vRan takes a leap of maturity: Samsung completed the first commercial live network call of a Tier‑1 operator in the United States using its cloud-native platform on a single Cots HPE server, with Intel Xeon 6 and Wind River cloud. The transition from the laboratory to the operational network changes the terms of the debate on the efficiency and complexity of the virtualized Ran, because it shows that a single-server topology can support carrier-grade requirements and, above all, concentrate multiple network functions in a single node, cutting energy and operational costs. The company links the milestone to the trajectory towards AI-native and 6G-ready networks, where local processing for artificial intelligence becomes an integral part of the radio architecture.
From the laboratory test to the network: why it really matters
Table of Contents
- From the laboratory test to the network: why it really matters
- The impact on networks: consolidation, energy and Tco
- AI‑native as a paradigm: from optimization to the new user experience
- Towards 6G: Why single-server accelerates the transition
- The actors’ statements
- What changes for operators: roadmap and risks to monitor
- The competitive framework
In 2024 Samsung had validated the call end-to-end in a test environment with Xeon 6; today it links that result to the first call on the commercial network, a crucial step to overcome doubts about the performance, latency and reliability of Ran on general-purpose servers. The session takes place on standard, non-proprietary hardware, and demonstrates that baseband, mobile core, transport and security can coexist on the same host without penalty. With more cores (up to 72) and built-in acceleration for AI and vRan, the platform enables more efficient use of memory and lower power consumption compared to the previous generation. Looking ahead, operators can scale via software, orchestrate capabilities where needed and speed up the release of advanced functions.
The impact on networks: consolidation, energy and Tco
The key principle is consolidation. With the vRan of Samsung on a single server, operators concentrate software elements traditionally distributed across multiple machines. The immediate benefit concerns site engineering and operations: less equipment to manage, simplified configurations, lower power budget, lower Capex and Opex. The effect si estende al lifecycle management: a common foundation enables coordinated updates, faster rollbacks, and pushed automation. At scale, the convergence of Ran and AI loads on the same host creates a computational fabric ready to host advanced optimization, scheduling and RRM algorithms, with real-time telemetry and adaptive energy-saving policies.
AI‑native as a paradigm: from optimization to the new user experience
The element of discontinuity is not just technological. AI-native networks shift the focus from hardware to the ability to learn from radio and transport data, to anticipate congestion, proactively choose beamforming and MIMO, and orchestrate resources with goal-driven logic. In this scenario the vRan it represents not only a different way of doing baseband, but the platform on which low latency AI engines are grafted, close to the point of emission of the signal. The expected result: stability, better perceived coverage, and more linear energy consumption during peak and valley hours.
Towards 6G: Why single-server accelerates the transition
6G aims for cloud-centric services with intelligent control loops at the edge of the network. Bringing Ran components and AI inference to a single server avoids traffic bounces, reduces overhead and improves flow deterministicity. Furthermore, the software-defined approach makes networks more programmable and aligns mobile infrastructure with cloud-native principles already established in the core. It is the basis for AI‑Ran functions that adjust power, bandwidth and QoS in response to context, preparing the adoption of 6G use cases such as integrated sensing and distributed edge orchestration.
The actors’ statements
“This milestone represents a major step forward in network virtualization and efficiency. It confirms the readiness of this technology in real-world conditions, demonstrating that single-server vRan deployments can meet the performance and reliability standards required by major operators,” he said June Moon, Executive Vice President, Head of R&D, Networks Business di Samsung Electronics.
“With Intel Cristina Rodriguez, vice president and general manager of Intel’s Network & Edge division.
What changes for operators: roadmap and risks to monitor
The industrialization of the model will depend on three variables. Primothe portability of functions on different Cots platforms, to avoid lock-in. Secondcoherence between cloud platforms and hardware accelerations, so as to obtain predictable benefits on different configurations. Thirdintegration with the existing fleet, including multi‑vendor and O‑Ran scenarios. Here vRan on a single server can help: it reduces the number of components, simplifies upgrade paths and enables autonomy in life cycle management. However, rigorous governance on observability, security and capacity planning will be needed to avoid resource contention between Ran flows and AI loads.
The competitive framework
With this stage, Samsung strengthens its positioning in the virtualized Ran compared to the large incumbents. The focus on efficiency and complexity reduction responds to historical concerns about the computational cost of cloud-based architectures. If the single-server model holds up across large deployments and across heterogeneous sites, operators will be able to accelerate RAN modernization without taking long time to replace dedicated hardware. In parallel, AI-ops-focused telcos will have a foundation to experiment with low-latency AI-Ran applications, from adaptive traffic steering to on-cell power management.
The networking of vRan on a single server marks a turning point: not only does it validate the technical feasibility, but it highlights a measurable operational advantage. Converging Ran, core and AI functions on the same host enables a faster innovation cycle and sets the stage for AI-native 6G built on software, automation and sustainability. For operators, the stakes are clear: simplify the infrastructure, contain consumption and make the radio network programmable, transforming it into a platform for intelligent services ready to scale.
