Summary of Key Points
Over the past two decades, the three major telecommunications operators have made their living by "operating data traffic" (selling internet bandwidth). With the advent of the AI era, they have shifted to using Tokens (the smallest units for processing information by large models, akin to the "fuel" for AI) as their new form of "traffic." Each operator has adopted a different approach to enter the Token business: China Telecom is focusing on underlying capabilities such as computing power and models; China Mobile is transforming SIM cards into entry points for AI services with AI-eSIM chips; China Unicom, on the other hand, emphasizes the real value of applications. The ultimate goal for all three is to seize the growth opportunities of the AI era and find a new source of revenue beyond traditional data traffic.
Detailed Analysis
1. Why Have Tokens Become the New Form of "Traffic" for Operators?
In simple terms, Tokens are the units used to measure the workload of AI systems. For example, when you ask an AI to write an article, it breaks down the text into numerous Tokens for processing. In the past, operators sold the "infrastructure for accessing the internet" (data traffic), but now AI services require Tokens as this "fuel." Tokens are tied to computing power, models, platforms that manage resources, and actual applications. Whoever controls these assets will profit significantly in the AI era. Therefore, by embracing Tokens, operators are transitioning from being mere providers of network infrastructure to sellers of AI service capabilities.
2. China Telecom: Building the Foundation Before Selling Packages
China Telecom's strategy is to strengthen its underlying capabilities first:
- Foundation System: The company has established a comprehensive framework consisting of a computing power network ("Xirang"), a data platform ("Xinghai" with over 100 industry-specific datasets), and multiple industry-specific models (e.g., for healthcare and manufacturing).
- Integrated Services: It offers integrated cloud services that combine computing power and models to provide a range of AI-based solutions.
- Implementation Model: The company adopts a "front-store, back-factory" approach, where the front end meets customer needs (e.g., providing AI quality inspection services for businesses), while the back end develops reusable models and tools. For instance, the Zhejiang fire warning project uses 4 billion Tokens daily, and industrial digital twin applications consume 230 million Tokens per month—both are profitable use cases.
3. China Mobile: AI-eSIM Chips to Transform SIM Cards into AI "Keys"
China Mobile is upgrading SIM cards with AI-eSIM chips:
- Chip Function: Traditional SIM cards serve only as identification for connecting to the internet, but AI-eSIM chips not only provide connectivity but also allow direct access to large models, essentially acting as a "key" to unlock AI services. Service packages include both data traffic and Tokens, with unified management for both purposes.
- Application Scenarios: These chips are already being used in smart transportation (e.g., in-vehicle AI), educational toys, robotics, drones, and other fields.
- Ecosystem Development: China Mobile has established a dedicated committee for AI-eSIM to set standards and promote hardware innovation, aiming to make these chips a standard feature in all AI-enabled devices.
4. China Unicom: Focusing on the Real Value of Applications
China Unicom prioritizes the practical benefits of applications:
- Concerns: Despite the rapid growth of Tokens, there is skepticism about how many truly help businesses improve efficiency or merely serve as symbolic assets.
- Actions: The company has upgraded its "AI Smart Selection Unicom Cloud" to integrate computing power, Tokens, and AI services on a single platform, ensuring compatibility with domestic software and improving training and deployment efficiency. Its enterprise platform, "Unicom Yunxi," serves nearly 50,000 companies, handling billions of Tokens daily, with a focus on whether these Tokens actually solve real problems (e.g., optimizing business processes).
- Cost Reduction: China Unicom is working to reduce costs through synergies between computing power and electricity usage, helping customers quickly develop customized models.
5. Behind the Tokens: Operators Redefining Their Roles
In the past, operators were merely providers of network infrastructure; now they aim to become integrators of AI services. Although their approaches differ, all three are striving to capitalize on the new opportunities offered by AI. This shift signifies a fundamental change in the competitive landscape of the telecommunications industry.
Conclusion
The Token business of the three major operators represents a battle for transformation in the AI era—moving from selling data traffic to providing comprehensive AI services. Although their approaches vary, they are all working to leverage the new opportunities brought by AI, indicating that the competitive dynamics in the communications industry are undergoing significant changes.