Ericsson launches carrier-grade AI subscription to boost 5G network performance

This shift to software-defined network intelligence allows operators to extract significant capacity gains from existing 5G assets. By embedding microsecond-level AI inference within the radio unit, Ericsson addresses the latency limitations of centralized AI while providing a scalable path to AI-native 6G. For the streaming ecosystem, this improves the reliably of high-bandwidth delivery in dense urban environments and reduces the energy footprint of massive data consumption. Industry observers should monitor how this purpose-built ASIC approach competes with general-purpose GPU alternatives from rivals like Nokia and Nvidia.

At Mobile World Congress in March 2026, Ericsson and Intel deepened their long-term partnership by aligning roadmaps to integrate Intel 18A process nodes and Xeon 6 processors with Ericsson’s software-defined stack. Per SatNews (March 2026), this collaboration aims to transform cell sites into high-density compute hubs capable of real-time AI inference at the network edge. The joint effort explicitly targets Level 4 network autonomy, positioning AI not as an add-on but as a fundamental component of the 6G standard.

Competitive pressure is mounting as rival vendors pursue differing architectural philosophies. According to Telecoms.com (March 2026), Nokia and Indosat Ooredoo Hutchison successfully demonstrated a GPU-accelerated 5G call using the Nvidia Aerial platform. While Ericsson advocates for purpose-built silicon to maximize energy efficiency, Samsung has taken a hybrid approach. Per Light Reading (April 2026), Samsung is validating AI-RAN on live networks in Korea while its leadership remains vocal about the high power consumption and thermal challenges associated with deploying full-scale GPUs at mobile sites.

Simultaneously, the broader telco ecosystem is moving toward "Physical AI" use cases. Per Ericsson and SoftBank (June 2026), the new software supports coordination between network and compute resources for low-latency robotics and autonomous services. This transition is highlighted by the launch of joint innovation hubs, such as the T-Mobile and Deutsche Telekom 6G project, which focuses on autonomous networks and high-precision sensing. Analysts at Mobile Experts noted in June 2026 that the software-based ROI for these upgrades is among the highest in years, as it allows operators to manage the "capacity gap" forecast for late-decade urban density.