Deploying private 5G networks in space-constrained environments often forces a trade-off between performance and physical footprint.

While industrial strategy frequently calls for private 5G to support automation and data sovereignty, the reality of fitting data centre-grade hardware into retail branches or small manufacturing sites halts many rollouts before they begin.

The challenge is not just connectivity, but the infrastructure overhead required to support it. Branch offices and remote industrial sites rarely possess the square footage or power budget to accommodate traditional server racks. Operators are consequently tasked with minimising wattage to control cooling costs while delivering the low latency required for modern applications.

Ji-Yun Seol, Executive VP and Head of Product Strategy, Networks Business at Samsung, explained: “If you are involved in deploying and operating private networks for enterprises, you might be familiar with this gap between promise and practicality—a challenge often encountered by system integrators, managed service providers, and operators.”

To bridge this gap, enterprises are moving away from proprietary hardware stacks toward virtualised architectures that can coexist with existing compute resources.

Consolidating the stack for private 5G networks

Samsung has responded to these density challenges with a ‘Network in a Server’ (NIS) configuration. The solution consolidates private network functions – including the mobile core, Radio Access Network (RAN), transport, and AI agents – onto a single platform.

This architecture abandons specialised telecoms hardware in favour of a software-driven model running on commercial off-the-shelf (COTS) servers. By virtualising containerised network functions, the system eliminates the need for separate physical appliances. It leverages standard compute resources to execute network operations and edge AI services concurrently.

The shift to general-purpose hardware alters the unit economics of deployment. Reducing the physical equipment count lowers shipping logistics, rack space requirements, and total power consumption.

For the enterprise, this consolidation translates to reduced operating expenses and a simplified support model with a single point of contact. Samsung applies its background in virtualising macro cell components to this enterprise-grade implementation.

Hardware and ecosystem integration

The performance of a single-server private 5G network relies on the underlying silicon. Samsung collaborated with AMD, Supermicro, and Wind River to engineer the NIS ecosystem.

The platform utilises an AMD EPYC 8000 Server CPU. This integration represents the commercial launch of Samsung’s virtualised portfolio on AMD processors. The server configuration supports GPUs alongside the CPU, enabling the hardware to manage specific vRAN processing loads while reserving capacity for AI and real-time analytics.

Embedding AI capabilities directly within the network server facilitates local data processing. This architecture is necessary for environments requiring immediate responsiveness, as it reduces latency by avoiding data round trips to a central cloud. Local data retention also supports compliance for government agencies and institutions with strict on-premises security mandates.

Samsung validated the private 5G system against industrial scenarios demanding millisecond-level network response times. These use cases depend on immediate analysis where latency creates operational risk:

• Video analysis: The system processes feeds from CCTV cameras on the private network to flag safety events. Applications include detecting missing hard hats on construction sites, identifying lost items in secure zones, or early fire detection.

• Integrated Sensing and Communication (ISAC): This technology utilises radio signals for sensing applications without disrupting connectivity. Use cases include parking space detection, drone tracking, and vehicular monitoring for road safety.

• Connectivity for emerging devices: The solution supports hardware such as AR glasses or XR headsets. These tools overlay contextual data in a user’s field of vision, such as player statistics at a stadium or participant details during business meetings.

Not your regular connectivity

As AI adoption accelerates in industrial settings, these services are becoming central to next-generation operations. Running physical AI and other applications on the same infrastructure used for connectivity allows enterprises to explore revenue models beyond standard access fees.

Combining network and compute resources offers a method to deploy private 5G in locations previously rejected due to space or power limits. Reducing physical complexity allows organisations to focus on applications that drive efficiency rather than managing the supporting hardware.

The convergence of RAN and AI compute on a single server alters the ROI calculation for private 5G. It reduces the deployment friction caused by space and power limits in edge locations like retail branches or small factories. However, successful implementation requires assessing whether a single-server architecture offers sufficient redundancy for mission-critical operations compared to traditional multi-node setups.

See also: 6G may arrive as sensing infrastructure, not just connectivity

Want to learn about the IoT from industry leaders? Check out IoT Tech Expo taking place in Amsterdam, California, and London. The comprehensive event is part of TechEx and is co-located with other leading technology events including AI & Big Data Expo and the Cyber Security Expo. Click here for more information.

Telecoms is powered by TechForge Media. Explore other upcoming enterprise technology events and webinars here.