Artificial Intelligence (AI) is increasingly transforming core sectors of the global economy, encompassing semiconductors, cloud infrastructure, and digital finance. However, instead of deepening international cooperation, AI has accelerated strategic competition between the two top economies and tech leaders, the United States and China. Drawing on neorealist theory, this study examines how AI, as a capability multiplier and a form of relative power, is utilised by states to enhance their tech advantages in economic and security arenas, thereby creating a techno-geoeconomic security dilemma that enables the US and China to leverage each other. It explores AI’s relative capability concerns, prompting balancing measures, such as trade restrictions, technology blocks, and domestic technological initiatives. The tech oligarchs in Silicon Valley’s American Silicon Valley monopolise AI, data, and related technologies in pursuit of global hegemony. China, amid US export control bans, is developing its indigenous technologies (Chips, LLM models, supercomputing, etc.) through self-help efforts to ensure technological sovereignty. This study employs process tracing for policy analysis, explaining strategic preferences and sector-specific dimensions of competition, providing concrete empirical evidence across defence and geoeconomic domains. Findings reveal that the US uses semiconductor and AI chip export controls, financial chokepoints (such as the SWIFT payment system), and tech alliances to maintain its global pre-eminence. In contrast, China, as part of its balancing act, responds by creating e-CNY and CIPS as means for BRI partners. It concludes that such competition is driven less by economic prosperity or absolute gains than by relative gains and tech sovereignty to secure structural advantage in an anarchic international system. Bibliography Entry Bashir, Sonia and Zahid Mahmood Zahid. 2025. "Artificial Intelligence as New Capital: Geoeconomic Competition and Reshaping of Global Order." Margalla Papers 29 (2): 33-53.
Bashir et al. (2026) studied this question.