AT&T and T-Mobile test Ericsson AI-native RAN feature

AT&T and T-Mobile tested Ericsson’s AI-native link adaptation feature in radio access networks (RAN) without using graphics processing units (GPUs), according to a Light Reading report.

TL;DR

• AT&T and T-Mobile tested Ericsson’s AI-native link adaptation feature in RAN.
• The feature aims to improve spectrum efficiency and network performance.
• The reported tests did not use GPUs.

Ericsson’s AI-native link adaptation feature applies artificial intelligence in the RAN, the part of a mobile network that connects user devices to cell sites. Light Reading reported that AT&T and T-Mobile used the feature in tests focused on spectrum efficiency and performance.

The report said the tests ran without GPUs. That detail matters because GPUs are commonly used to accelerate artificial intelligence workloads, while radio networks typically rely on other processing architectures in deployed equipment.

The source description did not provide test dates, performance figures, deployment locations, or whether the feature is in commercial use. The reported participants were AT&T, T-Mobile, and Ericsson.

Related questions

• What is AI-native link adaptation in a radio access network?
• Why would operators test AI features in RAN without GPUs?
• Which companies were named in the reported RAN AI tests?