Global telecommunications giant Ericsson has launched a new artificial intelligence-driven solution designed to transform mobile network performance, efficiency, and automation as demand for AI-powered services accelerates worldwide.
The new offering, dubbed AI in RAN, is a software-based subscription that integrates telco-grade AI models directly into radio access network (RAN) infrastructure, enabling real-time optimisation without the need for additional hardware.
According to the company, the solution marks a significant step toward the evolution of AI-native networks, providing communications service providers (CSPs) with immediate performance gains across 5G systems.
AI in RAN introduces advanced capabilities including real-time AI models embedded within basebands and radios, continuous learning powered by high-quality data, and agentic AI support to automate network operations.
Mårten Lerner, Head of Networks Strategy and Product Management at Ericsson, described the innovation as a breakthrough in mobile network evolution.
“Ericsson is redefining what’s possible in mobile networks by bringing powerful AI capabilities to service providers,” he said. “With AI in RAN software, we are taking a major step toward AI-native networks.”
The system is engineered for ultra-low latency processing at microsecond speeds, ensuring reliability and efficiency even in complex and dynamic network environments.
Ericsson says early deployments and trials in more than 15 markets have demonstrated significant improvements, including up to 20% higher downlink throughput and 10% gains in spectral efficiency.
The technology also enables networks to support up to twice as many high-traffic users, while delivering up to 95% accuracy in coverage prediction and up to five times greater precision in user positioning.
In addition to performance improvements, the solution is designed to enhance energy efficiency through optimised AI inference, helping operators reduce operational costs while maintaining high service quality.
Leading telecom operators including SoftBank, Bell, SK Telecom, and Rogers have welcomed the development, highlighting its potential to improve real-time network optimisation, boost efficiency, and support next-generation AI-driven services.
Industry analysts also view the solution as a high-return investment opportunity for operators seeking to maximise existing 5G infrastructure through software upgrades rather than costly hardware expansion.
AI in RAN is compatible with Ericsson’s 5G Advanced and Cloud RAN platforms, enabling flexible deployment across both dedicated and cloud-based network environments.
The first set of features is expected to roll out in the second quarter of 2026, with additional enhancements planned later in the year. These include AI-native scheduling, intelligent beamforming, macro positioning, and enhanced observability tools.
The launch builds on Ericsson’s broader AI strategy, which has seen the company steadily integrate AI capabilities into its network solutions since the 4G era, with significant advancements in AI-ready hardware and software announced in recent years.
As global demand for high-performance, low-latency connectivity continues to rise, solutions like AI in RAN are expected to play a central role in shaping the future of telecommunications and paving the way for next-generation technologies such as AI-driven 6G networks.
