Transforming Radio Access Networks with AI

The RAN market has seen a downturn in previous years, yet a positive outlook persists as global RAN revenues are projected to stabilize in 2025, according to Dell’Oro Group.

This comes at a crucial time, as the telecom industry adapts to the developments and challenges in growth and monetization. As the industry navigates a transformative period, AI and RAN technologies are poised to revolutionize how mobile networks are deployed and managed.

The RAN Evolution

AI-RAN holds the potential to enhance telecommunications by accelerating network evolution and unlocking new possibilities. Its significant evolution is set to transform the industry, enhancing network performance, reducing operational costs, and enabling innovative services.

Despite past hurdles, the RAN market appears to be promising, with the global AI and RAN traffic optimization market projected to surge from USD 2.2 billion in 2024 to USD 27.2 billion by 2034, growing at a compound annual growth rate (CAGR) of 28.60% from 2025 to 2034. The market’s impressive growth is driven by the increasing demand for efficient network management, the rapid growth of mobile data traffic, and the increasing usage of smart devices.

The emergence of AI has changed industries in ways previously unimaginable, driving digital transformation, enabling cutting-edge innovations, and optimizing network performances. AI in RAN technologies will enable a myriad of capabilities as well, streamlining network operations by enabling advanced traffic management, optimizing distribution and expanding 5G coverage, while also facilitating predictive maintenance and real-time network analytics.

AI will substantially impact RAN technologies by reducing energy consumption. With 87% of an operator’s energy consumption driven by RAN, integrating AI will offer significant energy reduction. The AI-powered system will be able to switch off radio components when they are not needed, ultimately leading to energy savings.

NVIDIA’s research reflected that AI-RAN not only improves the network’s adaptability and efficiency but also enables telecom operators to enhance asset utilization by allocating resources across dedicated workloads in real time.

Furthermore, AI-RAN will facilitate automated network planning, configuration, and optimization. Traffic demand can then be predicted, improving operations, enhancing network configuration consistency, and minimizing errors.

AI-RAN will promote intelligent automation and orchestration, which will eventually lead to more comprehensive decision-making for future technologies, particularly 6G networks. AI-RAN will drive innovative use cases, such as ultra-reliable low latency communication (URLLC), ubiquitous low latency broadband communication (uMBB), and enhanced mobile broadband communication (eMBB), and massive machine-type communication (mMTC).

Also Read: How the Telecom Industry May Be Recovering from the RAN Market Downturn

Challenges in AI-RAN

AI's integration with RAN technologies offers transformative opportunities; however, several challenges must be addressed to ensure widespread adoption.

In Q1 2024, Dell’Oro Group highlighted a 15-30% decline in the overall 2G-5G RAN market. Sharp declines were observed in the North American, European, and Asia-Pacific RAN markets.

Given the diverse nature of telecom data, AI-RAN can potentially create inconsistent data quality, leading to inaccurate AI predictions. Telecom operators require highly reliable AI solutions to prevent service disruptions, making system availability a critical necessity. Despite their potential, the industry is concerned that AI models could possibly misinterpret data, potentially leading to incorrect suggestions and inefficiencies. Algorithmic biases may also incur regulatory and ethical issues, making transparency a critical key to promote responsible AI usage. Similarly, security and privacy concerns also persist, particularly in managing user-sensitive data in a distributed environment and collecting comprehensive data.

Moreover, high implementation costs and the lack of a knowledge-equipped workforce pose further barriers to widespread deployment. Given the global technical skills shortage, a workforce with knowledge in AI-enabled RAN and skilled in AI, network infrastructure, and data analytics would be required to accelerate future developments.

Advancements in AI-RAN

The industry is experiencing significant developments in the RAN market which highly contribute to the sector’s promising future.

In 2024, the AI-RAN Alliance was formed, gathering industry leaders including Amazon Web Services (AWS), Ericsson, Microsoft, Nokia, NVIDIA, and T-Mobile, among others. The AI-RAN Alliance aims to enhance mobile network efficiency, reduce power consumption, and lay the groundwork for AI-driven telecommunications.

According to Alex Jinsung Choi, Chair of the AI-RAN Alliance, and Principal Fellow of SoftBank Corp.’s Research Institute of Advanced Technology, the alliance’s main mission is to “weave AI right into the fabric of the radio access network.”

We are not just about making networks faster and more flexible; we want to transform them into self-organizing, self-optimizing, self-managing systems that can handle real-time changes and anticipate maintenance and efficiently manage resources.

In the same year, Ericsson, T-Mobile, and NVIDIA collaborated to establish the joint AI-RAN Innovation Center to accelerate standardization and industry-wide adoption of AI-RAN technologies. By developing AI-RAN, the initiative is poised to deliver transformational network experiences to customers.

Similarly, SoftBank partnered with Ericsson, exploring joint AI-RAN architecture to develop AI-integrated RAN solutions that can significantly enhance network efficiency and performance. In addition, SoftBank launched the AITRAS—an innovative solution enabling AI and RAN to operate on the same NVIDIA-accelerated computing platform. The AITRAS implementation allows telecom operators to invest in existing RAN infrastructure while building AI capabilities, creating more efficient network operations.

Early this year, Saudi-based telecom operator, Zain KSA, and global technology giant, Nokia, partnered to drive 5G Cloud RAN and AI-RAN solutions by advancing automation capabilities and real-time data analytics. The initiative is poised to deliver enhanced user experience and operational efficiency.

To reduce energy consumption across network operations, Umniah Jordan collaborated with Ericsson. By utilizing Ericsson’s intelligent RAN power saving solution, Umniah’s energy management and operational efficiency will be enhanced, while operating costs and carbon dioxide emissions (CO₂) will be reduced.

The adoption of AI-RAN technologies has also shown promising results in energy efficiency with Taiwan’s Far EasTone saving 25% in daily RAN energy consumption using Ericsson’s Service Continuity AI app suite.

A New Era of Sustainable Connectivity

Combining AI with RAN represents a paradigm shift in telecommunications, unleashing new levels of advanced capabilities set to transform network solutions and drive new revenue streams.

This integration transcends beyond operational enhancements, solidifying its significance in the future of telecommunications infrastructure. As communication service providers (CSPs) face the social responsibility to reduce the global energy demand, AI-RAN will significantly decrease the environmental footprint in an intelligent way, advancing sustainability.

AI-RAN will emerge as a critical tool for operators to meet the growing demands for 5G and will lay the groundwork for futuristic 6G technologies. Intelligent networks equipped with AI-RAN capabilities are bound to usher the industry into a new era of connectivity, delivering exceptional user experiences while maintaining operational efficiency and supporting sustainability.

Ooredoo Qatar, Ericsson Collaborate on 5G RAN and Microwave Backhaul Solution

The Race to 6G: What’s Coming in 2025 and Beyond?