Middle powers seek flexible AI partnerships to avoid superpower lock-in

The 'Plugged In' Strategy

Countries across Southeast Asia and the Middle East are navigating the growing rivalry between the United States and China by seeking flexible partnerships. Rather than choosing a side, these middle powers are attempting to remain 'plugged in' to both ecosystems without being 'locked in' to one. This approach allows them to select the best tools and investments based on specific national needs.

Alejandro Reyes, a professor at the University of Hong Kong, explained that these nations are exercising agency by partnering with Beijing and Washington on an issue-by-issue basis. They want to avoid being forced into a binary choice that could limit their economic or technological growth. By maintaining multiple connections, they can balance the competing interests of the two superpowers.

The goal for these nations is to find a technology stack that earns their trust while remaining compatible with other partners. Reyes noted that the most successful systems will be those that allow countries to remain free to work with various international players. This strategy prevents any single superpower from gaining monopolistic leverage over a nation's digital infrastructure.

Infrastructure in Practice

Real-world examples of this balancing act are already appearing in major tech hubs and regional economies. In Malaysia's Johor state, American companies like Google, Microsoft, and Oracle are building data centers and cloud services. At the same time, Chinese tech giants such as ByteDance, Tencent Holdings, and Alibaba Group Holding are establishing similar facilities in the same region.

The United Arab Emirates serves as another prominent example of this dual-track approach to AI development. The UAE works with Washington to acquire advanced chips and AI models, while simultaneously utilizing China for affordable cloud services and digital connectivity. This allows the country to benefit from high-end innovation and cost-effective scale at the same time.

These developments show that middle powers are not just making theoretical statements but are actively diversifying their physical infrastructure. By hosting hardware and services from both sides, they create a buffer against the risk of being cut off from essential technology. This geographical and economic distribution of tech assets is a core component of their current strategy.

Defining Network Sovereignty

To manage these complex relationships, Reyes argues that middle economies should pursue a concept known as 'network sovereignty.' This involves maintaining domestic control over the critical aspects of technology that are closest to citizens, such as data, energy, and public-service AI. By keeping these core elements local, countries can protect their national interests while still participating in the global market.

Because it is nearly impossible for any single country to control every layer of the technology stack, the focus shifts to 'freedom within interdependence.' This means relying on a diverse range of upstream suppliers to ensure that no single country can exert total control over a nation's digital life. It is a move toward a more decentralized model of technological reliance.

This strategy requires a careful balance between opening up to international innovation and securing domestic systems. By diversifying who provides the raw materials and high-level models, middle powers aim to insulate themselves from geopolitical pressure. This approach seeks to maximize the benefits of global AI while minimizing the risks of dependency on a single political ally.

Building a Collaboration Stack

While the technology stack is advancing rapidly, there is currently a lag in the 'collaboration stack,' which refers to the rules and frameworks for using AI. Reyes suggests that policymakers need to coordinate governments, the private sector, universities, and local communities to test governance in the real world. This involves moving beyond theory to practical, multilateral rule-making and implementation.

Proposed priorities for this regional cooperation include pooling computing resources and creating open models specifically for Asian languages. There is also a call for a standing network to respond to cross-border AI incidents quickly. These initiatives would help create a shared safety net for countries that are not aligned with a single superpower's rules.

Former Chinese ambassador Cui Tiankai emphasized that AI governance must address the specific needs of all members to encourage widespread participation. He called for the US to set aside domestic political divisions to seek cooperation with China. The ultimate goal is to ensure that AI progress in the Asia-Pacific region promotes shared prosperity rather than creating new divisions.

Testing Grounds for Governance

The Greater Bay Area has been identified as a potential testing ground for these new forms of multilateral governance. This region combines the industrial and technological capacity of Shenzhen with the research, financing, and unique legal systems of Hong Kong. Together, they offer a diverse environment to experiment with cross-border AI regulations.

The region could be used to test practical governance measures, such as joint safety evaluations and incident response protocols. By testing these frameworks in a high-capacity area, policymakers can determine what works before scaling the solutions to other partners. This iterative approach allows for practical adjustments based on real-world data and results.

Successful testing in the Greater Bay Area could provide a blueprint for the rest of the Asia-Pacific. It offers a way to harmonize different regulatory systems while maintaining the flexibility that middle powers desire. As these nations continue to navigate the US-China rivalry, such collaborative hubs may become essential for establishing global AI norms.

The Path Forward

The current landscape shows a shift away from exclusive blocks toward a more nuanced, multi-aligned approach to AI. Middle powers are increasingly savvy about the risks of total dependency and are carving out spaces where they can interact with both Washington and Beijing. This 'plugged in' reality will likely define the next decade of international technology policy.

For the global community, this means that AI standards will not be dictated by a single set of rules. Instead, we may see a patchwork of regional agreements and 'collaboration stacks' that prioritize local needs and shared safety. The success of this model depends on the ability of middle powers to maintain their agency in a highly competitive environment.

As the competition for AI supremacy continues, the focus will shift toward who can provide the most reliable and compatible infrastructure. Countries that can successfully manage these diverse partnerships will be best positioned to harness AI's benefits. The move toward network sovereignty suggests a future where technology is global, but control remains local.