Ericsson has added artificial intelligence capabilities to its radio access network (RAN) software, aiming to improve how operator 5G networks are optimised.

The vendor has embedded AI directly into its RAN software portfolio rather than offering it as a separate analytics layer. The stated goal is 5G network optimisation: using machine-driven analysis to tune network performance in ways that manual configuration or traditional network management tools cannot match.

The move reflects where the radio access market currently sits. 5G rollouts have moved past their initial peak in most mature markets, and vendors are competing less on new radio hardware and more on software features that promise operators lower operational costs and better network performance from infrastructure they have already deployed.

For operators, AI-driven RAN optimisation addresses a practical problem. Modern multi-band, multi-layer 5G networks generate far more configuration and performance data than engineering teams can process manually. Automating that work — adjusting parameters, predicting issues, balancing load across cells — is the commercial argument Ericsson is making.

The announcement also fits the wider industry pattern. rival RAN vendors have been pushing similar AI and automation narratives for several years, positioning intelligence in the network as the next competitive battleground once 5G coverage build-outs saturate. Standards bodies and industry groupings have likewise been working on architectures that make RAN software more open to intelligent, automated control.

As with all vendor claims in this space, the distinction to watch is between demonstrated operator results and portfolio positioning. Ericsson is framing AI in RAN software as a core capability of its 5G offering; the proof will come in operator deployments and any measurable efficiency gains those customers report.

The company has not tied the announcement to a specific operator deployment or rollout timeline, so the near-term measure of its impact will be which service providers adopt the AI-enabled RAN software and what performance or cost results they disclose.