Private 5G beside Wi-Fi
Wi-Fi stays the plant floor network; private 5G earns a place beside it as a second radio that carries control traffic and nothing else, and its worst case still has to be measured like any other link.
Wi-Fi carries the plant floor. Private 5G earns its place beside it, not instead of it, when it gets one job: the control traffic.
The question does not change with the radio. What is the longest gap a device will see, and is it shorter than the timeout of the connection riding on it?
Two radios, two jobs
On a shared channel, prioritization only improves the odds. 802.11 cannot assure strict priority between access categories, or a minimum bandwidth for any of them. Separation is the stronger lever, and on a vehicle the cleanest separation is two radios: one for control and safety traffic and nothing else, one for everything else. The airtime guide makes the full case.
Two radios, two jobs
A map download on the second radio cannot queue in front of an I/O packet on the first.
Schematic, not to scale
On one shared channel, 802.11 cannot promise strict priority between access categories or a minimum bandwidth to any of them, so prioritization improves the odds and reserves nothing. Two radios separate the traffic instead. Conceptual: no products, timings, or throughput implied.
Source: RFC 8325.
The control radio can be a second Wi-Fi radio on its own band and channels. That is the first option to try, because your vehicles and your team already run Wi-Fi. Private 5G is the other candidate, and the rest of this page is about when it fits.
Whichever carries control, keep it lightly loaded. On one factory-floor private 5G testbed, mean one-way latency was 3.9 ms uplink and 4.6 ms downlink at low load1. When the same study saturated the uplink, mean uplink latency rose to 211 ms, or 123 ms with priority settings. The radio that carries I/O should carry nothing that could fill it.
The second radio is often a separate box already. In both measured testbeds, industrial equipment reached 5G through external 5G routers2 or plug-in 5G modules on host computers, not radios built into the controllers. On a vehicle that already carries a Wi-Fi client, that is one more radio to power, mount, and manage.
Vehicle makers split the same way. MiR’s network guide lets a robot route through an external router or 5G CPE, a cellular modem box, connected over Ethernet3. Boston Dynamics sells a compute payload for Spot with a built-in 5G/LTE modem and CBRS support for private networks4. Ask which one your vehicle needs before you design the second radio around it.
What a coordinated band buys the control radio
Wi-Fi is unlicensed. In 6 GHz, low-power indoor devices must use a contention-based, listen-before-talk protocol. Anyone can carry a transmitter into your building, which is why NIST tells factories to survey for employee-owned devices and mobile telephone hot spots. You control the air only as far as your policy reaches.
CBRS is coordinated, not exclusive. In the U.S., the Citizens Broadband Radio Service covers 3550 to 3700 MHz in three tiers5. Every base station must register with and be authorized by a Spectrum Access System (SAS) before it transmits, and run only at the power and location the SAS allows. Priority Access Licenses (PALs) are protected from the general tier but must accept interference from incumbents, and they are county-wide, with ten-year renewable terms. General Authorized Access users must accept interference from all other users, including other GAA users.
5G-ACIA, an alliance that promotes 5G in industry, is frank about this kind of sharing: it is primarily intended to protect incumbent users, and IIoT systems may suffer interruptions in such bands.
Other countries license locally. Germany, for example, allocates 3.7 to 3.8 GHz, up to 100 MHz, for private 5G. Rules differ by country, so check your regulator.
5G in unlicensed bands inherits Wi-Fi’s problem. 5G-ACIA notes no guarantee of very low latency in unlicensed spectrum unless only your own equipment uses the carrier.
Coordinated spectrum keeps the stranger’s hotspot off the control radio. It does not remove steel, shadowing, or your own traffic.
The handover is still a gap
A 5G handover is decided by the network from the device’s measurement reports, after which the device detaches and random-accesses the target cell. A conventional Wi-Fi roam is break-before-make: the client leaves its access point, then associates and re-authenticates with the next. Either way, the vehicle is briefly on no network.
The 3GPP evaluation assumption for a baseline or conditional handover is an interruption of around 80 ms6, twice the 40 ms default limit of a CIP Safety input connection7. Release 16 added Dual Active Protocol Stack (DAPS) handover, which is make-before-break and cuts the interruption to about 2 ms, but is not supported in the higher FR2 bands. Release 18 targets interruptions as low as 1 ms. Wi-Fi’s make-before-break roaming sits in the 802.11bn draft, with final approval projected for May 20288.
A feature in a release is not a feature in your module. Ask which mechanism your base stations, access points, and device modems actually implement, then measure the gap.
That gap is where Wi-Fi fleets hurt today. In one car plant, AGVs logged 1,131 disconnections of 3 seconds or more in 3.5 months, and 85.84% began with a roam attempt9. No comparable peer-reviewed field study of a private 5G AGV fleet was found for this page.
A vendor claim, labeled as one. Celona, a private 5G vendor, states that at one site that moved 400 AGVs off Wi-Fi it saw more than 98% successful handovers and handovers of about 110 ms against about 450 ms for Wi-Fi10. The figures come from the vendor’s own deployment summary. Taken at face value, 110 ms is still longer than a 40 ms default safety connection limit.
Specified versus shipping
What 5G specifies is impressive. URLLC targets air-interface latency under 1 ms at reliability above 99.999%11, and with URLLC features enabled and properly configured, 5G-ACIA puts end-to-end latency at 1 to 2 ms at 99.999%. Release 16 also added time-sensitive communication with bounded latency and non-public networks.
Measurements on commercial equipment tell a smaller story. A 2023 factory-floor study on commercially available Release 16 equipment noted that URLLC features were still not commercially available. A 2024 private 5G testbed measured an average round trip of 20.4 ms between 5G routers over a 20-hour test.
At low load, single-digit milliseconds with a tight maximum is a good control radio. But those are not URLLC numbers, they move with load, and an average says nothing about the one gap that trips a connection.
What the second radio costs to run
- A credential for every device. On public networks and public-network-integrated private networks, devices authenticate with a USIM, usually a SIM card or eSIM. Standalone non-public networks may use other credentials. Either way, a credential lifecycle joins the operating model.
- Coordination for every base station. On CBRS, SAS registration and authorization before transmitting, and a PAL if you want priority over the general tier.
- A core on site. Keeping the core’s user plane local adds less latency than the radio, which puts core equipment in your building.
No cost figures appear here, because no source found for this page publishes comparable ones. Price the second radio against the same coverage, the same vehicles, and the same staff.
Field note. Private 5G costs much more than Wi-Fi and takes more management.
When a second radio is worth it
Field note. Two radios is a newer idea, and it is likely to matter more as time goes on. Deciding where control goes means assessing how both networks perform, and many dependencies shape the answer. Where it goes from here depends on cost and on what Wi-Fi delivers next.
Work through these in order:
- Separate within Wi-Fi first. A dedicated band and channel set for control traffic is the cheapest second radio. Move on only if its worst case still does not fit the timeout.
- What your vehicles can carry. A radio your vehicles cannot mount and power is not an option this year.
- Who else transmits in your space, and whether you can enforce a spectrum policy.
- Indoors or outdoors. Low-power indoor 6 GHz Wi-Fi stays indoors, and outdoor 6 GHz needs AFC-controlled standard power in part of the band. CBRS allows outdoor Category B base stations at up to 47 dBm per 10 MHz under SAS control.
- Which make-before-break mechanism your hardware implements, and the gap you measure with it.
- Which operating model your team can run on the night shift.
Neither radio is a safety device
5G-ACIA treats 5G as a black channel, like almost all industrial communication, whose components do not have to implement functional safety standards. Safety protocols poll the black channel and put the machine in a safe state when it is unavailable, so a network with 10% availability would still be safe and almost never running12. The same holds for Wi-Fi. What you buy from either radio is availability.
Give control traffic a radio of its own, keep everything else off it, and measure its worst case in your building.
About this page
Built from 17 sources: 4 standards bodies and labs, 2 regulators and government sources, 2 protocol owners and alliances, 5 research papers and theses and 4 vendor documents. Researched and drafted with AI assistance, then reviewed and approved by Ben Rutter on . How pages are made
- First published
- Last updated
Change history (1)
- First published
Cite this page
Plain
Ben Rutter. "Private 5G beside Wi-Fi." OT Wireless, published October 5, 2026. https://otwireless.com/technologies/private-5g-vs-wifi/
APA 7
Rutter, B. (2026, October 5). Private 5G beside Wi-Fi. OT Wireless. https://otwireless.com/technologies/private-5g-vs-wifi/
BibTeX
@misc{rutter2026private5gvswifi,
author = {Rutter, Ben},
title = {{Private 5G beside Wi-Fi}},
year = {2026},
howpublished = {\url{https://otwireless.com/technologies/private-5g-vs-wifi/}},
organization = {OT Wireless},
}