Ericsson adds AI in RAN software for 5G network optimisation

Ericsson has introduced a software subscription called AI in RAN, adding telco-grade AI models to its radio access network products for 5G networks.

The software is designed to run AI functions inside basebands and radios without requiring operators to add new hardware. Ericsson said the first AI in RAN features will be available in the second quarter of 2026, with more updates planned later in the year.

Software-based AI for 5G networks

AI in RAN is built to work with Ericsson 5G Advanced across purpose-built RAN and Cloud RAN platforms. The software applies AI models in different parts of the radio network, including functions linked to performance, energy use, and automation.

The system uses Ericsson Silicon for AI inference in radios and the latest generation of RAN Compute. Cloud RAN software portability supports deployment across partner platforms.

The launch follows earlier Cloud RAN work involving AT&T, Ericsson, and Intel. In March 2026, Ericsson said the companies demonstrated AI-native Link Adaptation on a Cloud RAN stack powered by Intel Xeon 6 SoC, with tests showing up to 20% higher throughput compared with legacy rule-based link adaptation.

Ericsson said its AI models are designed for microsecond-level inference for radio network functions that depend on low latency. The models were developed with domain expertise, trained on high-quality data, and built for changing RAN environments.

Initial software features include an AI-native scheduler for link adaptation, AI-powered macro positioning, and AI-managed beamforming. Other features cover multi-layer coordination, performance management event schema files, and augmented observability for AI in RAN.

Link adaptation determines how the network adjusts transmission settings based on current radio conditions. Ericsson’s AI-native version uses real-time models to respond to changes in signal quality, interference, and other factors.

Trials and operator deployments

Ericsson has reported more than 15 deployments and trials worldwide. According to the company, the technology has delivered up to 20% higher downlink throughput and up to 10% better spectral efficiency.

It also reported support for up to twice as many high-traffic users, 90% to 95% coverage prediction accuracy, and up to five times better user-positioning precision.

Ericsson has also tested AI-based coverage prediction with Optus in 5G Standalone networks. The model predicts whether a device on one frequency layer is also within coverage of another, information used in network decision-making.

Ericsson named SoftBank, Bell, SK Telecom, and Rogers among the operators working with the software. The operators referred to network optimisation, spectrum efficiency, energy use, and AI-driven services in their comments.

SoftBank said the software could support low-latency services, including physical AI use cases that require coordination between network and compute resources.

SK Telecom said its work with Ericsson combines research, real-world validation, and software development as part of preparations for AI-native 6G.

Bell and Rogers referred to performance and energy efficiency in their comments. Rogers said it plans to use AI in RAN for real-time network optimisation and lower energy use.

Mobile Experts Principal Analyst Joe Madden said the software upgrade gives operators a way to increase capacity and improve observability using existing 5G network assets. He also pointed to more accurate location-based services.

Ericsson has added AI functions to its network products since the 4G era. In 2021, it introduced AI-ready acceleration in RAN Compute.

In February 2026, Ericsson unveiled Neural Network Accelerators in its Massive MIMO radios. Ericsson said those accelerators increased AI inference capability by ten times.

(Photo by Scott Rodgerson)

See also: T-Mobile and Ericsson test AI-RAN on live 5G Advanced network

Want to discover how IoT is transforming telecoms and connectivity? Join the IoT Tech Expo in Amsterdam, California, and London. Explore how innovations in 5G, edge computing, and IoT are shaping the future of networks and services. The event is part of TechEx and co-located with other leading technology conferences, click here for more information.

Telecoms Tech News is powered by TechForge Media. Explore other upcoming enterprise technology events and webinars here.