AI-powered radio intelligence is stealing the spotlight in telecoms – and China is stealing the show. While global telcos experiment, Chinese operators are delivering Level 4 automation at scale, with real-world impact.
The news | Tier one vendors innovate with RAN intelligence
At MWC25, mobile network intelligence and automation was yet again a key topic, with generative artificial intelligence (gen AI), machine learning (ML), and various other forms of AI taking center stage to optimize telco networks and help mobile operators enhance their profitability. The promise of radio intelligence lies in its ability to automate network operations – from real-time traffic-aware power tuning and spectrum scheduling, to mitigating interference and supporting differentiated services. The ultimate goal: empower mobile operators to onboard new services and business models with agility and confidence. Two main ways of achieving this have appeared in the market over the past few years, and many relevant products were announced at MWC.
On one hand, the RAN Intelligent Controller (RIC) and the Service Management Orchestration (SMO), both O-RAN Alliance concepts, continue to evolve – with several vendors presenting early RIC-based platforms, although few had reached general availability, and even fewer showed real-world deployments in existing networks. Fragmentation, lack of interoperability, and abstract use case targeting remain significant hurdles.
Moreover, macro industry trends have resulted in a few vendors decommissioning their RIC/SMO products, including Broadcom/VMWare stopping development of its RIC and SMO. Tier-one vendor automation software that has multi-vendor integration capabilities, including Ericsson’s Intelligent Automation Platform and Huawei’s Intelligent RAN, are now emerging as commercial and mature alternatives.
Coupled with China’s 5G market maturity, its national decision to deploy Standalone (SA) from the beginning of 5G deployments and aggressive technology adoption all mobile operators are now showing signs of Level 4 (L4) automation.
The impact | Chinese operators reach Level 4 (L4) automation
China Mobile has formally published its autonomous network L4 strategy and implementation path, kicking off a national initiative known as the “gold-pointing action.” This program aims to accelerate the commercial realization of L4 capabilities through targeted scenarios, including fault management, performance optimization, and energy efficiency. China Mobile has commissioned Huawei’s Intelligent RAN 2.0, which utilizes a multi-agent architecture to drive automated, closed-loop decision-making. The operator is also working with Huawei and has deployed the vendor’s Watt Master, an intelligent power management solution that uses real-time environment perception and multi-objective simulation to fine-tune energy consumption. In Hubei, this system delivered 10.4 percent energy savings, all while maintaining service key performance indicators (KPIs) such as access success rate and call drop stability.
China Telecom has taken a more standards-based approach, announcing more than 20 self-intelligence standards for cloud-network operations and achieving L4 automation in multiple high-value domains at MWC25. These include energy savings, SLA fulfillment, and customer experience assurance. The operator has also started working with Huawei and has deployed the vendor’s Fault Management and Engineering (FME) Mate, which uses a telecoms foundation model trained with network data to manage errors more effectively. The operator expects to improve Mean Time to Repair (MTTR) by 30 percent.
China Unicom has implemented L4 practices in targeted high-value scenarios such as coverage optimization, spectral efficiency, and fault avoidance. The operator has released a live broadcast package for consumers that places additional requirements on the network and requires intelligent processes to ensure a high-quality live broadcast experience. The operator is now assessing which capabilities it needs to deploy in its network to satisfy these requirements.
The message | The West has to learn from the Chinese experience
What sets China’s approach apart is not just the technology, but the industrial scale and systematic planning behind the automation journeys of all mobile operators. These operators are not dabbling in AI – they are building frameworks, defining maturity models, and measuring impact through structured KPIs and business value indicators aligned with the TM Forum IG1256 and IG1345 standards. As global operators convened at MWC25, the story from China provides a powerful case study on how to go from pilot to production, from vision to execution, and from manual operation to autonomous value creation.
At the core of this transformation lies a holistic framework that combines advanced AI-native architectures, multi-agent systems, real-time telemetry, and standardized KPI models – all working in concert to enable intent-driven automation. While many global carriers are still in early experimentation or isolated proof-of-concept stages, China’s carriers are already demonstrating commercial viability, operational benefits, and replicability across diverse geographies and service environments.
For global operators, the message is clear: the path to scalable, sustainable, and service-centric 5G – and beyond – must include intelligent, closed-loop automation. And that path must be approached not just as a technology deployment, but as an organizational transformation that bridges operations, engineering, and business strategy. China’s operators have shown that with the right vision, framework, and ecosystem, it is possible to make radio intelligence commercially effective, not just technically feasible. Operators in other regions are following, including AIS in Thailand and e& in the United Arab Emirates (UAE), but Chinese operators are leading the way.
