On July 15, a significant milestone was achieved in China's AI smartphone industry. The Cyberspace Administration of China (CAC) unveiled the filing details for seven services that offer on-device generative AI capabilities for mobile phones. These services include Apple Intelligence, Huawei's Xiaoyi AI Large Model, Xiaomi's Surge AI, Samsung's Galaxy AI, OPPO's AndesGPT, vivo's BlueLM On-Device Large Model, and Nubia's Doubao Mobile Large Model.

On the same day, the Interim Measures for the Administration of Human-like Interactive AI Services were officially put into effect. The simultaneous occurrence of these two pivotal events marks China's on-device AI transitioning from a phase of "unregulated growth" to a new era of "licensed operation."

As of June 30 this year, a total of 988 generative AI services across the nation have completed the filing process. However, this marks the first instance where a distinct list has been formulated specifically for on-device services. Numerous industry experts believe that the increased policy certainty has finally alleviated the industry's concerns, while also hinting at intensified competition at the product level.

01. Filing Implementation: On-Device AI Says Goodbye to 'Unlicensed Operation'

The most notable distinction in this filing compared to previous ones is the term "on-device." Previously, generative AI service filings mainly targeted cloud-based large models and standalone applications. On-device large models, which are deployed locally on devices, generate content without relying on any servers, rendering traditional cloud-based review and interception mechanisms ineffective. This specialized filing formally establishes an independent regulatory framework for the emerging phenomenon of "AI residing in mobile phones."

The creation of a dedicated filing category for on-device mobile AI underscores the regulatory authorities' focus on this burgeoning sector. Since ByteDance and ZTE jointly unveiled the first engineering machine equipped with an AI smartphone assistant at the end of 2025, AI smartphones have become a focal point throughout the industry. Domestic smartphone manufacturers have accelerated their deployments, while overseas tech giants and smartphone brands have also announced plans to launch AI smartphones or related features. Meanwhile, controversies surrounding AI smartphones persist. Many legal experts point out that traditional social governance rules face new challenges in the AI era. The fact that on-device large models can generate content without passing through the cloud significantly complicates the review process, highlighting the urgent need to regulate privacy and security issues.

For domestic manufacturers such as Huawei, OPPO, vivo, Xiaomi, and Nubia, the filing confirms that their technical approaches comply with national regulations, enabling them to expedite product launches within a compliant framework. For international brands like Apple and Samsung, the filing holds equal significance—regardless of their origin, companies that adhere to China's rules can gain legitimate operation qualifications in the world's largest consumer market.

02. Apple's AI Localization Collaboration Comes to the Fore

The approval of Apple Intelligence in this filing list has garnered particular attention. Since its initial announcement at the 2024 Worldwide Developers Conference, this AI feature, specifically tailored for the Chinese market, has undergone over a year of preparation and has finally secured the qualifications for compliant operation. This means that the long-awaited AI functions for iPhone users in China are now just one step away from an official launch.

Following the completion of the filing, Apple's roster of AI partners in China has also been fully disclosed. Alibaba confirmed that its Tongyi Qianwen large model will be seamlessly integrated as the underlying AI capability for Apple Intelligence, serving iOS, iPadOS, macOS, and visionOS users in China. In the future, users of Apple devices in China will be able to directly invoke the Qianwen large model at the system level without switching between applications, enjoying native intelligent experiences such as text and image understanding and content generation. Meanwhile, Baidu will also play a significant role in the localization of Apple's AI, focusing on developing AI-powered search functions to comprehensively upgrade the search and Q&A experience of the Chinese version of the Siri voice assistant.

Industry analysts believe that by collaborating with Alibaba's Qianwen and Baidu, two major domestic AI players, Apple not only meets regulatory requirements for data security and content compliance but also effectively addresses the shortcomings of iPhones in China regarding Chinese-language search capabilities and underlying large model performance.

03. Accelerated Penetration Rate Unlocks Scalable Space for a Trillion-Yuan Market

With policy support, the pace of AI smartphone scalability is accelerating comprehensively. Third-party institutions predict that models supporting generative AI will account for 45% of global smartphone shipments by 2026 and are expected to reach 52% by 2027, with generative AI poised to become a standard feature across the entire smartphone industry. The market size for on-device AI is projected to surge from RMB 321.9 billion in 2025 to RMB 1.22 trillion in 2029, representing a compound annual growth rate of 40%.

Meanwhile, the 2026 World Artificial Intelligence Conference will be held in Shanghai from July 17 to 20. According to official disclosures, the event will feature the global debut of several top-tier AI achievements, including a consumer-grade terminal product labeled as the "world's first AI agent smartphone." Market speculation suggests that this product may be the mass-produced second-generation Doubao smartphone, developed through deep cooperation between Nubia and ByteDance. The first-generation engineering machine generated significant market enthusiasm, with its entire stock of 30,000 units selling out immediately upon launch and second-hand market prices surging by over 40%, validating strong user demand for AI-driven proactive service smartphones. The new model will feature a complete Doubao AI system, enabling cross-application and cross-platform autonomous operation, marking the smartphone's evolution from a "passive tool" to a "proactive intelligent assistant."

However, industry insiders also acknowledge that AI smartphones currently lack truly killer applications, and commercialization remains in its early exploratory stages. The transition from "filing implementation" to "mass scalability" will require time for validation. Leading industrial chain companies are also closely monitoring how AI features drive demand for device upgrades, with the true inflection point for scalability potentially awaiting the emergence of more groundbreaking application scenarios.

04. Systematic Industrial Chain Upgrade: From Chips to Thermal Management

AI smartphones impose systematic hardware upgrade requirements. The local deployment of on-device large models raises technical thresholds for computing power, storage, sensing, and thermal management, presenting iterative innovation opportunities across various industrial chain segments.

In terms of data security and biometric recognition, on-device AI demands far higher privacy protection than traditional smartphones. Multiple terminal manufacturers have collaborated with chip and sensor companies to establish a hardware security foundation, while the AI software security sector has also become a hotspot for capital attention. With on-device large models generating content without passing through the cloud, the importance of local security chips and encryption technologies in devices continues to grow.

Regarding thermal management and temperature control, influenced by the trends of thinner, lighter, and foldable smartphones, AI phones face far greater unit heat dissipation pressures than traditional models. Large model inference requires higher sustained computing power output, leading to increased chip power consumption that traditional cooling solutions can no longer meet. Novel thermal management materials, represented by graphene thermal films and vapor chambers, are seeing rising penetration rates, with their market space expected to expand rapidly.

In optical imaging, AI smartphones fundamentally reshape the strategic role and hardware specifications of cameras. Three major trends—professionalized shooting, stereoscopic perception, and deepened software-hardware collaboration—are driving a comprehensive upgrade of imaging systems. AI algorithms impose new requirements on hardware metrics such as sensor pixel count, aperture size, and stabilization precision, causing the average cost of smartphone camera modules to continue rising.

From an overall industrial landscape perspective, amid the year-on-year decline in global smartphone shipments, AI capabilities are becoming the core differentiator for manufacturers to achieve a competitive edge. With the establishment of a specialized filing system for on-device AI, a standardized development pathway is taking shape across the entire industrial chain, spanning chips, sensors, device manufacturing, and software ecosystems. Despite persistent challenges, the implementation of policy frameworks and the collective preparedness of the industrial chain are paving the way for large-scale adoption. The first year of AI smartphones may indeed be arriving.

- End -