At the 2026 Cloud栖大会 (Yunqi Conference) in Hangzhou, Alibaba CEO Eddie Wu (Wu Yongming) announced a significant commitment to artificial intelligence infrastructure. Citing strong customer demand for AI, Wu declared that Alibaba Cloud will fully invest in AI infrastructure, with a goal to exceed 20GW in global data center scale by 2032.
Alibaba’s AI Ambitions Take Center Stage
The 2026 Yunqi Conference, which opened on September 22nd, highlights the growing importance of AI and cloud computing in the industry. Nvidia is a premier partner for the event, with Intel and AMD also securing top-tier partnerships. The conference features a strong lineup of domestic and international companies in cloud computing, AI, and smart manufacturing, emphasizing the critical role of computing power and the AI industry chain.
Alibaba Group CEO Eddie Wu and Liu Dayiheng, head of Alibaba’s Qianwen LLM project, are scheduled to deliver keynote speeches. A major focus of this year’s conference is the collaboration within the computing power ecosystem.
Rethinking Human and Machine Intelligence
Wu shared a compelling vision for the future of intelligence, stating that while machines currently handle 99.9% of physical work, human cognitive load will continue to increase. However, machine intelligence is expected to expand at a much faster pace, eventually undertaking 99.9% of cognitive tasks. He believes that advanced thinking capabilities will transition from being a luxury to a scalable commodity, enabling millions of AI scientists and experts across various fields. This shift, Wu posited, will bring about profound changes to human society.
Three Pillars of Machine Intelligence
Alibaba’s strategy is centered on three foundational pillars for the machine intelligence era: AI models, AI chips, and AI cloud services. Wu affirmed that Alibaba will steadfastly invest in these three areas, characterizing it as a long-term strategic choice for the company.
Full-Stack AI Development
Joining the discussion, Alibaba Group Chairman Joe Tsai (Cai Chongxin) noted that AI development has entered a new phase, moving from cutting-edge technological breakthroughs to the creation of scaled value in real-world applications. AI is already making significant inroads in autonomous driving, embodied intelligence, and world models, effectively boosting productivity across industries.
Tsai revealed that Alibaba is continuously strengthening its full-stack AI system, establishing a comprehensive five-layer architecture that spans chips, cloud models, model services, and intelligent agent applications. This integrated system aims to create a closed loop from fundamental technology research and development to widespread industrial implementation, accelerating the transformation of AI technological achievements into tangible industrial progress and fostering deep integration between AI and the real economy.
Ambitious Model Training Plans
Further underscoring its commitment, Alibaba plans to train an AI model with a parameter count ranging from 5 trillion to 10 trillion.
