Alibaba Cloud maps AI stack and new cloud regions

Alibaba Cloud unveiled a full-stack AI roadmap at its Apsara Conference, covering chips, cloud infrastructure, models and agents, while setting a goal for global data center capacity to surpass 20GW by 2032. Chairman Joe Tsai framed the strategy around moving AI from technical breakthroughs into business value, and CEO Eddie Wu said the company sees a large growth runway for “Machine Thinking.”

The company said Qwen 4 is in training and outlined future Qwen 4.5 and Qwen 5 model series projected to scale up to 5 to 10 trillion parameters. Alibaba also introduced updates across speech, audio and vision, including Qwen3.8-LiveTranslate, Qwen-Audio-3.1-TTS-Next and the forthcoming Qwen-Image 3.1, alongside Qwen Intelligence, a smartphone-focused agentic platform for phone makers.

T-Head introduced the Zhenwu V900 AI training and inference processor, scheduled for mass production and commercial release in the first quarter of 2027. The company said the chip offers 216 GB of GPU memory and 1,200 GB/s of inter-chip bandwidth, and it showed an upgraded supernode server supporting a cluster of up to 500,000 cards.

Alibaba Cloud will open its first cloud regions in Türkiye, Finland and the Netherlands over the next 12 months, while expanding data center capacity in Malaysia, Germany, the United Arab Emirates, France and Hong Kong. New international tools include Smart Studio for Model-as-a-Service platforms, Smart Fusion for multi-model deployment and Smart Video for AI-generated video production.