AI in Open RAN: Role, requirements, and challenges

How can artificial intelligence help telecom operators build more efficient, scalable, and interoperable Open RAN (O-RAN) networks? As O-RAN adoption grows, AI is playing a crucial role in optimizing performance, automating network functions, and managing complex multi-vendor ecosystems. However, challenges remain in integrating AI into telecom infrastructure, balancing real-time processing needs and addressing regulatory concerns.

Industry leaders recently explored what it takes to achieve a truly open, intelligent, and efficient RAN ecosystem in a panel moderated by Abe Nejad, Publisher of Network Media Group (NMG). The session featured experts from Vodafone, AMD, and Rakuten Symphony, discussing the role of AI in Open RAN, the requirements for large-scale deployment, and the challenges operators must overcome.

Speakers:

• Muslim Elkotob - Principal Architect, Vodafone
• Gilles Garcia - Senior Director Business Lead Wired and Wireless Group, AMD
• Mohammed Talaat - Global Head of Presales & Solution Architecture, Rakuten Symphony

Watch the full video here.

‍

The discussion highlighted that AI is essential for managing the high level of interoperability and automation required in O-RAN. Unlike traditional RAN architectures, where hardware and software come from a single vendor, O-RAN networks integrate components from multiple suppliers. This creates new operational complexities, and AI is emerging as a solution to address them.

One of the key areas where AI is proving its value is in real-time network optimization. AI-powered automation enables operators to adjust network parameters dynamically, improving spectral efficiency, reducing interference, and enhancing user experience. AI is also being used to automate traffic management, predictive maintenance, and resource allocation, ensuring networks operate smoothly with minimal human intervention.

Challenges in deploying AI-driven Open RAN

While AI is helping unlock new efficiencies, integrating AI into O-RAN networks can be challenging due to:

• Data availability and standardization – AI relies on large datasets to train models, but telecom data is often fragmented across different systems and vendors. Ensuring that datasets are standardized and accessible is critical for AI-driven O-RAN deployments.
• Computing constraints at the edge – AI processing can be compute-intensive, and edge locations often lack the necessary power and infrastructure to support real-time AI workloads.
• Security and trust in AI models – With O-RAN networks involving multiple vendors, ensuring trust and security in AI-generated decisions is critical. Operators must address concerns around data privacy, regulatory compliance, and multi-vendor collaboration to make AI work seamlessly.

Operators need to be strategic about where and how they deploy AI to maximize impact without overloading network resources.

Monetizing AI in Open RAN networks

Beyond operational improvements, AI is also creating new revenue opportunities for telcos. The discussion covered how AI-driven O-RAN can support:

• AI-powered network slicing – Enterprises can purchase dedicated network slices optimized for their specific needs, such as low-latency industrial applications or high-reliability financial transactions.
• Energy-efficient network operations – AI helps telecom operators reduce power consumption by dynamically managing network resources, leading to cost savings and sustainability benefits.
• Automated service assurance – AI-driven predictive analytics help operators proactively detect and resolve issues, improving network reliability and customer satisfaction.

These innovations are positioning telecom operators not just as connectivity providers but as intelligent network service enablers, offering new capabilities beyond traditional mobile services.

The conversation wrapped up with a discussion on the long-term role of AI in Open RAN. While AI is already proving its value in network automation and optimization, future advancements will focus on:

• AI-driven self-healing networks that detect and fix faults autonomously.
• Greater regulatory clarity on AI in telecom to address cross-border data concerns.
• Hardware acceleration for AI workloads at the edge, ensuring AI can operate efficiently without requiring centralized cloud processing.

The panelists agreed that AI in O-RAN is still in its early stages, but as AI models become more sophisticated and deployment challenges are addressed, the industry will see even greater automation and intelligence in telecom networks.

"If we focus initially on some of the vital AI use cases that are relevant to the telco industry, then we can start to build them, step by step.”

- Mohammed Talaat, Global Head of Presales & Solution Architecture, Rakuten Symphony

Follow us

Stay tuned for more insights from industry leaders on the future of telecom and technology. Follow us for updates on upcoming discussions.