Alibaba pushes open-source AI as key to widespread adoption, highlights Qwen model success

At the 2026 World Internet Conference in Wuzhen, Alibaba co-founder and chairman Joseph Tsai announced that the company has released over 470 Qwen models, achieving cumulative global downloads surpassing 4 billion, alongside more than 330,000 derivative models. Tsai asserted that open sourcing is essential for making AI accessible to the masses, noting it can lower research and development costs, particularly benefiting smaller enterprises and developing nations. He likened Alibaba's open approach to a top university offering global access and financial aid. Furthermore, Tsai previously advised Europe, according to an October Bloomberg report, that embracing open-source AI and building domestic computing infrastructure is the optimal route to reduce reliance on the US and China for computational power. He projects that within five years, AI will become invisible infrastructure, deeply integrated into business, a trend he believes open source is key to enabling by releasing AI capabilities beyond a few large data centers.