include China, Russia, Brazil, Indonesia, Malaysia, Pakistan and Laos, alongside other countries from Southeast Asia, Africa, and Latin America. The absence of any Group of 7 (G7) countries reflects a deepening fracture in a global AI order shaped by alliance politics, export controls, and competing approaches to critical infrastructure.
WAICO's establishment has the potential to cut across defence alliances, reshape regulatory frameworks, and influence investment flows in data centres across the Asia-Pacific region.
To date, AI governance has been shaped predominantly by Western-led institutions. The EU AI Act, adopted in 2024, imposes mandatory, risk-tiered compliance with extraterritorial reach, but focused on the AI product and use cases, rather than the infrastructure that powers it. The OECD AI Principles represent a voluntary soft-law standard endorsed by 46 countries. The G7 Hiroshima AI Process, which produced international codes of conduct for advanced AI developers, has expanded to 49 participating countries.
Alongside these sits Pax Silica, a US-led coalition of allied and partner economies focused on AI and supply chain security - from Dutch chipmaking machinery and Japanese chemicals to Korean memory chips and American chip design. India formally joined during 2026, reinforcing the initiative's growing Indo-Pacific orientation.
Against this backdrop, a 2024 UN report found that 118 countries, primarily in the Global South, are parties to none of seven prominent international AI governance initiatives. WAICO can be viewed as one response to concerns that many Global South states remain under-represented in existing AI governance initiatives.
The WAICO founding agreement is light on detail, but commentators have suggested that WAICO may eventually address issues such as technical standards, data governance, localisation, and infrastructure procurement. WAICO is anticipated to help countries adopt AI - albeit through Chinese-infrastructure supply chains aligned with Chinese open-weight models. Measured against multilateral norms, WAICO's early design raises questions: China is likely to supply the models, infrastructure, and expertise, while key governance details - voting rules, funding, and leadership selection - remain undefined. Critics argue that WAICO could create technological dependencies for countries that build their digital future around Chinese AI ecosystems.
As WAICO and Pax Silica crystallise into competing institutional structures and decisions about AI infrastructure take on strategic importance, the question is no longer whether bifurcation will occur, but how deeply it will shape the technological and geopolitical trajectory of Asia-Pacific countries.
Data centre construction illustrates how this divide will reshape the investment landscape. Approximately US$800 billion in data centre investment is expected across the Asia-Pacific by 2030, with data centre investment reaching a record US$11.6 billion in 2025 and hyperscalers making multi-billion-dollar commitments across Singapore, Indonesia, India, and Japan. Each investment threatens to deepen the divide between WAICO and Pax Silica.
The United States maintains a tiered system of access to advanced computing semiconductors. Tier 2 countries including Singapore, India, and most ASEAN nations can only acquire chips with validated end-user approval, while China sits in Tier 3 where approvals are generally refused. The US Foreign Direct Product Rule extends these restrictions to chips made anywhere using US-origin technology, software, or equipment. In practice, access to advanced semiconductors may depend on whether countries and companies operate within Pax Silica-aligned supply chains.
The United States has also recently used export controls to temporarily limit access to frontier AI models to US nationals. The order, the first of its kind targeting a specific frontier AI model, appears to reinforce the broader pattern in which Pax Silica membership increasingly determines which AI capabilities are available, and to whom.
Conversely, China leads in cooling systems, electrical infrastructure, and AI model deployment throughout the developing world. Its technology companies occupy critical positions in AI stacks across Southeast Asia, Africa, and Latin America, while Chinese open-weight models now account for up to 46% of tokens on major developer platforms.
Indonesia illustrates this dynamic. Jakarta quickly endorsed WAICO despite being traditionally non-aligned in US-China technology competition. Chinese firms have invested heavily in Indonesia's e-commerce platforms and 5G network infrastructure, providing a foundation for Beijing to layer AI services and governance institutions. Across the region, WAICO member states receiving Chinese AI training, infrastructure, and open-source models may be more likely to adopt regulations consistent with WAICO standards.
For developers, investors, and operators, these developments mean infrastructure choices carry regulatory and geopolitical consequences well beyond project economics. A data centre built in Indonesia or Malaysia using Chinese compute infrastructure and Chinese-standard cooling systems will operate within an entirely different regulatory and supply chain ecosystem, and may attract different sets of users from one built within the Pax Silica-aligned framework of trusted vendors and US-origin technology. The same logic applies to undersea cable networks where routing decisions, landing-site permits, and vendor selection are becoming arenas of strategic bargaining. On the other hand, China has argued that WAICO advances inclusion and capacity-building.
Many Asia-Pacific states, including Singapore, therefore seek to remain interoperable across both ecosystems rather than align exclusively with either.
Defence considerations compound these risks. China's National Intelligence Law of 2017 has been interpreted as obliging Chinese manufacturers and citizens to support, assist and cooperate with national intelligence work. Cambodia, Indonesia, Laos, Malaysia, and Myanmar are founding WAICO members whilst simultaneously participating in US-aligned security arrangements, including Indonesia's growing defence cooperation with Washington, or Malaysia's participation in the Five Power Defence Arrangements. AI systems trained on Chinese architectures, governed by Chinese-aligned standards, and hosted on Chinese cloud infrastructure may raise interoperability, security-assurance and procurement questions for participants in certain defence and intelligence arrangements.
Energy is a further dimension of this securitisation as Asia-Pacific data centre capacity expands from 32 GW to a projected 57 GW by 2030. Access to reliable power - from LNG, cross-border grids, or renewables - is becoming strategically important and no longer purely commercial for operators and investors. Where power comes from and who the suppliers are may ultimately determine whether a facility is interoperable with Pax Silica or WAICO-aligned standards.
The launch of WAICO signals that AI supply chain issues have moved from political rhetoric to alternative visions of institutional frameworks with real-world regulatory and procurement consequences. For Asia-Pacific businesses, this is a live commercial issue. Businesses must factor these rival systems into supplier, compliance, and infrastructure decisions, because a choice made in one system may limit access to the other. Governments and enterprises must navigate this fracture, keeping room to move where they can while managing the legal, commercial, and geopolitical risks of operating in a world where the rules of AI infrastructure are being written twice.