Private 5G Networks in Factories: When It Makes Sense
Table of contents
- Key takeaways
- What a private 5G is
- Typical architecture
- Cases where private 5G shines
- Comparison with Wi-Fi 6/7
- Open RAN: lowers the entry barrier
- Co-located edge compute
- When it doesn't make sense
- Conclusion
- Frequently asked questions
- When is Wi-Fi 6/7 enough, and when do I need a private 5G network in the plant?
- How much does a private 5G deployment cost and how long does it take for a medium plant?
- Do I need my own spectrum, or is a slice of the operator's network enough?
- Sources
Private 5G networks deliver high capacity, low latency, and thousands of connected devices for factories that do not want to depend on a carrier or settle for Wi-Fi. They make sense in large campuses with mobility or high IoT density; for medium plants with fewer than 50 devices, Wi-Fi 6/7 remains more cost-effective.
Private 5G networks are the connectivity proposal for industrial environments that don’t want to depend on public operators or settle for Wi-Fi. They bring high capacity, low latency, broad coverage, thousands of connected devices, and criticality slicing. After years of hype, there are real installations with measurable metrics. This article covers architecture, comparison with alternatives, costs, and when investment makes sense.
Key takeaways
-
A private 5G network is radio + core operated by the enterprise on dedicated spectrum, not an operator "slice".
-
Private licensed spectrum (3.7-3.8 GHz in Germany; auction-based concession in Spain) gives real independence vs. public-network slicing.
-
Cases with proven ROI are large campuses with mobility, critical URLLC latency, or high IoT sensor density.
-
Wi-Fi 6/7 remains the right choice for most medium-size plants with fewer than 50 mobile devices.
-
Typical startup cost is €500k-2M for a medium plant; scarce talent is the most underestimated barrier.
What a private 5G is
A private 5G network is a radio + core deployment operated by the enterprise (or contracted provider) over dedicated spectrum. Three modalities:
-
Private licensed spectrum: national allocations for campus networks, such as the 3.7-3.8 GHz band the German Bundesnetzagentur licenses individually to each plant. In Spain the equivalent mechanism is a private-use concession awarded by auction[1]. The 3800-4200 MHz band has also been harmonised at EU level as shared low/medium-power spectrum, applicable from 2026. It is a less exclusive model than Germany’s but one that widens the available options.
-
Shared spectrum: CBRS in US, opportunistic use.
-
Public network slicing: a "virtual slice" within the operator’s network. Less independent.
For serious industry, private licensed spectrum is the preferred option.
Typical architecture
Essential elements:
-
Radio Access Network (RAN): antennas and radios distributed across the plant.
-
Edge core: the 5G core (UPF, AMF, SMF) on-site or near-edge to minimise latency.
-
SIM/eSIM management: manage device identities.
-
IT/OT system integration: tends to run over VLAN or IP.
Leading providers: Nokia, Ericsson, Siemens + Nokia, Athonet (Hewlett Packard), Celona, Druid, and the Open RAN family (Parallel Wireless, Mavenir).
Cases where private 5G shines
Scenarios with proven ROI:
-
Autonomous vehicles in large campuses: AGVs, mine trucks, drones. <20 ms latency and reliable handover are key.
-
AR/VR for maintenance: augmented glasses with edge-served technical content.
-
Large IoT sensor density: thousands per km². 5G scales better than Wi-Fi at high density.
-
Mobile / outdoor plant: construction, ports, logistics over large areas.
-
Redundancy and determinism: URLLC for critical control where packet loss is unacceptable.
There are four real, verifiable cases. BMW Group at its Landshut plant[2] uses 5G for autonomous logistics, with self-driving forklifts that process camera data in the cloud instead of onboard. Bosch[3] runs a private 5G campus network with Nokia at its Industry 4.0 lead plant in Stuttgart-Feuerbach with augmented-reality glasses for assembly.
The Port of Hamburg[4] trialled network slicing with Deutsche Telekom and Nokia (the EU-funded 5G-MoNArch project) for traffic control and environmental monitoring. Finally, Lufthansa Technik[5] has run private 5G with Nokia since 2020 for remote high-resolution video engine inspections.
Comparison with Wi-Fi 6/7
| Aspect | Private 5G | Wi-Fi 6/6E/7 |
|---|---|---|
| Range | Km (outdoor) | <100m typical |
| Latency | <20 ms URLLC, <5 ms ideal | 5-30 ms variable |
| Per-cell capacity | Thousands of devices | Hundreds |
| Determinism | High (with slicing) | Limited |
| Handover | Robust between cells | Weak |
| Capex cost | High | Moderate |
| Spectrum | Licensed / shared | Unlicensed ISM |
| Enterprise experience | New | Mature |
Wi-Fi 6/7 remains the default for office and limited-industrial environments. Private 5G wins in broad outdoor coverage, high mobility, and critical determinism. The most common choice in a medium-size plant is Wi-Fi 6 for offices and stable production zones, reserving private 5G for the outdoor logistics area or the longer-range AGVs.
Open RAN: lowers the entry barrier
Open RAN reduces dependence on Nokia/Ericsson:
-
Alternative providers: Parallel Wireless, Mavenir, Altran, dozens more.
-
Interchangeable components: radio from one vendor, core from another.
-
Lower cost with commoditised equipment.
-
Challenge: more complex integration and operation.
This movement fits the broader trend toward as-a-service models in industry, where machinery manufacturers want to control the connectivity of their field assets.
Co-located edge compute
A private 5G network without edge compute under-uses the investment. The typical pattern:
-
5G core + MEC (Multi-access Edge Computing) in the same rack.
-
Kubernetes at the edge for industrial workloads.
-
AI/ML inference close to the sensor for minimal latency.
-
Integration with central IT via MPLS, SD-WAN.
References: Azure Private MEC, AWS Outposts + Private 5G, Red Hat OpenShift at the edge.
When it doesn’t make sense
Honestly:
-
Plant with fewer than 50 connected devices: Wi-Fi 6 is more efficient.
-
No real URLLC cases: if you don’t need <20 ms latency, don’t pay the premium.
-
No significant mobility: cable + Wi-Fi covers.
-
No budget for sustained ops: 5G requires continuous expertise.
A well-executed project takes 12 to 18 months from business case to production, with a limited 10-20 device PoC before scaling.
Conclusion
Private 5G networks are real and have cases where ROI justifies investment, mainly large campuses with mobility, URLLC cases, or high device density. For most plants, Wi-Fi 6/7 remains the right choice. The decision should be based on concrete needs, not hype. Companies adopting private 5G without clear cases find high capex and low return; those adopting for real needs get significant operational advantages that complement their Industry 4.0 and digital thread strategy.
This article is also available in Spanish.
Sources:
- BMW Group PressClub — Intelligent connected factory with 5G technology: autonomous logistics at BMW Group Plant Landshut[2]
- Bosch Media Service — Bosch applies for local 5G licenses[3]
- Nokia — Nokia deploys 5G private wireless network for Lufthansa Technik virtual inspection trial[5]
- 5G-MoNArch — Smart Sea Port testbed, Port of Hamburg[4]
- CNMC — Statement on non-operator companies deploying their own 5G networks[1]
Frequently asked questions
When is Wi-Fi 6/7 enough, and when do I need a private 5G network in the plant?
Wi-Fi 6/7 remains the right choice for most medium-size plants, especially with fewer than 50 mobile devices, no significant mobility, and no cases requiring latency under 20 ms. Private 5G wins in broad outdoor coverage (kilometre range versus under 100 m), high mobility with robust handover between cells, thousands of devices per cell, and URLLC determinism. The most common choice in a medium plant is Wi-Fi 6 for offices and stable production zones, with private 5G only for outdoor logistics or longer-range AGVs.
How much does a private 5G deployment cost and how long does it take for a medium plant?
Typical startup cost is €500k-2M, and a well-executed project runs 12-18 months from business case to production, with a limited 10-20 device PoC before scaling. On top of capex comes sustained operation: 5G requires continuous expertise, and scarce talent is the most underestimated barrier. Without co-located edge compute (5G core plus MEC in the same rack, Kubernetes at the edge) the investment is under-used.
Do I need my own spectrum, or is a slice of the operator's network enough?
For serious industry the preferred option is private licensed spectrum, because public-network slicing keeps you dependent on the operator at the most critical moments. In Germany the Bundesnetzagentur licenses the 3.7-3.8 GHz band individually to each plant. In Spain the mechanism is a private-use concession awarded by auction. The 3800-4200 MHz band has been harmonised at EU level as shared low/medium-power spectrum, applicable from 2026.