Apple Trains Its Own AI Model for China, Becoming the First Foreign Company Approved

Apple has trained its own large language model for the Chinese market with help from Alibaba, according to Reuters, making it the first foreign company to receive Chinese government approval to offer a proprietary AI model in the country. That represents a deliberate break from Apple’s previous approach of relying entirely on domestic Chinese models to bring generative AI features to its devices sold there.

Apple Foundation Models framework

Reuters describes the move as a “dual-track strategy,” with Apple maintaining its existing Alibaba Qwen integration while also developing its own China-specific model. Three sources described the new model as a conscious departure from depending on third-party options. How exactly the proprietary model will coexist with the Qwen integration is not clear yet, and Apple has not publicly disclosed implementation details.

Apple’s path to AI in China has been anything but straightforward. Chinese law requires foreign companies to partner with a domestic AI provider, forcing Apple into a completely different regulatory framework than the one it uses everywhere else. Apple originally lined up Baidu as its primary China AI partner in early 2025, but by February of that year Alibaba had stepped in as the lead after Baidu’s models reportedly fell short of Apple’s quality requirements.

By mid-July 2026, China’s Cyberspace Administration had cleared Apple Intelligence for launch, registering Apple’s on-device generative AI service on its list of approved providers. Alibaba reported that its Qwen model would power Apple Intelligence functions across iOS, iPadOS, macOS, and visionOS for Chinese users, covering both text and image generation. The Qwen integration requires macOS 26.6 or later, users must sign into their own Qwen account, and Alibaba is barred from using submitted materials to train or improve its own models.

A preview of the impending rollout came on August 8, when Reuters spotted a Chinese-language guide on Apple’s support site explaining how Mac users in mainland China could connect Qwen to Siri and Writing Tools. The guide was quickly removed, and no equivalent was published for iPhone or iPad, suggesting Apple is preparing deployment procedures ahead of a broader launch. Apple had published a separate iOS 26.6 and iPadOS 26.6 update earlier this year that began laying groundwork for iOS 27 performance features.

Domestic rivals like Huawei had AI features baked into their devices well before Apple could offer anything comparable in China, and Apple is still playing catch-up. That context makes the 24.4 percent year-over-year iPhone growth Apple recorded in China during the second quarter of 2026 all the more striking, with Apple ranking as the fastest-growing smartphone brand in a market that was otherwise contracting.

A functional, Apple-built AI model could help sustain that momentum. Apple’s overhauled Siri is scheduled to arrive publicly with iOS 27 this fall, though the exact timeline for China remains unconfirmed. No other U.S. Tech company currently holds an approved proprietary AI model in China, which gives Apple a regulatory foothold its competitors lack entirely.

About the Author

Imran Hussain is the founder and editor of iThinkDifferent, which he launched in 2008 to cover Apple news, reviews, and how-to guides. He has spent over 15 years writing about iOS, macOS, and the wider Apple ecosystem, with a focus on hands-on guides - installing developer betas, troubleshooting, and walking through new features on his own devices. Based in Dubai, he also loves to cover photography, gaming, and the tech industry more broadly on his social media profiles.

Leave a Reply