Summary of Key Points
Recently, the AI industry has witnessed a highly counterintuitive shift: The ordinary C-end users, who were once the target of massive investment by various companies, have now become a liability for AI manufacturers, as over 700 million of them use AI services daily and are difficult to drive away. As a result, the entire industry has stopped the costly competition in the C-end market and has turned its focus to the B-end (enterprise-level) AI solutions. The new strategy for the C-end is to "not invest heavily, not chase users, and simply occupy the market while waiting for new opportunities."
---
Detailed Analysis
1. The Counterintuitive Phenomenon: "Users Who Can't Be Driven Away" Have Become a Burden
There used to be a golden rule in the internet industry that has persisted for over a decade: once you attract and retain users with free services, you can generate revenue effortlessly. Products like WeChat and Alipay grew into giants based on this principle. However, the AI industry has taken the exact opposite approach. Currently, the four leading AI products in China—DouBao, QianWen, DeepSeek, and iFlytek—combined have over 710 million monthly active users, with nearly half of Chinese internet users relying on them for tasks such as writing content, researching information, and chatting. These users are so loyal that they are difficult to lose. Yet, manufacturers are no longer willing to invest heavily in acquiring new C-end users. DeepSeek has even stated that it is no longer focusing on C-end development and considered stopping C-end maintenance. The costly tactics used in the past, such as sponsoring the Spring Festival Gala and bidding for advertisements, have ceased. The high-retention users that were once highly sought after in the internet industry have now become a burden in the AI industry, completely overturning previous industry norms.
2. All Three Paths to Monetization Have Failed
AI manufacturers have tried every possible C-end monetization method over the past two years, but none have been successful:
- Direct Membership Fees: The payment rate is extremely low. Even the most successful AI product, ChatGPT, has only 50 million paid users out of its 900 million monthly active users, with less than 6% paying. Official predictions suggest that the number of paid users will further decrease from 44 million to 9 million. Domestic users are even less willing to pay for basic AI services. Even Kimi, which has the best subscription performance in China, earns 70% of its revenue from providing technical services to businesses, not from individual subscriptions.
- Advertising: The classic free-to-use model of attracting users through advertising has failed in the AI industry. OpenAI's advertising efforts generated less than 2.5% of its total revenue, which is barely significant. Moreover, AI-generated ads can lead to misleading recommendations. For example, fake reviews can be used to promote non-existent products. Regulatory authorities are focusing on addressing this issue, and the advertising revenue generated is not enough to cover the costs of violations.
- AI Hardware: AI has become a bonus feature in hardware products. Companies once launched AI-enabled phones, glasses, and learning devices, but consumers mainly bought the hardware itself, with the "built-in large models" being merely a marketing gimmick. With declining global smartphone shipments and longer replacement cycles, the profit from hardware sales is thin, and AI manufacturers cannot profit significantly from this approach.
3. AI Solutions That Make Money in the C-End Often Leverage Existing Needs
Not all C-end AI services are unprofitable. For instance, Baidu Wenku has over 40 million paid users after integrating AI, and iFlytek's AI learning devices have become the leaders in the education hardware market, generating substantial revenue. However, these successes do not rely on users paying for the AI itself. Baidu Wenku had a long history of paid subscriptions even before adding AI, and the AI just enhanced the value of its membership. Similarly, parents were already willing to spend on educational devices for their children, and the AI added value to these products. In contrast, most free AI chat tools are used for simple tasks like searching and chatting, and there is no historical precedent for users paying for these services.
4. The Industry's Shift to the B-End
All AI manufacturers are now focusing on the B-end market because it is a clear source of revenue. Enterprises are willing to pay for AI office solutions, which can save significant costs by automating tasks such as meeting notes and data processing. Since enterprises are less likely to switch to cheaper alternatives, and the cost of switching to a new AI service is high, serving enterprises is more cost-effective than serving hundreds of millions of free users.
5. Why Not Drive Away C-End Users? Manufacturers Are "Occupying the Market While Waiting for the Next Wave"
AI manufacturers are not foolish; although they cannot make money from the C-end for now, they do not want to drive away users. They adopt a strategy of "not investing heavily and not chasing users." There are two main reasons for this:
- They fear missing the next major C-end AI trend. Who knows if a future service that can handle all work tasks, schedule appointments, or control smart home devices will emerge? If they lose the C-end market now, they might be unable to adapt when such a service becomes popular.
- There is a tacit agreement among manufacturers not to invest heavily in competitive development. Since most companies already have basic C-end features, they choose to keep them free to avoid additional research and development costs and prevent user churn. This approach ensures that they have a basic level of coverage while waiting for the next opportunity.
For us as ordinary users, free AI tools will continue to be available for a long time, and basic functions like writing and researching will remain free. However, disruptive new AI services are likely to take longer to emerge, as manufacturers are currently focusing on generating revenue from enterprises.