Li Huidi, Executive Vice President of China Mobile, delivered a keynote speech titled “AN L4 Powering the IQ Era” sharing the company’s latest insights and practices in advancing Level 4 (L4) Autonomous Networks (AN).
Li noted that with the deep integration of AI and networks, the two are converging into a self-accelerating flywheel. “AI re-architects how networks operate, and networks expand the scalable frontier of AI,” he said. Together, they propel AN from concept to large-scale deployment, and from efficiency optimization to value creation.
According to Li, L4 is now entering real networks, operations, and services. Networks are beginning to understand user intent, make decisions proactively, and transition from passive response to active decision-making. The large-scale deployment of Autonomous Networks worldwide, he emphasized, has become a defining feature of the industry’s current stage of development.
From isolated intelligence to holistic intelligence
As China Mobile advances its AN journey toward L4, Li identified two major trends accelerating this transformation.
The first trend is the evolution of AI from isolated intelligence to holistic intelligence. As Networks for AI continues to mature, the industry is building a complete ecosystem to support AI-driven services and operations.
This transformation is reflected in three dimensions.
AI in Networks.
China Mobile has proactively prepared for 6G development and updated more than 400,000 Wireless AI base stations equipped with intelligent boards. Together with the NWDAF, telecom networks are evolving from “manageable and controllable” toward fully autonomous systems.
AI in Operations.
By leveraging digital twin technology and millisecond-level network sensing, China Mobile has built a closed-loop decision-making system that integrates virtual and physical networks. In its IP network, traffic volumes approaching 300 Tbps are dynamically scheduled on a daily basis.
AI in Services.
AI is enabling instant, intelligent, and personalized offerings. China Mobile’s 5G New Calling service now serves more than 74 million users, reflecting how intelligence is directly embedded into commercial services.
From models to real applications
The second trend Li highlighted is the industry’s shift from models to real applications.
Li said. “AI Agents are now becoming the primary vehicle for AI applications. They can understand intent, act autonomously, interact with the environment, and close the loop end-to-end — from service provisioning to fault demarcation.”
To enable efficient and secure multi-agent collaboration across domains, China Mobile is leading industry partners in collaboration with TM Forum to advance the A2A-T (Agent-to-Agent Telecom) framework. Designed specifically for telecom network operations, the framework effectively reduces multi-agent interaction latency and supports the evolution toward an agentic architecture.
Building the infrastructure foundation for the IQ era
Li also introduced China Mobile’s progress in building the new-generation information infrastructure foundation to support the IQ Era.
5G-Advanced (5G-A) has achieved large-scale commercial deployment across urban areas nationwide. 10-Gbps fixed access is being rolled out in parallel. The company’s self-owned AI Computing power now exceeds 60 EFLOPS, forming a nationwide, integrated and coordinated deployment of computing and network resources.
On this foundation, China Mobile is comprehensively upgrading AN to L4 across 21 High-Value Scenarios (HVS), including service provisioning, complaint handling, and fault management. The company has restructured 31 core workflows and streamlined 45% of manual and sequential tasks. As a result, efficiency in key operations has improved by 70%, saving 6,000 person-years annually and reducing electricity consumption by more than 8 billion kWh per year.
A cross-domain fault management breakthrough
Li highlighted a cross-domain fault management case as a concrete example of AN in action and joined industry partners in unveiling the achievement.
China Mobile has established “shortening MTTR and reducing manual intervention” as its core KBI/KEI. Addressing the long-standing challenge of cross-domain fault correlation, the company has built a three-phase cross-domain autonomous closed loop—“service assurance, fault repair, and network restoration”—based on its in-house developed JIUTIAN Network Large Model, creating a new “carbon-silicon symbiosis” operating model.
The solution has completed provincial-level pilots, serving 110 million users. After live-network verification, MTTR has been reduced by 25%, and annual labor savings exceed 100 person-years. The solution will be further scaled nationwide.

The end of the beginning
“L4 is not the end. It is not even the beginning of the end. But it is, perhaps, the end of the beginning,” Li concluded.
He emphasized that today’s L4 marks a turning point, moving Autonomous Networks from exploration and pilots into holistic, large-scale deployment. As 5G-Advanced continues to expand and AI capabilities further mature, connectivity will extend from relatively static links between people and things to dynamic capabilities spanning communication services, computing services, and intelligent services.
China Mobile stands ready to work with global partners to deepen open collaboration and co-create value, aiming to make Autonomous Networks the core operating system of the IQ Era.
The summit brought together leading players in the global Autonomous Networks ecosystem, including TM Forum, Deutsche Telekom, Telefónica, Orange, Huawei, and AsiaInfo Technologies, who jointly explored pathways to accelerate AN L4 deployment and advance Autonomous Networks from industry consensus to global-scale implementation.

ITW Asia 2026
ITW Asia brings together the whole connectivity and digital infrastructure industry to get business done. Join 1700+ leaders from carriers, MNOs, cloud solution providers, hyperscalers, content service providers, data centres, satellite operators, investors, regulatory authorities and more, to define the future of connectivity in Asia.





