Liu Guiqing stated that China Telecom is fully embracing AI and advancing its corporate strategy toward the “Cloudification, Digital Transformation and AI for Good” upgrade, consistently placing technological innovation at the core of its corporate strategy and driving the company’s transformation from a traditional telecommunications operator into a technology-oriented enterprise. China Telecom’s eSurfing Cloud has already become the world’s largest carrier cloud service provider and China’s largest hybrid cloud service provider, and is now stepping into a new phase of intelligent cloud development.
Liu Guiqing noted that, in advancing the commercial deployment of 5G networks, how to achieve industrial coordination and promote green, sustainable development is a question that operators around the world have been continually exploring. China Telecom and China Unicom have jointly explored 5G co-construction and sharing, overcoming a series of world-class technical and engineering challenges to build the world’s first and largest 5G SA co-built and shared network, providing invaluable experience for large-scale 5G deployment globally. The two parties have now shared over 1.54 million 5G base stations and over 2 million 4G base stations, cumulatively saving USD 56.5 billion in investment, reducing annual operating costs by USD 6.5 billion, and cutting carbon emissions by 13 million tonnes per year.
Liu Guiqing introduced that, in the era of AI, China Telecom is fully leveraging the operator’s integrated advantages in “computing power + algorithms + data” to build its core technology “Xirang” and construct a five-in-one intelligent cloud system encompassing “computing power, platform, data, models, and applications.” At the IaaS layer, it has built a computing power layout covering general computing, intelligent computing, supercomputing, and quantum computing, achieving three-dimensional coverage across all-optical networks, computing-power internet, mobile communications networks, and satellite networks. At the PaaS layer, it provides one-stop computing scheduling and AI development services. At the DaaS layer, it builds high-quality datasets and a trusted data circulation toolchain. At the MaaS layer, it independently develops AI large models, establishing leading advantages in the fields of semantics, speech, vision, and multimodal capabilities. At the SaaS layer, it builds standardized AI products and launches industry-specific large models and intelligent agent services. Simultaneously, it is actively building a mutually beneficial and win-win cooperative ecosystem and a security framework spanning models, data, and applications, releasing the “Jianwei” security large model and open-sourcing China’s first foundational security guardrail for large models.
China Telecom is advancing cloud-network integration to provide customers with integrated computing-network services offering ultra-strong computing power, ultra-low latency, and full-domain scheduling. It is continuously upgrading intelligent computing data center construction, with a total data center rack scale exceeding 590,000 racks. It is continuously upgrading fundamental network capabilities, building a millisecond-access computing network, deploying AIDC-centric inter-node networks and lossless intra-node networks, and completing the world’s largest 100G/400G all-optical network, reducing inter-hub node round-trip latency to 12 ms. It is continuously upgrading the integrated computing-network system — the Xirang integrated intelligent computing service platform — with schedulable computing power reaching 87 EFLOPS. The Triless platform architecture has been launched, achieving triple decoupling of resources, frameworks, and tools to provide users with flexible scheduling of cross-domain and heterogeneous computing power. China Telecom is also deepening “computing-power and electricity coordination.” In Shanghai, it has established the world’s first subsea data center, deployed directly on the seabed and powered directly by an offshore wind farm, with a green electricity utilization ratio exceeding 95% and electricity costs reduced by 50%.
China Telecom is working to become a provider of data and foundational large model services. It has built a Data Intelligence Middle Platform that aggregates proprietary, open-source, and third-party data — including 10 trillion tokens of telecommunications-industry data and over 350 TB of industry data spanning 14 sectors — to empower model training and applications, and to provide customers with dataset and annotation services. It independently develops the Xingchen large model system and the intelligent agent service platform, and actively introduces third-party foundational large models and various industry-specific large models to meet the needs of different industries and customers for large model selection and application scenario innovation, accelerating the drive toward more inclusive model services.
Liu Guiqing emphasized that only by continuously expanding the breadth and depth of applications can AI bring about qualitative change. China Telecom integrates AI into the core processes of its own network operations, customer service, and technology R&D, using AI to transform corporate workflows and comprehensively enhance operational efficiency. For example, in network operations, it has built a cohort of digital employees based on network large models to handle repetitive and foundational operational tasks, reducing the average monthly number of on-site repair visits by field technicians by 35%. Leveraging R&D large models, AI-generated code now accounts for 40% of all code produced, improving R&D efficiency by 20%. AI is comprehensively reshaping core services, with the development of a series of intelligent products including eSurf Smart Ring, eSurf IntelliHub, AI Cloud Computer, and AI Phone. For instance, the “eSurf IntelliHub” has reengineered the traditional “FTTR + IPTV” home service model, becoming the unified gateway for home AI and integrating full-scenario services including security and protection and healthcare. It is also driving AI empowerment of enterprise clients’ core processes, providing AI services to customers in industrial manufacturing, education, healthcare, and other sectors, facilitating the digital-intelligent transformation of the economy and society.
Liu Guiqing stated that operators inherently possess the resource endowments of extensive network connectivity and rich data scenarios. How to transform these core advantages into forward-looking core competitiveness in the era of AI is a question requiring in-depth exploration and collaborative resolution. Drawing on China Telecom’s own practice, Liu Guiqing put forward five proposals:
- 6G standard innovation and network deployment must fully account for the rapid development of AI. Global operators should consolidate industrial efforts to build a globally unified 6G standards framework. 6G standard-setting must proactively adapt to the rapid iteration of AI development, achieving deep integration between network connectivity and the intelligent engine.
- Cloud-network integration will play an ever greater role in the AI era. Global operators should fully leverage the core driving force of “cloud-network integration,” actively advancing the upgrade of DCs to AIDCs, providing customers with integrated computing-power interconnection and scheduling services, fulfilling the strategic mission of AI infrastructure providers, and building a solid computing-power foundation for AI.
- AI security governance will become a mandatory topic for global operators, and is also a watershed defining the strength of operational and service capabilities in the intelligent era. Security has become a vital cornerstone of AI — “no security, no intelligence.” Global operators must collaborate in governance, actively participating in the improvement of AI security governance frameworks, and steering the healthy and orderly development of AI in a beneficial, safe, and equitable direction.
- Computing-power and electricity coordination capability will become the key to the sustainable development of intelligent computing infrastructure. Global operators must jointly address the challenges posed by energy supply, coordinate and schedule computing-power resources, deepen computing-power and electricity coordination — strengthening computing power through electricity — and promote the integrated development of AI and green, low-carbon practices.
- The flourishing development of AI applications requires operators to open up and cooperate with greater force. Global operators should deepen communication and collaboration, promoting the open sharing of AI models and the inclusive prosperity of AI applications, so that the fruits of AI innovation benefit global industry partners.
Finally, Liu Guiqing stated that China Telecom is willing to use this Congress as a bond to join hands with GSMA and global industry partners, and together build a bright future for operator development in the era of AI.
Hashtag: #ChinaTelecom #LiuGuiqing #MWC2026
The issuer is solely responsible for the content of this announcement.
