Plugged in, not locked in: middle powers navigate US-China AI rivalry

Middle powers navigating the US-China competition in artificial intelligence (AI) favour flexible partnerships over closed alliances, demanding a “plugged in, not locked in” approach, an AI governance panel heard on Saturday.

Alejandro Reyes, adjunct professor in the department of politics and public administration at the University of Hong Kong, said that middle powers were exercising agency in the AI era, selectively partnering with both Beijing and Washington “issue by issue” rather than aligning with a single camp.

“Countries in the middle do not want to be locked in; they want to plug in,” Reyes told a forum in Shenzhen co-hosted by the Shenzhen Loop Area Institute and the Centre on Contemporary China and the World (CCCW) at HKU.

“The stack that earns their trust will be the one that combines easily with others and leaves them free to work with other partners too,” said Reyes, who is also a senior fellow at CCCW.

The US and China are competing for influence in AI infrastructure investment in regions such as Southeast Asia and the Middle East.

In Malaysia’s Johor state, American enterprises including Microsoft, Google and Oracle are establishing data centres and cloud services alongside similar facilities built by Chinese tech giants ByteDance, Tencent Holdings and Alibaba Group Holding. Alibaba is the owner of the South China Morning Post.

The United Arab Emirates is another high-profile example, with Washington supplying advanced chips and models and China providing affordable cloud services and digital connectivity.

02:33

The Indonesian island where residents compete with AI data centres for fresh water

Reyes, who previously advised the Canadian foreign ministry, argued that middle economies should pursue “network sovereignty” – maintaining domestic control over critical aspects closest to citizens, such as data, energy and public-service AI, while diversifying upstream suppliers to prevent any single country from wielding monopolistic leverage.

“Few countries can control every layer of the technology stack … What countries can have is freedom within interdependence,” he said.

The US and China are vying for AI supremacy. Washington relies on venture capital and tech giants, while Beijing favours cost-effective models and prioritises practical deployment through state planning.

Governance frameworks are similarly diverging. Beijing has launched the China-based World Artificial Intelligence Cooperation Organisation, championing beneficial, safe and fair AI development. Washington’s 24-nation Pax Silica initiative expands AI norms to semiconductor supply chain security and aims to foster innovation.

Reyes said policymakers must build a “collaboration stack” – coordinating governments, the private sector, universities and local communities to test governance frameworks in practice.

SCMP Plus is a new premium news platform that gives you an

all-inclusive edge to stay ahead on China news.

To access our exclusive content you’ll need to subscribe.

Already a subscriber?

LOG IN

“We have been building the technology stack much faster than the collaboration stack,” he said, citing a lag in multilateral rule-making and implementation.

To address this, he proposed priorities for regional cooperation, including pooled computing resources, open models for Asian languages and a standing network for responding to cross-border AI incidents.

Cui Tiankai, a former Chinese ambassador to Washington who attended the forum, said AI governance must address members’ concerns and needs while encouraging widespread participation and collaborative oversight.

He said the use of AI in the Asia-Pacific region should benefit all members, “promoting shared progress and common prosperity”, and called on the US to set aside domestic political divisions and seek AI cooperation with China.

Reyes identified the Greater Bay Area as an ideal testing ground for multilateral governance.

“Shenzhen brings technological and industrial capacity. Hong Kong brings research, international connections, a different legal and regulatory system and financing,” he said.

“Together, they could test practical forms of cross-border AI governance – from joint safety evaluations to incident response protocols – and share what works with partners across the Asia-Pacific.”