AI and RAN Deep Integration: Driving Telecom Business Model Innovation

Amid the global digital transformation wave, the deep integration of AI and RAN is reshaping the telecom industry’s commercial landscape. With traffic dividends fading and competition intensifying, operators face the challenge of “increasing traffic volume with flat revenue growth”: although mobile data traffic surged 245% year over year, it contributed just 3% to the revenue growth. Converting data “volume” into business “value” has become a critical challenge.

Rapid advancement in AI is unleashing the potential of RAN. AI and RAN convergence not only tackles pain points, including limited spectrum resources, high energy use, and complex O&M, but also drives business model innovation and ecosystem restructuring. It accelerates the shift from “connectivity as a service” to “intelligence as a service,” unlocking new value growth curves for the industry.

Business Model Innovation: From Traffic Pipeline to Capability Supermarket

Traditional operators rely on traffic and voice revenue, a single, rigid model that struggles to meet diverse user needs and fierce competition. Other industries lead the way: logistics offers service tiers (next-day, same-day, even hourly delivery), while transportation offers business, first, and economy classes—both sharply boosting value through differentiation.

AI’s deep integration with RAN is transforming mobile networks into diverse, intelligent, ecosystem-driven platforms. Through the synergy between “AI for RAN” and “RAN for AI”, network capabilities are monetized, data is translated into revenue streams, and service capabilities are diversified. This creates an open, flexible, value-sharing digital ecosystem, unlocking entirely new growth paths for the telecom industry.

Network Capabilities Monetization: From Dumb Pipes to Golden Pipes

By deeply integrating AI and RAN, operators greatly enhance the value of connectivity and drives value-based operations. AI-enabled intelligent scheduling and resource allocation boost efficiency and user experience, breaking the limitations of the traditional dumb pipe. AI-driven precise service identification enables personalized offerings, transforming connectivity into an intelligent “golden pipe.”

RAN capabilities (e.g., bandwidth allocation, latency optimization, and network slicing) can be encapsulated and exposed via APIs, enabling operators to build a flexible “capability supermarket.” This shifts the RAN from providing mere connectivity to building an ecosystem, empowering enterprises, developers and third parties to rapidly innovate and meet diverse industry needs, while opening new revenue streams for operators.

l AI for RAN: AI algorithms optimize RAN resource allocation (spectrum, power, load balancing, and user steering) based on traffic load, service type and radio conditions. In high-load scenarios, AI can provide differentiated experience assurance tailored to services and scenarios, delivering Quality-on-Demand (QoD) services. For example, gaming platforms can subscribe to QoD APIs to provide their members with guaranteed low-latency, smooth gameplay, thereby improving user satisfaction.  

l RAN for AI: RAN provides low latency, high-reliability connectivity and edge computing to support real-time AI inference. For instance, in autonomous driving, RAN offers tailored network slices via service APIs to ensure stable, secure, ultra-low-latency transmission, fully enabling AI-driven applications.

Data Monetization: From Connectivity to Value Mining

Massive data generated in RAN system can be transformed into tradable data assets after AI-driven anonymization and compliance processing. These assets deliver deep insights for enterprise, enable data-driven business models, unlock hidden value, and create new revenue for operators. For example,  mobility insights can help retailers optimize store location selection, improve layout planning, and design targeted promotional strategies.

AI for RAN: AI algorithms analyze user behavior, traffic patterns and network status to deliver high-value insights. It identifies traffic hotspots, enables targeted experience packages for events (concerts, major matches), provides venue-specific experience guarantees, and supports data-driven urban planning and site selection.

RAN for AI: RAN edge nodes enable real-time data processing, cutting transmission costs and enhancing privacy. For example, edge AI can process IoT data locally to instantly generate industry reports and predictive models to support enterprise decision-making.

Service Diversification: From Cloud to Cloud-Edge-Device Collaboration

With the booming development of AI applications, human-machine interaction is moving towards multimodal and personalized experiences. Interaction models are shifting from singular text or voice to multi-agent collaboration, encompassing both communication among AI agents and natural language interaction with humans. This transformation significantly enhances the intelligence and diversity of interaction.

In real-time interaction scenarios (e.g., AI glasses), the traditional cloud-based collaboration model is limited by network latency, making it difficult to meet low-latency requirements. Meanwhile, terminal devices are constrained by the "impossible triangle" of power consumption, size, and cost, making it challenging for them to independently support complex AI tasks.

The cloud-edge-device collaboration model breaks through these limitations by integrating the powerful computing capabilities of the cloud, the efficient processing of the edge, and the real-time responsiveness of the device. This provides users with a low-latency, highly immersive interactive experience, enabling  seamless, personalized delivery of new AI services.

AI for RAN: AI algorithms enhance RAN performance through intelligent resource allocation, proactive identification and resolution of potential issues before they impact service. It delivers reliable, high-bandwidth connectivity for AI agents, ensuring seamless handling of massive data workloads (e.g., video analytics and edge inference) with millisecond-level responsiveness.       

RAN for AI: New AI applications, such as AI agents, require low latency and strong uplink speeds (e.g., ubiquitous 20 Mbps) for instant HD image and video uploads. As the network element closest to users, RAN seamlessly integrates AI workflows to guarantee stable, low-latency connections through precise resource orchestration, while enriching AI application with real-time RAN data, such as enabling personalized recommendations and predictive insights.

The cloud-edge-device collaboration supports flexible deployment, differentiated experiences (e.g., multimodal interactions), and rapid expansion of the entire AI services ecosystem.

Global Practices in AI and RAN Integration

Global operators are embracing AI. Currently, AI for RAN is focused on boosting efficiency and differentiated experience, which is the fastest path to AI monetization. RAN for AI targeting new business models remain exploratory and scenario-dependent.

European operators: They focus on improving user experience and operational efficiency, pragmatically advancing AI for RAN. For instance, Deutsche Telekom’s Guardian Agent markedly improves O&M efficiency.

Chinese operators: The Chinese market is monetizing AI for RAN, quickly achieving monetization in terms of experience, energy efficiency, and O&M efficiency through the addition of intelligent computing boards to existing BBUs. China Mobile has deployed these boards across hundreds of thousands of sites by the end of 2025, leading global AI monetization.

Japanese & Korean operators: These operators are highly active in GPU-based AI-RAN. SoftBank, a member of the AI-RAN Alliance, launched the world’s first GPU-based AI-RAN in November 2024 to enable edge AI monetization, However, viable use cases remain scarce, and monetization prospects are uncertain.

The integration of AI and RAN injects strong momentum into business model innovation within the telecom industry. By achieving network capability monetization, data monetization and service diversification, operators are transforming from traditional connectivity providers into coordinators of the digital ecosystem, creating a vibrant "capability supermarket." This customer-centric model delivers personalized services, significantly enhancing user experience and satisfaction. It not only drives revenue growth and ecological collaboration but also provides solid support for the development of frontier fields like smart cities and autonomous driving, ultimately fueling the vigorous growth of the AI-driven digital economy.