The global AI competition is shifting from U.S.-China rivalry to open-source versus closed-source models.

Huo Xing Finance reports that Fortune magazine recently published a commentary noting that as China’s open-source AI models rapidly rise, the global AI competitive landscape is shifting—the real contest has moved from U.S.-China geopolitical rivalry to the clash between open-source and closed-source models. Citing data from the independent evaluation platform Artificial Analysis, the article states that DeepSeek V4 Flash’s intelligence index is only 1 point lower than GPT-5.6 Luna, and even after OpenAI reduced prices by 80%, the former still offers a 60% lower cost per task. The author argues that while U.S. export control policies aim to restrict China’s AI development, they have inadvertently spurred Chinese companies to innovate at the architectural level, compelling them to adopt open-source, low-cost approaches. Chinese models attract global developers to improve them by releasing open weights, while shifting inference costs to overseas cloud providers, enabling a low-cost global deployment strategy. Regarding the notion that open-source models cannot be profitable, the article likens their business model to open-source software: users pay for hosted services, and most enterprises prefer not to self-deploy. The article also notes that several prominent figures in the U.S. tech sector—including former White House AI official David Sacks—have recently begun supporting open-source initiatives, and NVIDIA CEO Jensen Huang has called on the industry to embrace open-source. The author urges cooperation between China and the U.S. in AI safety, proposing that open-source AI be treated as a “global public good” rather than descending into a trillion-dollar zero-sum competition.