Towards a Smarter RAN: Ericsson Unveils New Software Embedding AI in the Optimal Location

This document is a translated summary of a press release issued on June 11, 2026.

With the growing adoption of AI-driven services, mobile networks are facing increasing demands for higher performance and intelligent automation. Today, Ericsson announced 'AI in RAN', a software subscription that embeds carrier-grade AI models into basebands and radios to enhance efficiency, performance, and energy savings. This scalable, commercially ready solution delivers immediate benefits to operators’ 5G networks and supports a seamless transition to AI-native Radio Access Networks (RAN) without requiring additional hardware.

Ericsson's 'AI in RAN' introduces the following capabilities:

- Carrier-grade AI models: Designed to operate in real time within the RAN. - Continuous learning capabilities: Software powered by scalable, high-quality data. - Support for Agentic AI: Enables advanced RAN automation and network operations.

Mårten Lerner, Head of Network Strategy and Product Management at Ericsson, said: "Ericsson is redefining the potential of mobile networks by delivering powerful AI capabilities to operators. With the 'AI in RAN' software, we are taking a major step toward AI-native networks, complementing the AI-enabled radios we announced in February."

Ericsson's carrier-grade AI models are designed to deliver ultra-low latency inference in microseconds, providing high reliability and robustness in diverse and dynamic RAN environments. Developed with deep domain expertise and trained on high-quality data, these models set new benchmarks for performance and efficiency.

"In the AI era, operators need high-performing, secure, and efficient networks. Through 'AI in RAN', Ericsson brings AI into the network, enhancing 5G performance and efficiency through large-scale, energy-efficient AI inference," said Lerner.

Proven in Commercial Networks

Ericsson's 'AI in RAN' works with both dedicated RAN and Cloud RAN platforms in conjunction with Ericsson 5G Advanced to enable new AI-driven services. This realizes Ericsson's 'AI for Networks' vision of leveraging AI to improve network performance and efficiency.

'AI in RAN' is designed to place the right AI models in the optimal locations within the radio network. It leverages Ericsson Silicon for energy-efficient AI inference within radios and the latest generation of RAN Compute, while Cloud RAN software portability enables deployment of AI capabilities across partner platforms.

Initial features of 'AI in RAN' will be available in Q2 2026, with further expansions planned for the second half of the year. Software features to be introduced include an AI-native scheduler for link adaptation, AI-based macro positioning, AI-controlled beamforming, AI-based multi-layer coordination, performance management event schema files, and extended observability for 'AI in RAN'.

Proven in over 15 commercial networks and trials worldwide, Ericsson's 'AI in RAN' increases downlink throughput by up to 20% and improves spectral efficiency by up to 10%. It also doubles the number of high-traffic users supported, enables 90–95% coverage prediction accuracy, and enhances user location estimation accuracy by up to five times.

Koji Oya, Executive Managing Director and CNO at SoftBank Corp., said: "Ericsson's 'AI in RAN' software is a crucial step toward broader AI implementation in radio networks. It enables real-time optimization of radio performance, spectrum efficiency, and user experience quality, turning AI innovation in commercial networks into tangible value. We greatly anticipate this foundation accelerating the societal deployment of new AI-driven services, including physical AI use cases that require low-latency, high-reliability connectivity and advanced coordination between network and computing resources."

Bruce Dean, Senior Vice President of Network Technology & Operations at Bell, said: "At Bell, we are constantly evolving our network to meet the growing demand for high-performance AI-driven services. Integrating AI directly into the RAN is a critical step in enhancing network intelligence and efficiency. By partnering with leaders like Ericsson to deploy these capabilities, we will deliver improved performance, better energy efficiency, and an enhanced experience for our customers."

Yu Takki, Head of Network Technology Office at SK Telecom, said: "SK Telecom is advancing AI-RAN in collaboration with Ericsson to improve network performance and energy efficiency while supporting more intelligent and automated operations. By combining research, real-world validation, and software innovation, we aim to strengthen our leadership in AI-driven network evolution and contribute to building the foundation for AI-native 6G."

Mark Kennedy, CTO at Rogers, said: "As the provider of Canada’s best 5G+ network, we are proud to partner with Ericsson to bring the latest 5G technologies to Canadians. With the deployment of 'AI in RAN', we will optimize network performance for our customers in real time while further advancing energy consumption reduction."

Joe Madden, Principal Analyst at Mobile Experts, said: "This could be one of the most cost-effective initiatives for operators in recent years. With a software upgrade, operators can extract more capacity, better observability, and more accurate location-based services from their 5G networks purchased years ago."

Ericsson is integrating AI capabilities across its entire portfolio of products from 4G onward and, in 2021, introduced AI into RAN Compute.