China's top artificial intelligence companies are debating whether to continue using open-source models or shift toward closed-source development [1].
This tension is critical because the decision affects how China competes globally in AI. While open-source collaboration accelerates innovation through community contributions, closed-source models allow companies to protect proprietary intellectual property, and potentially monetize their technology more effectively.
Alibaba, which holds a market valuation of $300 billion [1], has begun experimenting with closed-source AI models. However, this shift may not reflect the broader trend within the domestic ecosystem. Catherine Thorbecke said, "Alibaba's shift towards closed-source AI models shouldn't be representative of the broader landscape in China just yet" [2].
Other domestic labs, including DeepSeek, remain heavily integrated into the open-source framework [1]. The broader ecosystem is supported by a strong community-driven development culture and lower costs associated with open collaboration [1]. Regulatory incentives in China have also made a wholesale move to closed-source models less feasible for many firms [1].
Industry perspectives on China's progress vary. Forbes said that China is winning the open-source AI race [3]. Conversely, Reuters said the situation is a growing dilemma for AI labs, suggesting that the path forward remains uncertain [1].
"To open-source or not may soon be the trickiest question facing China's top artificial intelligence companies," Reuters Breakingviews said [1]. The conflict pits the speed of collective development against the control of private ownership—a balance that will define the next phase of Chinese AI growth.
“"To open-source or not may soon be the trickiest question facing China's top artificial intelligence companies."”
The struggle between open and closed AI in China represents a pivot point for the global tech landscape. If China maintains an open-source trajectory, it may continue to lead in accessibility and rapid iteration. However, a shift toward closed models by giants like Alibaba suggests a growing desire for commercial moat-building, which could mirror the proprietary strategies used by leading AI firms in the U.S.
