Cook Confirms Apple Intelligence to Integrate Alibaba's Qwen, China Beta Launch at End of October

Apple Intelligence's China Plan Takes Shape

According to an RSS source, Apple Board Executive Chairman Tim Cook confirmed during a media interview in Shanghai that Apple Intelligence and the related Alibaba Qwen large model have received approval and will launch a beta version in China in late October. This news marks substantive progress for Apple in AI localization compliance and deployment.

Cook revealed in the interview that the first batch of features includes real-time translation, writing tools, and photo cleanup tools, among other core capabilities. However, the highly anticipated Siri AI is not included in this initial release. Cook was positive about this, stating: "But we're very excited that the first batch of core Apple Intelligence features can gradually roll out." This phrasing suggests Apple has adopted a phased approach, prioritizing stable availability of foundational features.

For Chinese users, the arrival of Apple Intelligence means the iOS ecosystem will finally gain a system-level AI experience, while the integration of Alibaba's Qwen large model addresses both data compliance and localization service requirements.

Why Alibaba's Qwen Became Apple's Choice

Apple's selection of Alibaba's Qwen as its large model partner in the Chinese market is no coincidence. From a technical standpoint, the Qwen series models are highly competitive in Chinese language understanding, multi-turn dialogue, and instruction following, and Alibaba Cloud has mature computing infrastructure and compliance qualifications in China.

From a compliance perspective, China implements a registration management system for generative AI services, and overseas models face clear restrictions for direct deployment. By partnering with Alibaba, Apple can both meet regulatory requirements and leverage Qwen's Chinese language capabilities to enhance Apple Intelligence's real-world performance in Chinese scenarios. This approach echoes Apple's dual-track strategy in overseas markets of proprietary models plus OpenAI collaboration.

You may not have noticed that Cook proactively mentioned the term "approved" during his media engagement in Shanghai, indicating that the partnership has passed the necessary regulatory approval process. For developers and users following Apple's AI progress, this means Apple Intelligence in China is no longer a question of "when it will launch" but rather "how well the first batch of features will perform."

Generative AI Registration Management is a regulatory system China implements for generative artificial intelligence services provided to the public. Under the "Interim Measures for the Administration of Generative Artificial Intelligence Services," any AI service that provides generated content such as text, images, audio, or video to the domestic public must pass registration review by local cyberspace authorities, and those involving public opinion attributes or social mobilization capacity must also undergo security assessments. Registration requires service providers to clarify the source of their models, training data compliance, content safety mechanisms, and user information protection measures. For foreign companies like Apple, directly deploying proprietary or overseas models presents obstacles. By partnering with an already-registered domestic large model (such as Alibaba's Qwen), they can leverage existing registration qualifications to quickly complete the compliance pathway, avoiding a lengthy duplicate approval process. This is the institutional meaning behind Cook's mention of "approval."

First Batch Features: Practicality First, Siri Absent

According to Cook's introduction, the first batch of Apple Intelligence features focuses on three high-frequency scenarios:

Real-time Translation — covering cross-language communication needs, callable in system apps like Messages and Mail. This aligns with the experience direction of current mainstream AI translation tools, but Apple's advantage lies in system-level integration.

Writing Tools — providing text polishing, rewriting, and summarization capabilities, directly embedded into iOS and macOS input scenarios.

Photo Cleanup Tools — using generative AI to remove unwanted elements from photos, representing a lightweight AI application for image editing.

Siri AI's absence is noteworthy. As the most imaginative part of Apple Intelligence, Siri's AI rebuild involves deeper system permissions and model invocations, with higher technical complexity and compliance assessment requirements. Choosing to "bypass Siri" for the first batch of features reduces launch risk on one hand and leaves room for subsequent iterations on the other.

Siri AI refers to the large-scale AI overhaul of Siri announced by Apple at WWDC 2024, which plans to inject large language model-based contextual understanding, cross-app operations, and personalized semantic capabilities, upgrading it from a traditional voice assistant to an intelligent agent with reasoning and execution abilities. Compared to single-point features like real-time translation and writing tools, Siri's rebuild requires invoking more system-level permissions (such as accessing calendars, emails, and third-party app data) and deep coordination with on-device and cloud models, thus facing higher thresholds in privacy protection, cross-border data transfer, and compliance assessment. Apple's choice to defer Siri AI in the first batch of China version features is a prudent approach under dual considerations of technical maturity and regulatory risk.

Strategic Considerations for Phased Rollout

Apple's approach of "beta version first, core features rolled out in batches" reflects several practical constraints. China's AI regulatory framework requires security assessments before service launch, and a beta version allows feedback collection within a controllable scope. Meanwhile, Apple needs time to verify the deep integration of the Qwen model with the iOS system, especially in terms of response speed and privacy protection.

From a competitive standpoint, domestic smartphone manufacturers have already completed multiple rounds of iterations on on-device AI features, and Apple Intelligence's late arrival puts it under pressure to catch up. However, Apple's brand appeal and system-level experience advantages remain key differentiators. Whether the first batch of features can build a good reputation through experience will directly impact user expectations when Siri AI eventually launches.

For developers, Apple Intelligence using Alibaba's Qwen model in China may mean that API-level capability invocation differs from overseas versions. This requires developers to pay attention to updates in Apple's official documentation when adapting their apps.

On-device AI refers to running AI models directly on the device locally (such as a phone chip's NPU), completing inference without uploading data to the cloud. Its advantages include fast response times, strong privacy, and offline availability, but limited by device computing power and memory, it can typically only run models with smaller parameter counts. In contrast to on-device AI is cloud AI, which can invoke larger-scale models for complex tasks but involves data transmission and compliance issues. Apple Intelligence adopts a device-cloud collaborative architecture: simple tasks are processed locally, while complex tasks are uploaded to the cloud or partner models. In the Chinese market, the cloud portion is handled by Alibaba's Qwen, which poses new design challenges for Apple regarding how to divide on-device versus cloud tasks and how to ensure data compliance.

What It Means for Users and the Industry

For ordinary users, the arrival of Apple Intelligence's China version means iPhones will gain system-level AI capabilities for translation, writing, and photo editing without relying on third-party apps. The beta release also signals that the official version will be gradually refined over the coming months.

For the industry, Apple's partnership model with Alibaba provides a referenceable example for international tech companies deploying AI services in China: technology partnership plus compliance registration plus phased release. Whether this model can be replicated by other manufacturers depends on each company's resource investment in model selection and regulatory communication.

Cook's Shanghai visit sent a clear signal — Apple Intelligence's advancement in the Chinese market has entered the execution phase. The launch of the first batch of features is just the starting point; Siri AI's eventual arrival is the real test.

One-line Summary

Cook confirms Apple Intelligence's integration with Alibaba's Qwen has been approved, with a China beta launching in October; first batch of features excludes Siri AI.

Paragraph Summary

Apple Board Executive Chairman Tim Cook confirmed in Shanghai that Apple Intelligence and Alibaba's Qwen large model have been approved and will launch a beta version in China in late October. The first batch of features includes real-time translation, writing tools, and photo cleanup tools, with the highly anticipated Siri AI not included for now. Apple chose Alibaba's Qwen as its partner primarily based on its Chinese language capabilities, computing infrastructure, and compliance qualifications to meet China's generative AI registration requirements. This partnership model balances technical deployment with regulatory compliance, adopting a phased rollout strategy to collect feedback and reduce risk through a beta version first. For users, iPhones will gain system-level AI capabilities; for the industry, this model provides a referenceable example for international tech companies deploying AI services in China.