China approves Apple Intelligence with Qwen
Beijing's Cyberspace Administration cleared Apple Intelligence for China using Alibaba's Qwen model after Apple's earlier talks with Baidu and DeepSeek stalled.
China’s Cyberspace Administration has approved Apple Intelligence for launch in the country, with Alibaba’s Qwen model handling the AI features across iOS, iPadOS, macOS, and visionOS. Alibaba describes the integration as covering “text and image understanding and generation.” Apple has not given a launch date.
The approval ends a long detour. Apple had explored partnerships with Baidu, DeepSeek, and ByteDance’s models before settling on Qwen, and the earlier talks reportedly stalled partly over how well those models adapted to Chinese-market requirements. The delay meant Apple Intelligence features that shipped in the US and other markets over a year ago have still not reached Apple’s China customers, its second-largest market: Greater China sales rose 28% to $20.5 billion in the most recent quarter even without them. Alibaba shares rose more than 6% following the announcement.
The arrangement matters beyond Apple’s own product roadmap. It is a concrete example of a major US consumer AI feature shipping on a Chinese-developed model inside China, the kind of pairing that export-control and data-sovereignty debates in Washington have been anticipating and, in some cases, trying to prevent. It also validates Qwen as a foundation capable enough to sit inside one of the most scrutinized consumer software stacks in the world, at a moment when Chinese officials have separately been weighing restrictions on overseas access to the country’s most advanced open models.
For operators running AI products across US and China markets, this is the clearest template yet for how that split gets handled: same product, same UI, a swapped model underneath tuned to local regulatory and data requirements. See open models outside your jurisdiction for the broader pattern of building around regional model constraints.