Alibaba Confirms: Qwen to Power Apple Intelligence in China

Bottom line

A long-pending question is settled: Qwen will be the model behind Apple Intelligence in China, across iOS, iPadOS, macOS, and visionOS. An Alibaba spokesperson confirmed it.

Key facts

• Alibaba's spokesperson: 'For users in China, Qwen will be integrated into Apple Intelligence across iOS, iPadOS, macOS, and visionOS.'
• Regulatory side: China's CAC had listed the 'Apple Intelligence' model (via Apple's Shanghai entity) as registered, applicable to iPhones.

Why it's a big deal

For Apple: the key China-compliance piece

Apple's generative-AI features can't launch in China without a hard constraint—they must use a registered domestic model. Foreign models can't directly offer generative services in China. Qwen's integration is precisely the precondition for Apple's AI to 'launch compliantly' in China.

For Qwen: one of the largest distribution channels

• Apple has a huge active device base in China. Powering system-level AI means Qwen reaches a massive, default, system-level user entry point—a distribution scale no standalone app can match.
• System-level powering also brings depth of data and scenarios: from writing and summarization to cross-app actions, Apple Intelligence's surface is broad.

For the industry: another case of US-China AI 'stack splitting'

Apple uses one model in the US and Qwen in China—essentially regulation-driven 'AI stack splitting': one product, different compliant models per market. This is becoming the norm for multinational AI products.

Worth watching

• Feature parity: whether China's Apple Intelligence matches the overseas version or has features trimmed.
• Experience and boundaries: the system model's latency, privacy handling, and coordination with on-device capabilities will shape real UX.

For Chinese model makers, entering Apple's ecosystem is the largest distribution play; for Apple, it's the ticket to shipping AI in China.