Private 5G or Wi-Fi on the Plant Floor: How the Choice Is Actually Made
The comparison is usually argued on throughput, which is the one specification that rarely decides it. What separates the two in practice is who owns the spectrum, how the network behaves at the edge of coverage, and what happens when a device roams.
Ask which wireless technology belongs on a factory floor and you will get a throughput argument. Wi-Fi 6E and Wi-Fi 7 quote enormous peak rates; 5G quotes enormous peak rates; both are correct and neither number has much to do with why one succeeds and the other disappoints in a given building. The differences that decide the outcome are less glamorous.
The first is spectrum, and it is the largest. Wi-Fi runs in unlicensed bands shared with every other device in range, including the ones your neighbours install without telling you. Interference is not a fault condition, it is the operating environment, and the protocol is designed around contention: devices listen, back off, and retry. That works well and degrades gracefully, but the degradation is not something you control. Private 5G runs in spectrum that is either licensed to you, licensed to an operator who has partitioned a slice for you, or shared under a coordination scheme. Where you hold the spectrum, nobody else's equipment can take capacity from you by being switched on. That is the substantive difference, and everything else follows from it.
The second is behaviour at the edge of coverage. Wi-Fi is designed for good coverage in a bounded area, and it thins out at the boundary — throughput falls off, retries climb, and a client at the margin can hold a cell's airtime while it struggles. Cellular scheduling is centralised: the base station allocates resources rather than letting devices contend for them, and it can trade throughput for reliability for a specific device. In a warehouse with long aisles of racked metal, or a yard where assets move between buildings, that difference shows up as the number of places where a link is technically present and practically unusable.
The third is roaming. A device moving across a Wi-Fi deployment associates with one access point at a time and re-associates when the signal justifies it. Fast-roaming extensions have improved this considerably, but a handover is still an event that a poorly-tuned deployment can turn into a gap of hundreds of milliseconds. Cellular handover was designed from the beginning for a device travelling at speed between cells, and it is measured in tens of milliseconds. For a fixed sensor this is irrelevant. For an autonomous vehicle crossing a site, or a crane, or a handheld terminal carried at walking pace across a large building, it is often the whole argument.
Against all that, Wi-Fi wins on cost, familiarity and device support, and those are not small advantages. Access points are inexpensive, your IT team already runs them, and every laptop, tablet, camera and barcode scanner speaks Wi-Fi natively. A private 5G deployment needs radios, a core, SIM management and a skill set most plants do not have in house, and every device needs a cellular module or a bridge to one. The 2026 market is around USD 2.8 billion growing at roughly a third a year — real, but small next to the installed base of enterprise Wi-Fi, and that gap shows up as the price of everything.
Two developments are narrowing the gap. The first is regulatory: more countries are opening mid-band spectrum for local licensing, which is what makes a private network something a factory can hold rather than something it has to rent. The second is silicon. 5G RedCap — the reduced-capability profile standardised in Release 17 — strips out the parts of 5G that an industrial sensor never needed, producing modules with a fraction of the complexity, cost and power draw of a full 5G modem while keeping the network behaviour. Release 17 RedCap modules are now shipping commercially, and analysts count dozens of operators investing in RedCap networks. That combination is what turns private 5G from a flagship project into an option for ordinary devices.
So the practical decision procedure. Start with the devices, not the network: if the things you need to connect are laptops, tablets, cameras and scanners, Wi-Fi is the default and the burden of proof is on 5G. If they are AGVs, cranes, mobile robots or anything that moves fast enough to hand over, look hard at cellular. Then look at the environment: a clean office-like space favours Wi-Fi; dense metal, long distances, outdoor yards and neighbours running their own unlicensed equipment favour licensed spectrum. Then check what spectrum you can actually get in your country, because that is a fact about your regulator rather than about your requirements. Finally, count the devices that would need a new module, and price that honestly.
Most large sites end up running both, and that is a sensible answer rather than a failure to decide. Wi-Fi carries the general-purpose traffic it is good at, and a private cellular network carries the specific workloads that justified it. What does not work is choosing on peak throughput, discovering the constraint was coverage or handover, and paying for a second network to fix the first.