E-Paper
Digital swing states: How middle powers can bridge the global AI divide
The global AI landscape is increasingly being seen as a two-horse race between the United States and China, where the vast majority of infrastructure, capital, and talent are concentrated.
The global AI landscape is increasingly being seen as a two-horse race between the United States and China, where the vast majority of infrastructure, capital, and talent are concentrated. The two superpowers control nearly 70% of the world’s top-tier AI researchers, with their private and public investments reaching hundreds of billions of dollars each. This dominance is not merely economic; recent developments suggest both nations are willing to weaponise their technological capabilities to maintain geopolitical edges.
This reality leaves significant powers–India, France, the UK, Japan and others–in a precarious position. They are understandably worried and unwilling to remain bystanders. However, the astronomical requirements for compute and capital in the frontier model era mean that no middle power can successfully go it alone. As UN Tech Envoy Amandeep Gill has warned, doing so risks a ‘Second Great Divergence,’ potentially leaving some nations far behind others, as happened during the Industrial Revolution.
What Can the Middle Powers Actually Do?
To navigate this, middle powers must shift their strategy from mere survival to actively leveraging the US-China competition. The solution lies in forming ‘minilateral alliances’–strategic small groupings that pool resources to preserve autonomy.
The world order is moving away from an era of multilateral bodies toward the age of the minilateral, and the AI world order should be no exception. A prime example–but by no means the only one–would be a partnership between India, France and Japan. India’s vast talent pool and vibrant startup ecosystem combined with France’s low-carbon compute and Japan’s precision hardware strengths could create a platform of capabilities that neither the US nor China could ignore.
As I argued in my book, The Great Tech Game, new friendships and alliances will increasingly be built on such technological complementarities. These tech alliances will act as a geopolitical shield: while a single middle power might be bullied by a superpower’s export controls or cloud monopolies, a collective bloc represents a market and research base too large to browbeat.
Pool Resources to Build Strategic Autonomy
Strategic autonomy in AI requires pooling talent, R&D budgets and compute clusters to develop autonomous AI stacks. This ensures that critical infrastructure is not reliant on black boxes controlled by Washington or Beijing..
History offers a roadmap for this kind of collaboration. In the 1960s, European nations realized they could not individually compete with American aerospace giants. By merging the efforts of France’s Aérospatiale, Germany’s Deutsche Airbus, and UK’s Hawker Siddeley to create Airbus, they broke the American monopoly and secured industrial sovereignty. Similarly, CERN (the European Organization for Nuclear Research, based in Switzerland) allowed multiple nations to share the astronomical costs of particle physics research. Middle powers today must apply this logic to AI.
Build a Credible Threat of Switching Technological Choices
Becoming a ‘swing state’ of the AI age is about building leverage through partnerships. By building credible AI capabilities together, middle powers would gain collective bargaining power and the credibility to potentially switch technological choices. This would create a credible threat of ‘switching’: if the US threatens to restrict access to frontier models or tighten chip exports, the middle-power bloc could pivot towards open-weight models, or European-designed RISC-V architectures. This leverage would force the superpowers to offer better terms, including more technology transfer with fewer strings attached.
Furthermore, because both superpowers are hungry for large consumer and enterprise markets, a united, coordinated middle-power bloc can implement ‘market access-for-tech transfer’ policies. In return for access, these countries could require the large AI companies to build data centers locally, share underlying model weights, invest in local R&D and train local engineers.
Beyond the Rhetoric: The Hard To-Do List
The execution of the logic outlined above is hard, as middle powers often face incentives to defect and seek individual deals, as we saw with tariffs. To overcome this temptation, these nations must operationalize AI partnerships quickly.
They should create a ‘Common Compute Reserve’– a shared network of high-performance computing centers available to the bloc’s startups and researchers, much like a strategic oil reserve. This reduces the dependency that currently tethers them to large hyperscalers and cloud providers. They should develop a strategically aligned yet mobile pool of top-tier talent, through fellowships and joint PhD programs that allow researchers in Tokyo, for example, to work seamlessly with labs in Bangalore. Joint investment vehicles could help scale companies emerging from such collaborative work.
Beyond infrastructure, middle powers can win by coordinating and leading on AI governance and safety. By establishing open, interoperable standards domestically and in the various international standards organizations (ISOs) like the International Telecommunications Union (ITU), they can create a regulatory environment that foreign firms must follow to operate in their markets. This would allow them to shape the global industry without necessarily owning the most GPUs.
The Imperative to Avoid Digital Colonialism
Middle powers are currently having a moment on the global chessboard, and they must capitalize on it. The India AI Impact Summit 2026 provides the perfect platform for such cross-country collaboration bridges to be built. By forming minilateral alliances, countries like India, France, and Japan can shift from being potential targets of ‘AI colonialism’ to being strategic, autonomous pillars of the new global AI order. That will help narrow the global AI divide.
Anirudh Suri is Managing Director at India Internet Fund, host of The Great Tech Game Podcast and a nonresident scholar with Carnegie India.
