Alibaba develops larger AI models and new powerful chip
At Alibaba Cloud's Apsara conference, the company said its Qwen team plans models scaling to 5 trillion to 10 trillion parameters, up from the current 2.4 trillion-parameter Qwen 3.8 Max, and that it is training Qwen 4 with later Qwen 4.5 and Qwen 5 planned to scale. Alibaba also revealed the Zhenwu V900 AI chip, claiming three times the performance of its M890 predecessor and the ability to link up to 500,000 chips in clusters, with mass production slated for Q1 2027. The firm set a goal for over 20 gigawatts of global data centre capacity by 2032 and reported robust AI demand, while noting supply chain limits.
Alibaba plans models of up to 10 trillion parameters.
Context
Chinese technology firms are racing to build local AI hardware after tighter US export curbs. Alibaba is expanding models, chips, and data centres to reduce reliance on foreign processors. Next steps include training Qwen 4 and bringing…
The full analysis
19 dimensions on this story — world impact, market read, and what happens next.
- Full ContextLocked
- Affected SectorsLocked
- Stock ImpactLocked
- Economic IndicatorLocked
- Investor RelevanceLocked
- Professional RelevanceLocked
- Watch PointsLocked
- Probability of ChangeLocked
- Debate PointsLocked
- Prerequisite KnowledgeLocked
- Follow-up QuestionsLocked
- Pros & ConsLocked