Summary of Key Points
In the first half of 2026, the competitive landscape in the global AI application industry underwent a fundamental shift: the focus has moved from competing on the number of model parameters and user base to evaluating actual implementation and commercial returns. The industry is no longer solely interested in the cool features that AI can offer; instead, it emphasizes whether AI can integrate into real business processes, complete specific tasks, and ultimately lead to measurable improvements in efficiency and revenue generation. The report compares the differences between domestic and international markets, the commercialization strategies of leading large models, the trends in market segmentation, and the ecological layouts of top companies. The core conclusion is that the AI industry has entered a period of "value revaluation," where the quality of growth (the ability to convert growth into stable revenue and sustained demand) is more important than mere scale expansion.
I. Global Market: China Relys on Mobile Devices for Traffic Acquisition, While Overseas Markets Balance Both App and Web Platforms
The growth paths in the AI markets in China and abroad are vastly different:
- Chinese Market: Mobile devices dominate growth, with almost all growth coming from mobile apps. In particular, the AI chatbot category saw a 65% increase in the first half of the year. The DouBao app has 468 million monthly active users, and QianWen has 180 million users, accounting for 88.7% of the total growth in the top 50 AI apps in China—meaning that most new users are attracted through these two platforms. However, this concentration of growth makes it difficult for smaller products to gain a share of the market.
- Overseas Markets: There is a balanced approach using both app and web platforms. Apps are used to attract general users (for chat, entertainment, etc.), while web platforms focus on productivity tools (programming, office work, corporate collaboration). For example, web-based office tools saw a 164% increase, and 3D tools grew by 85%. The advantage of this dual-platform approach is that it allows for user acquisition through apps and generates high-value payments from business and professional users. Chinese companies excel at acquiring traffic, but the overseas "web-based productivity ecosystem" is worth learning from.
II. Leading Large Models: More Users Do Not Necessarily Equal Higher Revenue; Different Paths to Success
In the past, the focus was on the number of users for large models. Now, it's about how they generate revenue. The report highlights that the four leading players each adopt different strategies:
- ChatGPT: Has shifted from selling models to offering subscription services. Although it still has the largest user base globally (1.264 billion), its revenue structure has changed: personal subscriptions account for 80%, corporate subscriptions for 17%, and API sales (licensing the model to other companies) for only 2%. This indicates that ChatGPT has evolved from a model provider to a global software platform, with users willing to pay a monthly fee for its services.
- Claude: Smaller but more targeted, with a focus on professional users who are willing to pay high fees. Although it has fewer users than ChatGPT (260 million), the web platform accounts for a significant portion of its revenue, as its users are programmers, researchers, and corporate knowledge workers. Its revenue growth was the fastest due to frequent use and high willingness to pay.
- Gemini: Benefits from Google's ecosystem, being integrated directly into Google Search, Workspace, and cloud services. For example, AI can automatically improve documents when used with Google Docs, making it more appealing to users.
- Grok: Has gained users through social platforms but has yet to find a stable revenue model. It has attracted 130 million users through social networks like X, but its revenue is volatile, and it needs to prove that users are willing to pay on a long-term basis.
In summary, large models generate revenue through a mix of methods, including subscriptions, enterprise services, advertising, and task-based fees.
III. Market Trends: Consumer-oriented Tools Are Popular, While Productivity Tools Generate Higher Revenue
There is a clear differentiation between domestic and international market trends, but one common pattern emerges: basic tools are being phased out:
- China: The fastest-growing categories include chatbots, image generation, education, and virtual assistants (e.g., AI-powered girlfriends). These tools have low entry barriers and large user bases, but users are less willing to pay.
- Overseas: Productivity tools, such as office collaboration and development tools, are seeing significant growth. For example, web-based office tools grew by 164% overseas, and similar tools in China also increased by 54%.
- Which Tools Are Being Phased Out? Simple writing tools, PPT templates, and basic question-answering apps are being integrated into more comprehensive AI assistants. For example, if you can use an AI tool like DouBao to write documents directly, the need for standalone apps decreases.
IV. Leading Companies: Moving from Single Products to Integrated Ecosystems
The competition among large companies now focuses on whether they can integrate their products effectively:
- Overseas: Google and Microsoft have built ecosystems that incorporate AI into search, office software, and cloud services, providing users with a seamless experience across various products.
- China: Companies like ByteDance (with DouBao for chat) and Alibaba (with QianWen and KuaiKe for video editing and content creation) are extending their offerings to education and image processing. Tencent combines chatbots with coding assistants, and Baidu offers web-based assistants and legal services. The key is the synergies between these products; for example, DouBao users can use CapCut to edit videos directly, enhancing user retention.
V. Quality of Growth: Focus on Retaining Revenue
The report emphasizes that high growth rates are not always a good sign (e.g., they may be due to a low base). The true "growth stars" are those that:
1. Have surpassed the scale threshold (e.g., continue to grow even with hundreds of millions of users).
2. Outperform their peers in their respective markets.
3. Can create sustained demand (users use their products regularly).
For example, Tencent Cloud's coding assistant in China saw a 994% increase in revenue because programmers rely on it for daily tasks, making it a essential service. Zoom Workplace overseas grew by 505% due to the demand for AI collaboration tools.
There are also new developments in China's AI exports: the focus has shifted from single-hit hits (like CapCut) to a dual-platform approach (apps and web), with visual tools remaining a core focus, while expanding into video generation and office tools.
Four Things to Watch in the Second Half of the Year
1. Whether leading assistants like DouBao and ChatGPT can convert their user bases into stable revenue.
2. Whether intelligent systems (capable of complex tasks like booking flights and hotels) can move from demonstrations to practical applications.
3. Whether traditional tools (e.g., PPTs, writing apps) can develop new professional barriers that differentiate them from competitors.
4. Whether Chinese AI companies can transition from their mobile-based traffic advantage to web-based productivity and global commercialization.
In conclusion, the AI industry has moved beyond the stage of showcasing technical capabilities; the focus is now on practical applications that can help businesses save money and users save time, ultimately generating revenue.
The value of this report lies in its ability to provide a sober perspective: we should not be misled by short-term growth rates but should consider the underlying business logic and long-term value. Only AI products that can survive and generate revenue will truly succeed.