Recently, QuestMobile, an authoritative third-party data agency in China, released the "Core Report on the Development of AI Application Layer in 2025". The report comprehensively reviews the evolution of the AI application market in China over the past year, pointing out that in the context of increasing homogenization of underlying technologies, the AI competition paradigm has shifted from "model internalization" to "ecological expansion".

It is worth noting that, compared with the independent native App track that has attracted great attention from the outside world in the recent "Red Packet Battle", the form of "In App AI" dominated by the head Internet platform is becoming the core main force carrying the daily AI experience of Internet users with its huge user base and natural scene advantages. Among them, Weibo Zhisou's active user base exceeded 66 million in December 2025, with a compound growth rate of 18.1%. Among the top 10 application plugins in terms of monthly active scale, Weibo Zhisou ranks first in terms of compound growth rate.

The activity of four types of AI applications has comprehensively increased, and the In App AI market is approaching 700 million

According to QuestMobile data, as of December 2025, the activity of four types of AI applications in China has comprehensively increased. Among them, the user base of "In App AI" is as high as 696 million, making it the AI form with the widest user reach currently; And native apps maintain a strong compound growth rate.

The differentiation in application forms reflects the industry's "multi track" strategy in AI entrance layout. The report points out that native apps are positioned by multiple platforms as the future 'gateway' and 'platform layer'. In App AI has become the optimal solution for consolidating the "scene entrance" and "functional layer". It deeply integrates AI capabilities into high-frequency social, e-commerce, and information business flows, improving efficiency and retention in high-frequency applications.

The data confirms the efficiency of plugin forms in cultivating user habits: in December 2025, the monthly average usage of application plugins increased by 229.8% compared to January, far exceeding the 80.1% increase of native apps. This shows,In familiar existing scenarios, "easily calling AI" is becoming a high-frequency and stable user habit.

Breakthrough in AI search dominated track: Weibo's intelligent search has over 66 million monthly active users, leading the compound growth rate in the same volume category

Among all AI landing scenarios, "search" has been validated as the most mature and high-frequency entry point, and will face high competition in 2025. According to the report, there are a total of 6 applications belonging to the AI search engine track in the "Top 10 Application Plugin with Active User Scale in December 2025".

Among which,Weibo Zhisou ranks 8th on the list with a monthly active user base of 66.35 million.What is even more impressive is its growth momentum: among the top 10 application plugins with active user scale in December, Weibo Zhisou's compound growth rate reached 18.1%, ranking first in terms of growth rate, and its user scale ranking also jumped 15 places, showing an extremely strong breakthrough momentum among products of the same level.

In this regard, the report points out that AI Weibo intelligent search has broken through with high growth rate, which is enough to prove that even in mature tracks and finding the right niche, there is still a chance to break the original pattern and achieve a ranking leap.

From model internalization to ecological barriers, plugins build strategic defense walls

The report points out that with the maturity and increasing homogenization of underlying technologies, the paradigm shift of AI competition in 2025 will shift from "model internalization" to "ecological expansion", and the key to enterprise competition and business success will directly point to products, experiences, ecosystems, and user reach capabilities. Simply comparing technical parameters has lost its significance, and the ability to grasp real industry data and user scenarios has become the key to winning the landing of AI applications.

In this trend, an irreplicable data ecosystem has become the key to creating differentiated growth advantages. Taking Weibo as an example, as the largest social media platform in China, Weibo aggregates a massive amount of real-time celebrity voices, frontline media information, and user generated discussions (UGC). This dynamically updated 'live water data' constitutes the core underlying barrier for Weibo's intelligent search in handling hotspot tracing, viewpoint aggregation, and public opinion analysis.

In addition, the report emphasizes the significance of the AI entrance battle at the "strategic defense" level, which is essentially a contest around the distribution right and commercial liquidity right of next-generation Internet traffic. Weibo has deeply embedded intelligent search into its content ecosystem, which not only improves the efficiency of netizens in obtaining real information, but also effectively strengthens its own traffic base.

Overall, the development of AI entering the deep water zone is the real status quo of the industry. The differentiated development of native apps and application plugins confirms the path selection under different business genes. In the future, how to build high ecological barriers based on native scenarios in the era of "technological equality" will be the core proposition that determines the commercial value of AI in various enterprises.