Vodafone and Cirrus360 have completed trials of an AI-powered digital twin system designed to predict how 5G network equipment will perform before deployment or upgrade.

The technology creates a dynamic virtual replica of a physical 5G Radio Access Network (RAN) site, allowing engineers to simulate performance scenarios in advance. By applying AI-driven analysis to the model, the system forecasts hardware stress points such as CPU usage, memory constraints, and latency, as well as anticipated customer traffic loads.

The platform, Cirrus360’s cloud-based Gabriel™ system, can also identify potential failures and recommend hardware or software adjustments before rollout. Vodafone said the approach could reduce time spent testing multiple configurations — a process that typically involves evaluating six to eight RAN setups before nationwide deployment.

Unlike traditional digital twins that rely on rigid programming, Cirrus360’s “declarative” model uses rule-based specifications to adapt automatically as a physical asset moves from design to testing and live operation. The system defines performance objectives rather than fixed instructions, enabling automatic reconfiguration while ensuring compliance with operational boundaries.

Vodafone expects the technology to streamline testing, optimize existing resources such as underutilized CPUs, and accelerate deployment timelines. The method is particularly relevant for Open RAN architectures, which integrate components from multiple vendors.

Joint trials were conducted in the United Kingdom and the United States. Vodafone said it plans to broaden collaboration by opening its test facilities in Malaga, Spain, and Newbury, UK, to additional vendors.