Work

"Breakthrough via Open Source" and the Reconfiguration of Structural Power: The Asymmetric Strategic Logic of US-China Artificial Intelligence Competition

Year: 2026
Type: preprint
Abstract: Drawing upon structural power theory, asymmetric competition, and the perspective of moral realism, this study systematically explores the strategic logic behind the U.S.-China rivalry in the field of artificial intelligence and its implications for international politics. The research reveals that the significant divergence in the AI development trajectories of China and the United States is, at its core, a confrontation between two distinct logics of structural knowledge power. The United States leverages its closed-source large language models and technological containment to construct a "black-box hegemony" that sustains its traditional technological monopoly, whereas China, constrained by limited computational resources, actively promotes the decentralization of knowledge production through an open-source strategy, embarking on an "asymmetric breakthrough." This study systematically demonstrates the deep theoretical isomorphism between China's open-source strategy and Mao Zedong's "surrounding the cities from the countryside" strategy. In terms of spatial strategy, China's open-source model avoids the "core market" dominated by U.S. closed-source giants, instead focusing on the "vast rural" computational resource landscape—comprising countries of the Global South, small and medium-sized enterprises' long-tail applications, and edge-side deployment; in terms of power dynamics, it effectively mobilizes millions of developers worldwide through its open-source ecosystem to engage in distributed innovation, thereby creating an asymmetric advantage in both scale and efficiency over the handful of leading laboratories constituting the "regular force"; in terms of the encirclement pathway, it follows a gradual approach— "edge customization → capturing the middleware ecosystem → challenging the foundational model layer" —to achieve the strategic objective of redefining the rules of the game through its ecosystem. The article concludes with a discussion on the governance challenges and the impact of these two strategies on the international landscape. The closed-source model is confronting structural dilemmas, including the widening global digital divide, growing concerns over "AI sovereignty," and the collapse of the "security myth"; whereas the open-source model, while possessing a moral supply advantage, also faces risks associated with the "tragedy of the commons," such as malicious fine-tuning and technological runaway. The study ultimately concludes that the Sino-US technological decoupling appears absurd when viewed through the lens of open-source principles; instead, the future global AI order will likely not follow a zero-sum, eliminative trajectory, but will instead be driven by a layered interplay between top-level closed-source approaches and mid-to-bottom-level open-source initiatives, fostering an evolution of the global AI landscape from a unipolar hegemony toward a more inclusive, multipolar governance framework.
Language: English
Open Access status: green
OpenAlex ID: W7235574140
Cites: 26
Cited by:
Related to: 10