Alibaba has unveiled Qwen3.8-Max, a flagship large language model with approximately 2.4 trillion parameters, positioning it among the world's most advanced AI systems. The preview launch at Shanghai's World AI Conference pushed Hong Kong-listed shares up 5.4% on Monday, marking the second rally in a week after Apple secured Chinese approval to integrate Alibaba's Qwen models into iPhones for on-device AI. The Apple partnership is expected to deepen cloud demand and unlock recurring AI revenue, though much of the earlier Apple-driven stock gain had faded.
The Qwen3.8-Max preview is accessible via Alibaba's Token Plan subscription and developer tools Qoder and QoderWork. The company reports that cloud revenue grew 38% to 41.63 billion yuan, with AI products now contributing 30% of external cloud sales. CEO Eddie Wu stresses that tech spending is starting to yield commercial returns, but heavy investments in AI infrastructure and quick-commerce are pressuring adjusted EBITA, raising profitability concerns for shareholders.
While Alibaba races ahead, Beijing-based startup Moonshot AI launched its own 2.8-trillion-parameter model Kimi K3, claiming it beats GPT-5.5 and Claude Opus 4.8 on benchmarks. Alibaba itself faces U.S. friction: the Pentagon added it to a military-linked blacklist in June, and Anthropic has accused the company of a “distillation” attack—improperly training on Claude AI outputs, which Anthropic calls the biggest such incident it has encountered.
