The integration of Artificial Intelligence (AI) into Radio Access Networks (RAN) is set to revolutionize telecom infrastructure, promising smarter, more efficient networks capable of handling the growing demands of 5G and beyond. From optimizing network performance to enabling real-time decision-making at the edge, AI is transforming how operators manage complex systems. In a recent discussion, industry leaders explored how AI-driven innovations—backed by the AI RAN Alliance and advancements in edge computing—are creating new opportunities to enhance RAN efficiency, unlock revenue streams, and redefine the future of connectivity.
In a recent interview Aji Ed, VP and Head of Cloud RAN at Nokia, discusses key initiatives, technical developments, and collaborative opportunities shaping the future of telecom infrastructure.
The AI RAN Alliance
The AI RAN Alliance, formed earlier this year by industry leaders like T-Mobile, SoftBank, and Nokia, focuses on integrating AI into RAN. The alliance is built around three pillars:
1. AI for RAN – Utilizing AI to improve network efficiency, such as optimizing beamforming and pattern adjustments.
2. AI on RAN – Supporting AI-driven applications that operate on top of RAN, like distributed AI processing and monetization platforms.
3. AI and RAN – Exploring infrastructure convergence to support both AI workloads and RAN functionalities, potentially reducing costs and enabling edge AI services.
Building on Existing Capabilities
When asked how “AI on RAN” leverages existing work, Aji explained that it evolves from current platforms rather than requiring a complete reinvention. For example, network monetization platforms today utilize APIs that can be enhanced with AI to deliver new applications, ensuring the approach builds on established industry foundations.
Telco Cloud vs. Hyperscalers
Addressing the interaction between telco cloud infrastructure and hyperscalers, Aji highlighted that most AI workloads currently run in centralized data centers. However, latency-sensitive tasks like AI inferencing could benefit from edge processing, especially for distributed AI applications. Operators and hyperscalers will need to weigh the costs and benefits of moving workloads closer to the edge versus keeping them centralized.
Edge Devices and Convergence
On the topic of new edge devices, Aji noted that future infrastructure may not require standalone equipment. Instead, RAN and AI workloads could converge on a shared platform, enabling edge processing capabilities while simplifying deployment.
Collaboration with Open RAN
Aji clarified that while the AI RAN Alliance and Open RAN are distinct initiatives, they share complementary goals. Open RAN focuses on standardizing interfaces, while the AI RAN Alliance explores AI-driven innovations. Both groups are expected to collaborate on topics like codification, ensuring alignment and progress in key areas.
Looking Ahead
As AI-driven innovations continue to reshape the telecom landscape, collaboration among industry players will be essential. The convergence of AI and RAN workloads, edge processing, and ongoing initiatives by the AI RAN Alliance and Open RAN signal a future rich with opportunity and innovation.
The telecom industry is poised for significant evolution by 2025, with AI playing a central role in creating smarter, more efficient networks.
