Private 5G gets pitched as a Wi-Fi replacement, and sometimes it is one — but the spectrum, core network, and device decisions underneath that pitch are genuinely different from anything a plant IT team has had to evaluate before. Here’s what private 5G actually is, the spectrum and deployment models available, and where endpoint hardware — including 5G RedCap — fits into the economics.
Written by E-Lins Engineering Team
What Private 5G Actually Is, in Plain Terms
A private 5G network for industrial manufacturing is a cellular network — radio access, core network, and device management — built to serve a single organization’s site rather than the general public. Instead of your equipment connecting to a public carrier’s macro cellular network shared with everyone else in the area, it connects to infrastructure your organization (or a partner operating on your behalf, such as an industrial IoT router manufacturer) controls, running on spectrum specifically allocated for that site, covering exactly the campus, plant floor, or facility footprint you need and nowhere else.
Structurally, a private 5G network has the same three core components as any cellular network: a Radio Access Network (RAN) — the small cells or indoor radio units that actually transmit and receive — a core network that handles authentication, mobility management, and routes traffic to your local network or the internet, and spectrum, the specific radio frequency band the whole system operates on. What makes it “private” is that all three of those are scoped and controlled for your site specifically, rather than being a slice of a public carrier’s shared infrastructure.
What private 5G is not, to be clear before going further: it is not simply “5G Wi-Fi,” it is not automatically cheaper than Wi-Fi for every use case, and it is not a single off-the-shelf product you buy from one vendor. Building a private 5G network involves spectrum acquisition or licensing, RAN and core network infrastructure (typically from vendors like Ericsson, Nokia, Celona, Athonet, or similar specialists — a different category of vendor from the routers and CPE covered elsewhere on this site), and endpoint devices. This article focuses on helping you understand the decision landscape and where endpoint hardware fits — not on selling you a full turnkey network, which E-Lins does not provide.
Why Industrial Manufacturers Are Evaluating Private 5G Specifically
Wi-Fi has served manufacturing floors for two decades, and for a large share of applications it remains entirely adequate. The specific gaps that push manufacturers toward private 5G tend to cluster around a handful of recurring pain points that Wi-Fi’s architecture makes structurally difficult to solve, rather than 5G being inherently superior in every dimension.
- Seamless mobility across a large campus. Wi-Fi handoff between access points can introduce brief connectivity gaps as a device roams; cellular handoff — including on a private 5G network — is designed around continuous mobility from the ground up, which matters directly for AGVs and AMRs moving continuously across a large plant floor.
- RF penetration through dense industrial environments. Steel structures, large machinery, and metal shelving degrade Wi-Fi’s higher-frequency signals more than they degrade the lower and mid-band frequencies commonly used for private cellular, which can mean fewer access points for equivalent coverage in a metal-dense facility.
- Deterministic QoS through network slicing. A private 5G core can guarantee latency and bandwidth characteristics to specific device classes — safety-critical control traffic gets a different service guarantee than a bulk file transfer — in a way that’s architecturally native to 5G rather than bolted on.
- Device density at scale. Facilities deploying hundreds to thousands of sensors, tags, and controllers can hit practical density limits on Wi-Fi access points faster than on a cellular architecture designed for high device counts per cell.
- Data sovereignty and network isolation. Traffic on a private 5G network with a locally hosted core can stay entirely on-premises, relevant for manufacturers with strict data residency or OT network isolation requirements.
None of these are universal requirements — a facility with a handful of fixed sensors and no mobile robotics may never need any of them, and Wi-Fi 6/6E remains a perfectly sound choice for a large share of industrial connectivity needs. The point of evaluating private 5G is matching it to a genuine requirement from this list, not adopting it because it’s the newer technology category.
Spectrum Options: CBRS, Shared, and Licensed — and Why It’s Jurisdiction-Specific
Private 5G spectrum options CBRS shared licensed vary meaningfully by country, and this is the first place manufacturers scoping a project outside a single home region run into complexity — a spectrum framework available in one country may not exist in the same form elsewhere. The three broad categories below cover the models most commonly referenced, but always confirm the specific framework, application process, and available bands with your local telecommunications regulator before planning a deployment.
CBRS (Citizens Broadband Radio Service) — United States
CBRS operates in the 3.55–3.7 GHz band in the United States under a three-tier shared access framework managed by a Spectrum Access System (SAS). The tier most relevant to industrial private networks is General Authorized Access (GAA) — usable without an auction or exclusive license, on a shared, coordinated basis — alongside Priority Access License (PAL) spectrum, which is auctioned in specific census tracts and provides interference protection from GAA users. CBRS has become a common starting point for U.S. private 5G projects specifically because GAA access lowers the barrier to entry considerably compared to acquiring exclusively licensed spectrum.
Shared and Locally Licensed Spectrum — Other Regions
Several other countries have established their own frameworks for local, site-specific spectrum licensing, generally allocated directly to the end-user organization (the manufacturer) rather than to a mobile carrier. Germany’s Bundesnetzagentur, for example, offers local licensing in the 3.7–3.8 GHz range specifically for campus networks. The UK’s Ofcom offers shared access spectrum across multiple bands for local network use. Japan has a local 5G licensing framework covering specific frequency ranges for site-specific deployments. Each of these programs has its own application process, eligibility requirements, and available bandwidth — treat any specific frequency or process detail as a starting point for your own regulator’s current published guidance, not a substitute for it, since spectrum policy changes over time.
Licensed Spectrum via a Carrier Partnership
Where local or shared spectrum licensing isn’t available, or where a manufacturer prefers not to manage a spectrum license directly, partnering with an existing mobile network operator to use a portion of their licensed spectrum — either through a dedicated on-site deployment or a network slice on the carrier’s existing infrastructure — is the remaining path. This generally trades some independence and control for a lower barrier to entry, since the carrier already holds the spectrum rights and, in many cases, existing macro network infrastructure near the site.
Confirm spectrum availability before scoping anything else. Which spectrum options exist, and the specific process to access them, is entirely dependent on your country and, in some cases, your specific site location. This step alone can take anywhere from weeks to many months depending on the framework and regulator, and it should be the first item resolved in any private 5G project — not something confirmed after RAN and core vendor selection is already underway.
Planning Checklist: Questions to Resolve Before Scoping a Private 5G Project
- What spectrum options exist in your specific jurisdiction and site location, and what is the realistic timeline to secure access to them? Confirm with your national telecommunications regulator before any other planning proceeds.
- What is the actual device population — fixed machines, mobile robotics, sensors, cameras — and what does each category need in terms of throughput, latency, and mobility? This drives both the core network’s slicing design and the endpoint hardware specification covered later in this guide.
- Do you want to own and operate the network directly, or engage a neutral host / managed service provider? This is as much an organizational and staffing decision as a technical one — running a private cellular core requires network engineering expertise most manufacturing IT teams don’t currently have in-house.
- What is your facility’s RF environment — steel structure, machinery density, existing Wi-Fi deployment — and has a proper RF survey been conducted for private 5G specifically, not just extrapolated from Wi-Fi coverage data?
- Does the project need full 5G throughput at every endpoint, or does a large share of the device population fit a 5G RedCap profile? This has a direct, often underappreciated impact on total endpoint cost at scale, covered in detail later in this guide.
- How will the private network integrate with existing OT and IT network segmentation, security policy, and your existing SCADA/MES systems? A private 5G network doesn’t replace network security architecture — it’s a new access layer that has to fit into it.
Core Network Deployment Models
The core network — the part of the system handling authentication, mobility, session management, and traffic routing — can be deployed in several architecturally different ways, each with different cost, control, and complexity tradeoffs.
Fully On-Premises Standalone Core
Both the RAN and the core network run entirely on-site, typically on dedicated servers in the facility’s own data center or a local edge compute rack. This offers the highest degree of control, the lowest inherent latency for on-site traffic (since nothing needs to route off-site to reach the core), and the strongest data sovereignty posture, at the cost of the highest upfront infrastructure investment and the need for in-house or contracted network operations expertise.
On-Site RAN With Cloud-Hosted Core
The radio hardware remains on-premises, but core network functions run as software in a cloud or regional edge data center rather than on local hardware. This reduces on-site infrastructure and can lower upfront cost and operational complexity, at the cost of some dependency on connectivity back to the hosted core and a latency profile that, while still generally very good, isn’t quite as tight as a fully local deployment for the most latency-sensitive applications.
Neutral Host / Managed Service Provider Model
A third-party specialist builds, owns, and operates the network infrastructure — RAN, core, and often spectrum access — and the manufacturer subscribes to network capacity much like a utility service, rather than owning the infrastructure directly. This substantially lowers the upfront capital investment and the need for in-house cellular network expertise, at the cost of less direct control over network configuration and a recurring service cost rather than a capital asset.
Carrier Network Slice
Rather than genuinely separate infrastructure, this model uses the public carrier’s existing macro network with a dedicated, QoS-guaranteed virtual slice allocated to the manufacturer’s devices. It’s the lowest-complexity option to get running, since it relies on infrastructure the carrier already operates, but it depends on that carrier’s macro network actually providing adequate coverage and capacity at the specific site — which is not guaranteed for every industrial location, particularly larger or more RF-shielded facilities.
“The model decision is usually less about the technology and more about which organization is going to operate this thing day to day. A plant with a strong internal network engineering team and a genuine need for full control gravitates toward an on-premises core. A plant that wants private 5G’s benefits without building an internal cellular network operations capability gravitates toward neutral host. Neither is the ‘right’ answer independent of that organizational reality.”— E-Lins Engineering Team, on private network scoping conversations
Five Device Requirements That Determine Endpoint Hardware Fit
1. SIM/eSIM Provisioning and Onboarding at Scale
A private network with hundreds or thousands of endpoints needs a provisioning workflow that doesn’t require manually configuring each device’s network credentials individually. eSIM and remote SIM provisioning, along with our previously covered automatic carrier/APN identification capability, become directly relevant at private-network scale, since they remove per-device manual provisioning steps that don’t scale to a large campus deployment.
2. Band-Matched Hardware for the Specific Spectrum in Use
This is one of the most commonly underestimated requirements in early planning: endpoint hardware must support the specific band your private network operates on — CBRS’s n48 band, for example, requires modem hardware explicitly supporting it, which not every generic 5G router does by default. Confirm band support explicitly against your specific spectrum allocation before procuring endpoint hardware at any scale, rather than assuming any 5G-capable router will work on any private network.
3. Fixed vs. Mobile Endpoint Architecture
Fixed machine-mounted endpoints — a CNC controller, a quality inspection camera, a fixed sensor array — have very different hardware requirements from mobile endpoints like AGVs and AMRs. Fixed installations benefit from RS485/Modbus serial integration for bridging legacy OT equipment (covered in depth in our Modbus gateway configuration guide) and DI/DO for machine status and alarm signaling. Mobile equipment needs vibration tolerance, wide-voltage DC input matched to the vehicle’s electrical system, and often GPS/positioning integration for indoor or campus-wide location tracking.
4. Network Slicing and QoS Differentiation Support
The endpoint device doesn’t create network slicing — that’s a core network function — but it needs to correctly participate in whatever QoS marking and slice assignment the core network applies, so that safety-critical control traffic and best-effort sensor telemetry are correctly differentiated end to end rather than the distinction being lost at the device.
5. Edge Compute Proximity for Latency-Sensitive Applications
For applications where local processing needs to happen close to the data source — real-time machine vision inspection, for example — the endpoint device’s role often extends to local data aggregation and pre-processing before transmission, which affects the compute and interface requirements of the gateway hardware at that specific point in the network, not just its cellular radio specification.
How 5G RedCap Fits Into Private Network Endpoint Economics
This section builds directly on our dedicated 5G RedCap vs full 5G comparison guide — if you haven’t read that article, the short version is that RedCap (NR-Light, 3GPP Release 17) is a reduced-complexity 5G specification designed for IoT device categories that need 5G-core integration but not full 5G’s multi-Gbps throughput. Private network deployments are precisely where this distinction has the clearest economic impact, because a typical industrial campus network’s device population is heavily weighted toward exactly the category RedCap was designed for.
Consider a representative private 5G deployment: a handful of high-throughput fixed endpoints (machine vision cameras, video-based quality inspection, high-frequency data acquisition) alongside a much larger population of lower-bandwidth devices (environmental sensors, asset tags, simple status monitors, predictive maintenance vibration sensors). Specifying full 5G hardware across that entire device population — because the network itself is “5G” — needlessly inflates the module cost and power draw of the majority category, where the additional throughput headroom goes unused regardless. 5G RedCap endpoint cost private network deployment economics improve specifically because RedCap’s reduced antenna count and narrower bandwidth requirement lower both the module cost and the power draw for that device category, while still delivering the 5G-core network slicing and QoS participation the private network’s architecture depends on.
| Endpoint Mix — Representative Private 5G Manufacturing Deployment |
| Full 5G tier (small share of device count): machine vision/quality inspection cameras, high-frequency data acquisition nodes, mobile AGV/AMR controllers needing sustained throughput headroom. |
| 5G RedCap tier (majority of device count): environmental and process sensors, asset tags, predictive maintenance vibration/temperature monitors, simple status and alarm reporting endpoints. |
| 4G LTE tier (where applicable): legacy equipment retrofit via Modbus/RS485 gateway where 5G-core participation isn’t required for that specific device — see our dedicated 4G vs 5G decision framework for this comparison in more depth. |
| Matching each device to the tier its actual requirements justify — rather than defaulting the whole deployment to full 5G because the network is “5G” — is typically where the largest avoidable cost sits in a private network endpoint budget. |
Where E-Lins Endpoint Hardware Fits: CPE and Gateway Devices, Not RAN or Core
To be direct about scope: E-Lins manufactures the endpoint routers and gateways that connect machines, sensors, and mobile equipment to a private 5G network — we are not a RAN or core network infrastructure vendor, and this guide isn’t intended to position us as one. For the RAN and core network itself, you’ll be working with a specialist private network vendor or a neutral host provider. Where our hardware fits is the device layer described in the criteria section above.
H900f / H900pf — Full 5G Fixed Endpoint / High-Throughput Gateway
For fixed high-throughput applications — machine vision aggregation, video-based inspection, a multi-sensor gateway needing sustained bandwidth — the H900f and H900pf series provide full 5G SA/NSA connectivity with the Ethernet and Wi-Fi capacity to serve as an on-machine or cell-level aggregation point.
H900frc — 5G RedCap Endpoint for the Sensor-Class Device Population
For the larger, lower-bandwidth device category described in the RedCap economics section above, the H900frc provides 5G-core participation at meaningfully lower per-device cost and power draw — directly relevant at the device counts typical of a large manufacturing campus deployment.
H685f — Compact Endpoint for Mobile Equipment and OEM Integration
For AGV, AMR, and other mobile equipment needing to be embedded directly into the vehicle’s electronics, the H685f‘s compact form factor, wide-voltage DC input, and full 5G capability suit mobile private-network endpoints specifically.
H720 / H750 — OT Integration Gateways for Legacy Equipment
Where existing machines and PLCs need to join the private network via Modbus RTU/RS485 bridging rather than native 5G integration, the H720 and H750 series provide the serial and DI/DO integration covered in our dedicated Modbus gateway configuration guide, bridging legacy OT equipment onto the private network’s IP infrastructure.
Confirm band compatibility before ordering. As covered in the device requirements section above, endpoint hardware must support the specific spectrum band your private network operates on. Confirm current band support for your specific spectrum allocation directly with E-Lins before procuring hardware at scale for a CBRS, locally licensed, or carrier-slice private network deployment.
Three Deployment Patterns Common on Private 5G Manufacturing Projects
| Case Pattern 1 — Automotive Assembly Plant, CBRS, AGV Fleet |
| A U.S. automotive assembly facility deployed a private 5G network on CBRS spectrum specifically to support a growing fleet of AGVs moving parts across a large plant floor, after Wi-Fi handoff gaps between access points had caused intermittent control disruptions as vehicles roamed. The private network’s continuous cellular mobility architecture eliminated the handoff-related disruptions, with H685f units embedded in each AGV providing the vehicle-side connectivity, wide-voltage DC input matched to the vehicle’s onboard power system, and GPS for indoor-adjacent position tracking feeding the fleet management system. |
| Case Pattern 2 — European Electronics Manufacturer, Local Licensed Spectrum, Machine Vision |
| A manufacturer in a country with a local 5G licensing framework deployed a private network specifically to support a bank of machine vision quality-inspection cameras that needed sustained throughput and low, consistent latency to a local edge compute rack running defect-detection processing — a workload that had outgrown the facility’s existing Wi-Fi infrastructure during peak production runs. H900f units at each inspection station provided the sustained throughput needed for the camera feeds, with the private network’s local core keeping all inspection data on-premises to meet the manufacturer’s data handling policy. |
| Case Pattern 3 — Large Industrial Site, Predictive Maintenance Sensor Network on RedCap |
| A large industrial site building out a predictive maintenance program deployed several hundred vibration and temperature sensors across its production equipment on a private 5G network, with an initial plan specifying full 5G hardware for the entire sensor population. Reviewing actual per-sensor data rates — all comfortably within RedCap’s throughput ceiling — against the deployment’s device count led to specifying H900frc RedCap units across the sensor fleet instead, reducing the hardware and power budget for that device category meaningfully while retaining the network slicing and QoS participation the predictive maintenance platform’s architecture required, consistent with the economics discussed earlier in this guide. |
Common Mistakes in Private 5G Manufacturing Projects
Treating Spectrum Access as a Formality Instead of the First Gating Step
Spectrum availability and the process to secure it varies substantially by country and can take considerably longer than the RAN and core procurement process itself. Confirm this first, not last — projects that scope hardware and vendors before confirming spectrum access risk a significant timeline surprise.
Specifying Full 5G Across the Entire Device Population by Default
As covered in the RedCap economics section, a typical manufacturing device population is weighted toward lower-bandwidth sensor-class endpoints that don’t need full 5G’s throughput. Defaulting the entire deployment to full 5G hardware because “the network is 5G” is one of the most common and most avoidable sources of inflated endpoint budget on these projects.
Assuming Any 5G Router Works on Any Private Network’s Spectrum
Band compatibility between endpoint hardware and your specific spectrum allocation — CBRS, a local license, or a carrier slice — is not automatic. Confirm explicitly before procuring hardware at scale.
Underestimating the Organizational Commitment of an On-Premises Core
A fully on-premises private core network requires ongoing network operations expertise that most manufacturing IT teams haven’t previously needed to build in-house. This isn’t a reason to avoid that model, but it should be weighed honestly against the neutral host alternative before committing, rather than discovered as an operational gap after deployment.
Not Involving OT Security and Network Segmentation Planning From the Start
A private 5G network is a new access layer onto your industrial network, not a replacement for existing security architecture. Integrating it into existing OT/IT segmentation policy from the beginning avoids retrofitting security controls onto a network that’s already in production.
Extended Reading
E-Lins H900frc 5G RedCap Router — Full comparison of RedCap vs full 5G throughput, power, and cost tradeoffs referenced throughout this guide.
E-Lins H900f Industrial 4G/5G Router — Full 5G fixed endpoint and aggregation gateway for high-throughput private network applications.
E-Lins H685f Compact 5G Router — Mobile equipment and OEM-embeddable endpoint for AGV/AMR private network connectivity.
E-Lins H720 Dual SIM 4G Router — Modbus/RS485 OT integration gateway for bridging legacy plant equipment onto a private network.
E-Lins Private Network Endpoint Enquiry — Share your private network’s spectrum band, core model, and device population mix for a direct endpoint hardware recommendation.
Frequently Asked Questions
Q1：Does E-Lins build or sell private 5G network infrastructure — the RAN and core?
No. E-Lins manufactures the endpoint routers and gateways — CPE devices — that connect machines, sensors, and mobile equipment to a private 5G network. The radio access network and core network infrastructure itself comes from specialist private network vendors or neutral host providers, a different category of vendor from industrial router manufacturers. This guide is intended to help you understand the overall private 5G landscape and where endpoint hardware fits into it, not to position E-Lins as a full-network provider.
Q2：Is CBRS available outside the United States?
No — CBRS specifically is a U.S. regulatory framework administered under FCC rules with a Spectrum Access System unique to that country. Other countries have their own local or shared spectrum licensing frameworks for private networks — Germany, the UK, and Japan each have established programs, for example — but the specific bands, application processes, and eligibility requirements differ by country. Confirm the framework applicable to your specific jurisdiction with your national telecommunications regulator.
Q3：Do I need to build and operate the private network myself, or can I use a managed service?
Both models exist and are common. A fully on-premises, self-operated core network gives the most control but requires in-house or contracted cellular network operations expertise most manufacturing IT teams don’t currently have. A neutral host or managed service provider model lowers the upfront investment and operational burden by having a specialist build and run the network on your behalf, typically billed as an ongoing service rather than a capital asset. The right choice depends heavily on your organization’s appetite and capacity for operating cellular network infrastructure directly, covered in more detail in the core network models section of this guide.
Q4：Can I mix Wi-Fi and private 5G in the same facility?
Yes, and this is common in practice — many facilities use private 5G specifically for the applications that benefit most from it (mobile robotics, wide-area coverage, deterministic QoS for critical traffic) while retaining Wi-Fi for applications where it remains entirely adequate, such as office areas or fixed workstations with no mobility requirement. The two aren’t mutually exclusive, and a hybrid approach is often the most cost-effective path rather than a full-facility replacement of one with the other.
Q5：How does 5G RedCap fit into a private network if the network itself is described as “5G”?
A private 5G network’s core infrastructure operates on the 5G standard, but individual endpoint devices connecting to it can use different device categories within that standard — full 5G NR for high-throughput endpoints, and 5G RedCap for lower-bandwidth IoT endpoints that still need 5G-core participation (network slicing, QoS) without the cost and power overhead of a full 5G modem. Both device categories connect to the same underlying private network; RedCap is a lighter-weight device specification within the 5G standard, not a separate or lesser network.
Q6：What’s the biggest planning mistake manufacturers make when starting a private 5G project?
Treating spectrum access as a formality to sort out after selecting a RAN/core vendor, rather than the first item to confirm. Spectrum availability and the process to secure it — whether CBRS, a local license, or a carrier partnership — varies by jurisdiction and can take considerably longer than other parts of project planning. Confirming this first avoids a scenario where hardware and vendor decisions are already underway before discovering a spectrum access timeline that reshapes the entire project schedule.
Conclusion: Private 5G Is a Set of Decisions, Not a Single Purchase
A private 5G network for industrial manufacturing isn’t a single product decision — it’s a stack of choices, starting with spectrum access specific to your jurisdiction, moving through a core network deployment model that matches your organization’s appetite for operating cellular infrastructure directly, and ending with endpoint hardware matched to what each device category in your facility actually needs, rather than defaulting the whole deployment to full 5G because the network carries that label.
Three things to resolve before committing to a private 5G project:
- Confirm spectrum availability and the access process in your specific jurisdiction first — before RAN, core, or endpoint hardware decisions, since this step’s timeline can reshape the entire project schedule.
- Choose a core network deployment model that matches your organization’s actual appetite and capacity for operating cellular network infrastructure, not just the model with the lowest theoretical cost.
- Map your device population honestly against full 5G, RedCap, and 4G LTE tiers before specifying endpoint hardware — the largest avoidable cost in most private network budgets sits in over-specifying the sensor-class device population.
Planning Endpoint Hardware for a Private 5G Deployment?
Tell E-Lins your private network’s spectrum band, core deployment model, and device population — fixed, mobile, and sensor-class. Our applications engineering team will confirm band compatibility and recommend the right endpoint hardware mix for your project.
