Sovereignty by Download

In this piece for the CIGI blog, author Rafal Rohozinski discusses how artificial intelligence is rapidly becoming a central instrument of geopolitical power, with control over AI models, computing infrastructure, chips, energy, and data emerging as a new form of national sovereignty. Using the rise of Chinese open-weight AI models such as Moonshot AI’s Kimi K2.6 as an example, the author argues that the global AI race is no longer simply about technological innovation but about who controls the “cognitive layer” that underpins modern economies.

The article contrasts two competing approaches to AI development. The United States has largely pursued a centralized, hyperscale model dominated by firms such as OpenAI, Google, Microsoft, and Anthropic, where advanced models remain under corporate control and are accessed as services. China, by contrast, has emphasized open-weight and locally adaptable systems that can be downloaded, modified, and deployed independently. According to the author, this strategy extends Chinese influence by offering countries and organizations an alternative to dependence on American-controlled AI infrastructure.

Rohozinski argues that similar dynamics are already visible in cryptocurrency adoption across the Global South. Just as decentralized financial systems have gained traction where traditional Western banking infrastructure is inaccessible or constrained, open AI ecosystems are gaining appeal among governments and businesses seeking technological autonomy. In both cases, he suggests that efforts by the United States to restrict access through sanctions, export controls, or platform governance may inadvertently accelerate the development of alternative systems.

The article concludes that AI sovereignty is becoming a strategic concern not only for emerging economies but also for America’s allies. Countries such as Switzerland, France, India, Saudi Arabia, and the United Arab Emirates are investing heavily in sovereign AI capabilities, recognizing that access to critical AI infrastructure can be altered by political decisions beyond their control. As AI becomes embedded across defense, finance, energy, logistics, healthcare, and public administration, the author argues that ownership of the cognitive layer—not merely access to it—will increasingly determine economic power, national resilience, and geopolitical influence in the decades ahead…

CONTINUE READING…