Artificial Intelligence (AI) is rapidly emerging as a defining arena of geopolitical competition. Beyond the race to develop advanced AI models, major powers are increasingly seeking to shape the institutions, standards, and partnerships that will govern the future of AI. China’s recent establishment of the World AI Cooperation Organisation (WAICO) and the United States’ Pax Silica Initiative reflect two distinct approaches to global AI governance. In this interview, we speak with Prof. Dr. Özgür Tüfekçi about the strategic significance of these initiatives, the evolving US-China technology rivalry, the role of partnering countries, and the prospects for cooperation or fragmentation in the emerging global AI order.
China recently established the World AI Cooperation Organisation (WAICO). What motivated Beijing to launch this initiative, and what strategic objectives does it hope to achieve?
WAICO should be seen in a broader context of the shift of artificial intelligence from a technological realm to an international order-making realm. China is no longer concerned solely with competing in the development of advanced AI models. It also wants to have a role and influence in the institutional framework for the development, regulation and dissemination of AI.
From Beijing’s perspective, there are at least three strategic considerations. First, China has little interest in seeing a global AI governance regime emerge that is primarily shaped by the United States and its technologically advanced partners. Second, WAICO offers a channel for China to further its technological and diplomatic cooperation with nations that might not want to be entirely tied to the United States or China. Third, it would enable China to shape AI governance around values it has sought to promote in its broader approach to global governance – sovereignty, development, inclusiveness, multilateral consultation and greater access to technology.
But I wouldn’t want to read WAICO merely as an anti-American institution. It is more fully understood as being part of China’s general push to establish its institutional presence in fluid zones where international institutions have yet to be fully developed. AI governance is still an emerging regime. Under such conditions, states can gain a great deal of agenda-setting power if they are early players in institution-building.
Some analysts argue that WAICO is more than an AI cooperation platform. To what extent is China seeking to shape the future rules and institutions of global AI governance through this organisation?
I believe that’s a sound interpretation, but there’s a difference between intention and the eventual capacity. WAICO is significant precisely because the institutional architecture of global AI governance remains unsettled. Key questions concerning AI safety, access to computing power and critical materials, intellectual property, cross-border data transfers, algorithmic accountability and data governance remain unresolved.
China thus has an opportunity to participate in – and potentially shape – the institutionalisation of global AI governance. Whether WAICO proves attractive to countries beyond China’s established strategic partners, however, will ultimately determine how much influence Beijing can exercise over the development of global rules. International legitimacy is not merely the result of powerful states establishing institutions, but the result of other states seeing participation as beneficial.
This question is especially crucial for middle powers and developing countries. Many of them do not view AI governance solely through the lens of US-China competition. They are concerned with access to low-cost technologies, infrastructure, skills, investment, data sovereignty and rule-making. If WAICO can address these issues responsibly and respond to these concerns, then the influence may be significant. Conversely, its institutional appeal will be more limited if it is perceived primarily as an instrument of geopolitical competition.
The United States introduced the Pax Silica Initiative in 2025, and China has now established WAICO. How do these two initiatives differ in their strategic objectives and visions for global AI governance?
The most crucial difference is that they start from different dimensions of the AI ecosystem. Pax Silica is mainly concerned with the material and strategic foundations of technological power. Semiconductors, critical minerals, energy, advanced manufacturing, data centres, logistics and complex supply chains are all essential to AI. The US strategy thus focuses on resilience, trusted partnerships, technological security, and minimising strategic vulnerabilities.
WAICO, by contrast, begins more explicitly from the institutional and governance dimension. It highlights international cooperation, participation, development and the building of governance structures for artificial intelligence in its discourse. But the difference shouldn’t be overstated. Infrastructure and governance are becoming more and more intertwined. Anyone who has an impact on semiconductor supply chains, cloud, technical standards, data ecosystems and AI platforms gains significant influence over the rules that govern these technologies.
It is possible, then, that we are witnessing two variations of institutional power that are becoming more similar. The United States is aiming to develop a network of strategically trusted, technologically sophisticated partners critical for AI supply chains. Meanwhile, China is working to build a wider multilateral arrangement that will resonate with nations looking to access more technology and have more influence over global AI governance. One will have to see whether these approaches are still complementary in certain areas or whether they develop into competing institutional architectures in the long run.
Do you see WAICO and Pax Silica as the foundations of competing AI ecosystems, or is meaningful cooperation between these two approaches still possible?
Indeed, there is always a danger that they might contribute to the formation of competing technological ecosystems. There are already some signs of fragmentation in semiconductor supply chains, in export regulations, in telecommunications networks, in digital platforms and in the regulation of data. This tendency might be exacerbated by AI. However, I would be reluctant to call a bipolar AI order a foregone conclusion.
Technologies are extremely dependent on each other. The semiconductor manufacturing industry is a complex network; the United States is linked to East Asia, Europe and the rest of the world. The same is true of critical minerals, energy, cloud infrastructure, research communities and global markets.
South Korea illustrates this complexity particularly well: it is a close security partner of the United States while maintaining significant economic ties with China. Such structural realities abound in many middle powers. Thus, strategic flexibility and diversification will often be preferred by governments over total technological decoupling.
Managed technological pluralism may therefore be the most likely scenario, where the most distinct ecosystems become more competitive, and there is more cooperation where there’s still a shared interest. AI safety, as one example, is the clearest case. The risks associated with highly capable AI systems do not respect national borders. It is quite important that minimum standards regarding safety, transparency, crisis communication and responsible deployment should be considered as a matter for dialogue even in the face of strong strategic rivalry.
China presents WAICO as an open platform for international cooperation, particularly for developing countries. How attractive do you think this initiative will be to the Global South compared with US-led technology partnerships?
Its potential attractiveness should not be underestimated. Many countries in the Global South are less interested in choosing sides than in securing affordable access to infrastructure, computing power, investment, technological expertise and opportunities for digital development. Many are also concerned about being reduced to consumers of technologies and rules developed elsewhere. China has ample expertise in meeting this demand through infrastructure financing, telecommunications, digital platforms, and broader development partnerships. WAICO may provide Beijing with an opportunity to extend a similar approach into the AI domain.
But it does not necessarily mean geopolitical alignment with China. This is a significant difference. A country can partner with China for digital infrastructure, the United States for advanced technology, the EU for standards and regional organisations for data governance. The Global South should not be understood merely as a theatre of great-power rivalry, but also as a domain of considerable state agency. Many governments will therefore seek to derive benefits from multiple partnerships rather than align exclusively with a single technological bloc.
What concerns me more is whether, with these new institutions of AI, the states will be able to have any real control over governance, or merely reproduce technological entanglements in other guises. Both Chinese and US-led initiatives will need to be successful to some degree in convincing the world that technological cooperation can produce increased developmental benefits, not just new areas of technological dependence.
Competition in artificial intelligence now extends beyond algorithms to semiconductors, critical minerals, data, cloud infrastructure, and computing power. Which of these domains will be the most decisive in shaping global leadership over the next decade?
I would not identify a single decisive factor, because AI is increasingly best understood as an ecosystem rather than a discrete technological capability. The most immediate strategic constraint today is access to advanced semiconductors. The semiconductor industry, in turn, relies on cutting-edge manufacturing tools, specialised materials, critical minerals, energy, capital, manpower and intricately complex global supply chains.
Over the coming decade, energy could prove to be a key issue. The provision of reliable and affordable electricity at scale will be a key factor in how competitive AI can be, as AI models and data centres demand more and more power. Data and human capital will also remain critical components of technological leadership. Leadership cannot be created without quality data, researchers, engineers, entrepreneurial ecosystems and institutions that will be able to turn technological innovation into economic productivity.
I would therefore characterise the emerging competition as a contest over the capacity to build and sustain integrated AI ecosystems. The states that most successfully achieve their goals will not necessarily have the best individual model at any one time. They will be the ones who can pull together semiconductors, energy, data, infrastructure, capital, research institutions, talent and international partnerships into a robust technological ecosystem. Alliances and partnerships are important, too, because of this. Even the most powerful players are unable to manage the whole AI value chain on their own.
Looking ahead to 2035, do you expect the world to move toward a common framework for AI governance, or are we entering an era of competing AI blocs led by different powers?
I do not expect either complete convergence or complete fragmentation by 2035. Rather, a multiplex AI order will likely be the norm. At one level, there is a possibility of the creation of opposing technological ecosystems for key powers, such as China and the USA. The configuration of these ecosystems can vary from one another with regard to supply chains, platforms, standards, data governance, and regulatory philosophies. At another level, some form of common governance will continue to be necessary. Issues such as AI safety, cybersecurity, autonomous systems, intellectual property, cross-border risks and the potential for misuse of highly capable AI systems are problematic and can’t be solved solely within a technological bloc.
The history of international governance suggests that strategic competition and institutional development can proceed simultaneously. During earlier periods of geopolitical rivalry, states competed intensely while simultaneously developing mechanisms for arms control, trade, communications and other cross-border challenges. So, the really important issue is whether there can be any meaningful institutional communication in the face of growing technological competition.
Middle powers may play a particularly important role in this process. Countries such as South Korea, embedded simultaneously in different economic and strategic networks, may help bridge competing perspectives and facilitate compatible governance structures. This is what I mean by a multiplex AI order: a system defined simultaneously by competition, interdependence, overlapping institutions and selective cooperation. The biggest challenge for the international community will be to prevent geopolitical rivalry from crowding out cooperation precisely in those areas where the risks associated with AI are inherently global.
Prof. Dr. Özgür Tüfekçi is a Professor of International Relations at Karadeniz Technical University, Türkiye and Director of the Strategic Research Centre at the university. He has a PhD from Coventry University and a master’s from the University of Sheffield. His research focuses on international order, geopolitics, grand strategy, Eurasian studies, emerging powers and Turkish foreign policy, with particular attention to shifts in the international distribution of power and state responses to geopolitical and institutional change. Professor Tüfekçi is also the Founder and Director-General of CESRAN International, Convenor of the CESRAN Conference Series, and Editor-in-Chief of The Rest: Journal of Politics and Development and Novus Orbis: Journal of Politics and International Relations.