Alibaba Chairman Joe Tsai said at the World Internet Conference in Wuzhen, Zhejiang province, on the 3rd that open source is effectively the only path to sustainable growth for the artificial intelligence industry, according to the South China Morning Post on the 4th, as the Chinese tech giant accelerates its push to expand global influence through an open ecosystem — the opposite of the approach taken in the U.S. “Only through open source can AI become mainstream,” Tsai said, adding that “only through widespread adoption can we build a sustainable and inclusive age of intelligence.” Alibaba’s in-house AI model Qwen now has more than 470 open-source versions, and the company has doubled computing-related capital expenditure each year over the past three years to support its open-source strategy. Tsai said a consensus around open-source ecosystems has been spreading among Chinese technology companies, and that open-source models can sharply lower research and development costs for small and medium-sized firms and for developing countries as AI demand surges while computing resources remain scarce. The Chinese government is also actively supporting open-source AI as part of its technological self-reliance strategy; DeepSeek and MiniMax recently won regulatory approval to release and license their open-source systems overseas, and authorities allowed leading domestic AI firms to distribute model weights abroad even amid moves to tighten export controls on AI-related technology. That stands in contrast to major U.S. AI companies: OpenAI and Anthropic have stressed a cautious approach, warning that high-performance open-weight