Doubao AI phone’s strong competitor is Xiaomi? Goldman Sachs: AI “system-level integration” faces challenges, which further validates Xiaomi’s long-term competitiveness.

Doubao AI phone’s strong competitor is Xiaomi? Goldman Sachs: AI “system-level integration” faces challenges, which further validates Xiaomi’s long-term competitiveness.

```

Author of this article: Long Yue

Source: Hard AI

Recently, ByteDance made a splash by launching the preview version of “Doubao AI Mobile Assistant,” aiming to reshape mobile interaction experiences with large language models (LLMs), drawing significant market attention.

Goldman Sachs’ latest research report released on December 4 analyzes that the launch of “Doubao AI Mobile Assistant” marks AI competition entering the operating system level. However, third-party AI agents face numerous challenges in obtaining system-level permissions, user data, and cross-app connectivity, further validating the long-term competitiveness of existing smartphone giants like Xiaomi.

Doubao AI arrives, but system-level integration faces three major challenges

According to the report, ByteDance’s “Doubao AI Mobile Assistant,” released on December 1, is a system-level graphical user interface (GUI) agent. It integrates the Doubao large model deeply into the mobile operating system and aims to realize visual interpretation of screen contents and complex cross-app task operations, such as cross-platform price comparison and automatic coupon collection as shown in the demo.

The report shows Doubao has become a popular large language model deployed by several original smartphone equipment manufacturers (OEMs) such as OPPO, vivo, and Honor. In May, the average daily tokens generated by hardware AI assistants on Doubao reached 1.3 trillion (accounting for 8% of Doubao’s total token consumption).

Although the technology demonstrates potential, Goldman Sachs points out that for third-party AI agents to achieve large-scale applications, three core challenges remain:

  1. System-level operational permissions: AI agents need permission to read screen content, simulate user behavior, and access system services. Currently, major smartphone OEMs still enjoy strategic advantages in the AI ecosystem and may not easily open their full system kernel and permissions framework to third parties.
  2. System-level memory capability: Major OEMs can access and store comprehensive user behavior and preference data, which is key to training and optimizing AI agents. Third-party AI is inherently at a disadvantage in this aspect.
  3. Cross-application interface connectivity: The cross-app operational capabilities of AI agents depend on the openness of third-party internet application interfaces. The report notes that some large companies aiming to build closed-loop ecosystems may restrict external agent use. Additionally, the frequent changes in app UIs (such as pop-up windows) also affect the stability of AI agents.

Goldman Sachs expects that in the foreseeable mid-term, value chain competition among consumer AI terminals (phone manufacturers), internet platforms, and third-party LLM/AI agents will persist.

Highly concentrated smartphone market, software disruption is extremely challenging

Goldman Sachs stresses in the report that a key premise to understanding AI’s impact on the smartphone industry is market structure. Data cited in the report shows that China’s smartphone market is highly consolidated, with the top six manufacturers (vivo, OPPO, Honor, Apple, Xiaomi, Huawei) accounting for more than 90% of shipment share.

This forms a sharp contrast with China’s new energy vehicle (NEV) market, which is much more fragmented and leaves room for new entrants with software differentiation (such as models powered by Huawei HIMA). In the smartphone sector, less than 10% of the market is occupied by niche brands and knockoffs, leaving minimal space for new players to disrupt the market.

This means the smartphone industry has extremely high competitive barriers, and the market positions of smartphone giants are relatively solid.

Xiaomi’s moat: vast ecosystem and deeply integrated AI agent

Goldman Sachs believes the challenges faced by Doubao AI precisely highlight the long-term advantages of companies like Xiaomi. Xiaomi is actively advancing its “People x Car x Home” full ecosystem strategy, with AI at its core throughout.

The report notes Xiaomi’s enormous investment in AI, projected to spend over 7 billion RMB on AI R&D in 2025, accounting for 22% of its total annual R&D expenses. The company has released professional LLMs covering vision, audio, speech, and more, and is committed to applying AI capabilities throughout its vast ecosystem—as of Q3 2025, Xiaomi has nearly 1 billion connected AIoT devices worldwide.

Xiaomi’s AI agent, “Super XiaoAI,” has achieved substantive progress. According to QuestMobile data, “Super XiaoAI” is among the top three OS-native AI assistants by monthly active users (MAU) in China, with a penetration rate as high as 71% among Xiaomi phone users (as of September 2025).

Through channel research, Goldman Sachs found that “Super XiaoAI” has achieved deep functional integration in several core scenarios:

  • Social media: Can post notes, send chat messages, like, and share content on platforms such as Douyin, Weibo, QQ (except for WeChat).
  • E-commerce: Can help users find goods or services all the way to selection/payment pages on nearly all mainstream platforms including Taobao, Meituan, JD, Pinduoduo, and Ctrip, plus confirm receipt and write reviews.
  • Productivity & content services: Supports image editing, real-time translation, text writing, and summarization based on users’ natural language input.
  • Memory capability: Can locally remember information, knowledge base, and user habits (including sensitive data, AIoT usage, medication, movie-watching times), for future application.

Goldman Sachs concludes that Xiaomi, leveraging its strengths in operating systems, hardware, massive AIoT ecosystem, and deeply integrated AI agents, has built a robust competitive barrier. This systemic advantage makes it hard for third-party AI agents to pose a substantial threat in the short term, instead verifying the long-term value of its “People x Car x Home” ecosystem strategy.

 

This article is from WeChat Official Account “Hard AI”. For more cutting-edge AI news, please visit here

Risk Reminder and DisclaimerThe market involves risks, and investment should be cautious. This article does not constitute personal investment advice, nor does it take into account users’ specific investment objectives, financial situation, or needs. Users should consider whether any opinions, viewpoints, or conclusions in this article are suitable for their particular situation. Investments made accordingly are at your own risk. ```