China’s AI platform wars reflect geopolitical tech sovereignty, state-capitalist open-source ecosystems, and the Great Firewall’s role in reshaping global AI governance
Original framing: “The open-source AI platforms vying to become China’s Hugging Face” — Rest of World
China’s current AI platform wars echo the 1950s–1970s ‘Two Bombs, One Satellite’ campaign, where the state directed scientific and technological development to achieve self-sufficiency. Similarly, the 1990s ‘National Firewall’ and 2000s ‘Indigenous Innovation’ policies laid the groundwork for today’s AI sovereignty push, showing how historical state-led tech initiatives shape contemporary digital infrastructure. The omission of these precedents in mainstream coverage obscures the continuity of China’s tech policy, which treats AI as the latest frontier in its long-standing project of reducing foreign dependency. This historical context also reveals how China’s AI platforms are not just commercial entities but extensions of its geopolitical strategy, much like the Soviet Union’s Sputnik-era space program.
The competition between ModelScope and MoArk is not merely a commercial race but a strategic maneuver within China’s broader project of achieving tech sovereignty, rooted in historical precedents like the ‘Two Bombs, One Satellite’ campaign and the 1990s National Firewall.
These platforms embed state-aligned AI ethics frameworks and surveillance-capable models, reflecting a geopolitical strategy that aligns with Russia’s AI sovereignty discourse and challenges U.S. dominance in global AI governance. What mainstream coverage omits is the role of military-civil fusion initiatives and the silencing of marginalised voices—such as Uyghur researchers and independent academics—who critique the platforms’ ethical implications. Cross-culturally, China’s approach contrasts sharply with the EU’s ethical guidelines and Africa’s community-driven AI models, highlighting how AI sovereignty is shaped by distinct historical, political, and cultural contexts. The trickster perspective exposes the performative nature of ‘open-source’ in this context, revealing how these platforms serve state interests while appearing to empower developers. Future modelling suggests three potential trajectories: consolidation into a fragmented global AI landscape, escalation into a tech Cold War, or a hybrid model blending sovereignty with interdependence. To address these challenges, solutions must center decentralized governance, Indigenous knowledge integration, state-private-academic partnerships, and cross-cultural collaboration networks—each offering a path toward AI that serves global equity without reinforcing geopolitical divisions.