Amdocs, 1Finity and Supermicro test an AI-RAN blueprint built on NVIDIA infrastructure, aiming for more autonomous network operations

Amdocs, along with 1Finity and Supermicro, said it has validated an AI-RAN (radio access network) blueprint using NVIDIA-accelerated infrastructure. The companies presented the effort as a step toward creating a unified, multivendor foundation for autonomous network operations, where software can observe network conditions, identify problems, and trigger optimization actions.

In the validation, the partners described how integrating the AI components into RAN operations can improve performance and help with anomaly detection, with the stated goal of tuning networks more efficiently. The group also said the resulting platform is cloud-native and intended to support latency-sensitive applications, including use cases expected to expand at the edge.

A key element of the announcement is that the capabilities they validated are codified as a Network Workflow inside Amdocs aOS, according to the report. In practical terms, this means the workflow is designed to let service providers deploy AI agents that can sense, reason, and act on network conditions rather than relying only on traditional monitoring and manual interventions.

The partners further positioned the blueprint as a framework that can help communications service providers (CSPs) move beyond connectivity-only operations. By embedding intelligence in the infrastructure that manages network behavior, the effort is presented as a way to make operations more resilient and more adaptable to changing traffic patterns or service demands.

The companies did not provide additional quantitative results in the reporting that was available for this write-up, such as measured reductions in downtime, changes in mean time to detect issues, or specific performance benchmarks for the validated workflow. They also did not disclose any timeline for general availability or specify which exact NVIDIA hardware configurations or software stacks were used in the demonstration beyond the report’s general reference to NVIDIA-accelerated infrastructure.

Amdocs is a provider of software and services for communications and media companies, with offerings that span network transformation, billing and monetization tooling, and cloud platforms. In its portfolio description included in the reporting, the company references products and agent-related initiatives, positioning its platform as a place where AI-driven functions can be operationalized for network and service workflows.

For NVIDIA, the announcement reinforces its broader strategy of supporting AI workloads that can run in data centers and at the edge. For telecom equipment and services partners like Supermicro and 1Finity, the message is that AI for RAN operations is moving from experimentation toward packaged workflows that can be deployed in real networks, at least in a test and validation setting.

What remains unclear from the available reporting is how widely the validated blueprint has been deployed, whether multiple operators are already running it, and what performance trade-offs might exist when moving from a validation environment to live commercial networks. The next items to watch are any details on deployment, partner ecosystems, and whether Amdocs aOS network workflows and the AI-RAN components are being integrated into larger customer programs.