Apple and Alibaba's AI collaboration has entered a substantive phase. The Mac user guide on Apple’s China website now features a page titled “Use Qwen with Apple Intelligence on Mac.” Qwen is not offered as a standalone app but is embedded as one of the foundational models within system-level AI interfaces such as Siri and writing tools in the China-specific version of Apple Intelligence. Apple has thus established a new AI supply chain model: maintaining control over its operating system and hardware interfaces, while selecting local suppliers for foundational models based on regional markets. For Alibaba, this means gaining a significant distribution channel for its models, enabling millions of Apple devices to access them through integrated system functions. For Apple, the delay in launching Apple Intelligence in China prompted the search for a local partner to accelerate development, potentially paving the way for a global landscape where different regional markets adopt distinct local models.

Article author and source: Wall Street Journal

Apple and Alibaba's AI collaboration is moving from rumors and regulatory filings into Apple's operating system.

Recently, Apple China's official website has added a prominent new page to the "Mac User Guide" titled "Using Qwen with Apple Intelligence on Mac."

According to Apple's documentation, Qwen is not introduced to the Mac as a standalone AI application, but is directly integrated into the core use cases of Apple Intelligence.

For example, when Siri cannot directly fulfill certain requests, users can further invoke Qwen to obtain more in-depth responses, including analysis of content related to photos and documents; in writing tools, users can also leverage Qwen to generate text or images based on textual descriptions.

In other words, Qwen is one of the external large models behind China's version of Apple Intelligence.

This also makes Apple's AI strategy in China more concrete for the first time.

From Apple partnering with Alibaba to Qwen truly integrating into the system

The collaboration between Apple and Alibaba has long been in the making.

In July this year, Reuters reported that Apple Intelligence’s generative AI services have completed registration with Chinese regulators. To meet regulatory requirements in the Chinese market, Apple has partnered with Alibaba and Baidu, integrating Alibaba’s Qwen model into Apple devices, covering iOS, iPadOS, macOS, and visionOS.

The Wall Street Journal previously reported that Apple began advancing related regulatory procedures in early 2025 and is collaborating with Alibaba and Baidu to develop certain AI features tailored for the Chinese market.

But at the time, what the outside world knew more about was “who partnered with Apple.”

The recent update to Apple's official manual for the first time specifies this collaboration at the user experience level:

Siri can invoke Qwen, and writing tools can also invoke Qwen.

This distinction is important.

This means that Qwen may not just be gaining a position as a pre-installed app, but rather entering Apple's most core system-level AI entry point.

Users don't even need to actively open the Qwen app.

When they ask Siri complex questions, analyze documents, or use AI writing tools, some of the underlying model capabilities may come from Qwen.

For Alibaba, the value of this distribution method far exceeds simply having users download another AI app.

Apple is building a "localized AI supply chain."

More noteworthy is that Apple is developing an AI model that is entirely different from its past approach.

In the era of traditional smartphones, Apple was accustomed to controlling the core software experience itself: designing its own chips, developing its own operating system, and creating its own Siri.

However, with the advent of the generative AI era, large models are being updated too rapidly, training costs are too high, and different countries have entirely different data and regulatory requirements.

Apple is designing Apple Intelligence as a more layered system.

At the core, Apple still relies on Apple Silicon and its on-device models to handle a large volume of personalization tasks. Apple continues to position the Mac as a device “built for AI,” emphasizing on-device AI processing and privacy protection.

However, when faced with more complex issues requiring stronger model capabilities, the system can invoke external models.

In overseas markets, Apple has previously adopted a similar approach by integrating third-party model capabilities into Apple Intelligence.

In the Chinese market, this role is taken on by domestic Chinese AI companies.

This means Apple may be building a new "AI supply chain": controlling the operating system, hardware, user access points, and experience, while selecting different suppliers for the underlying large models based on different markets.

You don't necessarily have to build everything from scratch, but you must control the entry point.

This may be the real aspect of Apple's AI strategy worth paying attention to.

Qianwen has secured a position more significant than "being installed on an iPhone."

For Alibaba, this partnership is equally significant.

Over the past year, competition among large models in China has become increasingly intense.

Companies are not only competing on model capabilities but also vying for genuine user access points.

Alibaba's own Qwen app is continuously adding features such as office productivity, lifestyle services, AI agents, and image and video generation, aiming to directly compete with consumers.

But Apple offers a completely different entry point.

Instead of prompting users to remember, “I need to use AI,” and then opening an AI app, integrate AI directly into the operating system features users are already using.

Siri, writing tools, and even more system-level functions in the future could all become distribution channels for large models.

This is equivalent to transforming the large model from an independent product into a foundational capability underlying the operating system.

More importantly, Qwen has recently continued to rapidly expand its model capabilities. The Financial Times of the UK, in its analysis of Alibaba’s latest model, Qwen 3.8-Max, noted that Alibaba is striving to transform AI into a new technological and business growth engine, with Alibaba Cloud playing a crucial role in commercializing these model capabilities.

Therefore, the significance of Apple's partnership with Alibaba goes beyond merely "gaining another major client."

It is actually a very important model distribution.

If tens of millions of Apple devices in China can invoke Qwen through system features in the future, Alibaba will gain not only invocation volume but also an extremely important industry endorsement:

Its large model is already capable of serving as the underlying AI provider for one of the world's largest consumer electronics ecosystems.

For Apple, this is even more of a long-overdue lesson.

Apple actually needs this partnership more than Alibaba.

After Apple Intelligence was launched globally, the delayed rollout in mainland China has gradually become a noticeable disadvantage in Apple's product competition.

Meanwhile, Chinese smartphone manufacturers such as Huawei, Xiaomi, and OPPO have rapidly integrated generative AI, intelligent agents, and system-level AI capabilities into their phones.

The Wall Street Journal previously noted that the delay of Apple Intelligence in the Chinese market has left Apple without a capability that is becoming increasingly important in its competition with local Chinese smartphone manufacturers.

This is also why Apple cannot simply transplant the overseas version of Apple Intelligence into China unchanged.

The model, data processing, and regulatory framework all need to be re-adapted.

The ultimate solution is to find a local partner in China.

Thus, this interesting combination has emerged:

Apple provides devices, chips, operating systems, and entry points, while Chinese companies like Alibaba offer local large model capabilities.

Each party contributes their strongest areas.

In the AI era, Apple's most important asset may not be having the strongest model.

Behind this issue, there is actually a larger problem:

In the AI era, must a tech company possess the world's most powerful large model?

Apple's current answer may be no.

Over the past two years, companies such as OpenAI, Google, Anthropic, and Alibaba have continuously pushed the boundaries of model capabilities, but the gaps between models are also evolving rapidly.

For Apple, with its ecosystem of over a billion devices, an alternative approach may be more practical than investing all resources in the model arms race:

Master user access points and select the model best suited for different markets.

In the United States, you can access U.S.-based models.

In China, you can integrate local models such as Qwen.

Apple continues to control its chips, operating system, privacy architecture, and the overall user experience.

If this path ultimately succeeds, what truly matters in the future may not just be who owns the strongest model, but who gets to decide when hundreds of millions of users invoke which model.

From this perspective, the seemingly ordinary line on Apple’s website—

Get more answers from Siri: For certain requests, Siri can leverage Qwen to provide more in-depth responses.

The underlying significance is much greater than "Apple integrating Qwen."

It means Chinese AI companies have, for the first time, truly entered Apple’s system-level AI gateway.

For Qwen, this may be even more significant than appearing on the iPhone—it is now beginning to appear behind Siri.