Summary of Key Points
The explosive growth of AI technology is driving telecommunications operators to transform from traditional providers of data bandwidth and connectivity services into intelligent infrastructure service providers. The approach to AI adoption varies significantly among global operators:
- China follows a “full-stack AI” strategy, covering the entire spectrum from computing power to models to industry-specific applications.
- The United States focuses on building a strong foundation in AI infrastructure, including networks, data centers, and edge computing.
- Europe emphasizes the importance of “trustworthy and compliant” AI solutions that meet strict regulatory requirements (such as those set by the EU’s Artificial Intelligence Act and GDPR).
- Japan and South Korea adopt a balanced approach of developing their own AI capabilities while also collaborating with international partners.
The ultimate success of these efforts will depend on the ability to convert AI into tangible improvements in efficiency and sustainable revenue, rather than simply accumulating computing power or releasing large-scale models.
Chinese Operators: Moving from Communication Providers to Full-Stack AI Service Providers
China’s three major telecom operators (CMCC, CTCC, and UNICOM) are all considering AI as a new growth driver, aiming to become comprehensive digital service providers that offer end-to-end solutions. Their approaches include:
- CMCC: Promoting the simultaneous development of computing power, models, and data resources. For example, its “Jiutian” large-scale model framework has been upgraded to version 3.0, and it has compiled datasets totaling 3500TB for 32 different industries, providing end-to-end services to government and enterprise clients. By 2025, AI-related revenues are expected to reach nearly 90 billion yuan.
- CTCC: Highlights the integration of cloud, network, computing power, data analytics, intelligence, and security services. Its “Tianyi Cloud” platform manages computing resources, and its “Xingchen” large-scale models serve as entry points for various applications, particularly for clients with high data security requirements (such as governments and state-owned enterprises). By 2025, CTCC plans to offer services through 110 industry-specific models to 37,000 customers.
- UNICOM: Offers a MaaS (Model as a Service) platform that allows companies to select, train, and utilize AI models. AI-related revenues are expected to grow by over 140% by 2025.
Challenges: While these operators focus on providing integrated services, they face the risk of homogenization due to their widespread adoption of full-stack approaches. To be profitable, they will need to transform customized solutions into standardized products.
Differentiated Approaches in Europe, America, Japan, and South Korea
Other regions have chosen different paths based on their unique strengths and market conditions:
- The United States: Does not pursue large-scale AI models but focuses on building a robust AI infrastructure. Companies like Verizon combine high-capacity networks, data centers, and security capabilities to serve cloud providers and large enterprises. With giants like Google and NVIDIA in the industry, operators can leverage their strengths in connectivity without competing directly with model developers.
- Europe: Places a strong emphasis on “trustworthy and compliant” AI solutions. Partnerships with companies like Microsoft are used to deliver customer services and marketing initiatives using external models. The EU’s strict regulations require secure, auditable, and locally deployed AI services, making compliance a key differentiator for operators.
- Japan and South Korea: Balance developing their own AI capabilities with international collaboration. Companies like SK Telecom develop both proprietary and lightweight models in Japanese, while NTT explores consumer-oriented AI applications. Despite their smaller domestic markets, they seek to maintain technological independence without becoming too closed-off.
Why the Different Approaches?
These differences are driven by various factors such as industrial structures, regulatory environments, and market demands:
- China: With large operator sizes and a focus on government and enterprise clients, China’s operators can adopt a full-stack approach to leverage their network and distribution advantages.
- The United States: Its mature AI ecosystem, with strong cloud, chip, and model companies, makes it more cost-effective for operators to invest in infrastructure.
- Europe: Stringent regulations require operators to offer compliant AI solutions that meet customer needs for privacy and security.
- Japan and South Korea: With limited domestic markets, these countries rely on technology innovation to maintain independence while also seeking opportunities through international cooperation.
The Future of AI Transformation
The success of telecom operators in the AI era will not depend on the size of their models or the amount of computing power they possess. Instead, the key factors include:
1. Smarter networks: AI will evolve from being a support system for operations to becoming autonomous, capable of automatically resolving network issues (e.g., reducing repair times from hours to minutes).
2. Edge computing: AI processing will move closer to users (e.g., at the location of base stations), improving response times and leveraging operators’ 5G and fiber optic networks.
3. Integrated models and intelligent agents: Companies will use multiple models, and operators will need to help them integrate these models and manage data security and computing resources effectively.
4. AI sovereignty: Both China’s focus on autonomy and Europe’s emphasis on data sovereignty mean that customers want AI solutions they can control. Operators must provide standardized services that meet these needs.
In conclusion, AI will transform telecommunications networks into intelligent platforms that connect computing power, models, businesses, and individuals. Those who successfully integrate AI into their services to create stable revenue streams and enhance customer experiences will gain a competitive advantage in the AI era.