The Gambit of the Century: AI in a Multipolar World

In August, media outlets reported that the US was preparing to inform its allies that they would eventually have to pick a side in what will surely be the gambit of the century. Apparent battle lines were drawn in July, when twenty-nine states gathered for the inauguration of the World Artificial Intelligence Cooperation Organisation (WAICO) in Shanghai, after twenty-five other states signed the Pax Silica Washington in December 2025.

Yet, despite the overtones, we should be careful not to view this emerging US-Sino schism through a Cold War lens of bipolarity, which is insufficient and reductive in light of modern multipolarity. Whilst nuclear weapons and Intercontinental ballistic missiles (ICBMs) were ultimately instruments of destruction and deterrence. Artificial Intelligence is a far more totalising technological system with disparate possibilities for reshaping production, warfare, communication, governance and everyday life. The gambit, then, is not simply over who controls AI, but what technological power will mean in this new world.

The US is operating under the logic of ‘peace through strength.’ Security is achieved by controlling, or securing, the material foundations upon which technological power depends. Reducing ‘coercive dependencies’ on semiconductors, critical minerals, energy manufacturing, and digital infrastructure is key for the US, which has seen the delicate chokepoints that characterise supply chains under globalisation.

Securing a consistent, reliable supply of semiconductors is particularly crucial for the US, which currently sources semiconductors from an eclectic, transnational supply chain. UK-based Arm provides foundational chip architecture, Dutch ASML supplies critical lithography equipment, Taiwan’s TSMC manufactures many of the world’s most advanced chips, whilst South Korean firms dominate key memory technologies.

The US government has expressed the view that semiconductors will be for the 21^st century what oil was for the 20^th. The importance of these little chips will only grow over time. It is estimated that currently 54 cents of every dollar spent on data centre construction is spent on semiconductors. Additionally, demand for chips is continually increasing, with sales exceeding $400 billion in the second quarter of 2026 – a 35.1% increase from the first. Ultimately, without a secure semiconductor ecosystem, the US would struggle to project power and influence in the multipolar world. It would also hinder its ability to continue developing artificial superintelligence, which they hope will give it the edge over China.  

The US relies on the dynamism of private enterprise to spur innovation, with several tech companies – including Microsoft, Meta, OpenAI and Apple, amongst others – being elevated to a status resembling technological fiefdoms. This approach not only blurs the traditional distinction between state and private power, but also reflects an increasingly selective, post-liberal approach to economic and technological security. Where the previous liberal paradigm sought stability through openness and universalised interdependence, Washington seeks security through selective interdependence between trusted partners. It is a state of play which increasingly reflects the uncertainty in today’s multipolar world.

China, meanwhile, is pursuing a different route to technological power. Whereas Washington is attempting to secure the material foundations of AI, Beijing aims to shape the political and normative environment in which those technologies develop. At WAICO’s inauguration, Xi Jinping spoke of a ‘people-centric’ approach to AI development, stressing transparency, accessibility, multilateralism, equality and adherence to the UN Charter, with Secretary-General António Guterres notably delivering a speech.

Beijing has attempted to give these principles material form. This is most apparent with its open-source AI models like DeepSeek and Qwen, which allegedly have been downloaded over ten billion times. China has also previously set up several forums, including the Global AI Governance Initiative, the AI Capacity-Building Action Plan for Good and for All and the AI+ International Cooperation Initiative.

However, these initiatives are not purely altruistic. China clearly recognises the growing gap between the Great Powers in their great gambit and those less-considered in the multipolar world, making WAICO particularly attractive among states in the Global South. 

China plans to bridge this gap by developing 5000 opportunities in AI training over the next five years, building cooperation centres with Global South intergovernmental organisations, and providing 30 states with MAZU, an AI-powered meteorological warning system. WAICO will therefore create technological relationships with states traditionally heavily dependent upon existing Western infrastructure, encouraging them to rely on Beijing instead.

Besides the material benefits these initiatives will yield, their strategic value is also immense. WAICO will ultimately give Beijing a platform to shape norms, standards and institutions governing AI, therefore influencing what legitimate technological development looks like in this emerging, critical paradigm. Unlike the realist strategy undertaken by the US, China appears intent on a constructivist approach where power derives not just from controlling technological capabilities, but from shaping the norms through which those capabilities are understood and legitimised. 

Eisenhower’s 1953 Atoms for Peace speech helped establish a language through which peaceful nuclear development could be imagined. Perhaps Xi hopes to do something similar with AI, albeit through a more explicitly pluralistic framework.

Even if Washington appears intent on making countries choose a side, reducing today’s technological order to yesterday's bipolarity risks overlooking the complexity of the system emerging beneath. A Cold War narrative with two ideological blocs, organised around rival political and military systems, only goes so far. The emerging AI paradigm is considerably messier. Its geography is disparate, its actors are both private and public, and its technologies depend upon supply chains that cross political boundaries.

Kazakhstan, for instance, is currently the sole signatory to both Pax Silica and WAICO, demonstrating the disparity between technological alignment and broader geopolitical allegiance. India is another contradiction: despite being a founding member of BRICS and currently holding the presidency, it has aligned itself with Pax Silica. 

Then there is the EU, which is entangled in regulatory and technological interests whilst attempting to retain its relationships with both sides. Diplomatic non-alignment may remain possible; technological non-alignment is considerably more difficult, as AI depends on chips, cloud infrastructure, capital, data and standards spanning national borders.

The test, then, will not be what Washington or Beijing want, but how the rest of the globe navigates between them. AI may produce neither a new Cold War nor the end of globalisation, but something more complex: a Globalisation 2.0. Interconnection will remain, perhaps even intensify, but through increasingly transnational networks of states, corporations, capital, infrastructure and technology. The gambit of this century may therefore not be about choosing between two technological orders, but about whether a multipolar world can accommodate them both.

Image: Wikimedia Commons/Xuthoria

Licence. 

No image changes made.