Wireless on a factory floor is usually procured on throughput, and throughput is almost never the binding constraint. A robot does not need hundreds of megabits per second to follow a route; it needs the same bounded delay on the thousandth packet that it got on the first. That is a tail-latency specification, and it is where unlicensed spectrum and licensed spectrum stop being substitutes.
Wi-Fi 7 brings wider channels, multi-link operation and much better scheduled access than earlier generations. In a static office it is excellent. On a floor with welding arcs, metal racks, rotating machinery and dozens of access points, the contention mechanism still has to listen before it talks, and the amount of listening is set by neighbours you do not control. A deterministic upper bound is hard to promise when the medium is shared.
Industrial wireless is bought on average throughput and judged on the worst millisecond of the month.
Part One: What is actually being compared
The two options are often described as competing implementations of the same thing, which is misleading. private 5G and Wi-Fi 7 make different promises about where the variability sits: one puts it in the capital plan, the other in the operating tail. A comparison that only measures the mean therefore misses the entire reason a team would choose one over the other in industrial wireless.
There is a quiet accounting asymmetry here. private 5G moves spending into capital and spreads it over the life of the installation; Wi-Fi 7 keeps spending in operations where it is easy to approve and hard to cap. Finance teams prefer the second, engineering teams prefer the first, and the argument is usually settled by the budgeting process rather than by measurement in industrial wireless.
Part Two: The arithmetic that decides it
Private 5G buys determinism with a scheduled, licensed or locally licensed carrier and a core that owns admission control. The base station tells each device exactly when it may transmit, so the tail stays bounded even under load. The cost is not just the spectrum and the radios; it is the SIM or eSIM lifecycle, the core, and an operations team that now has to run a small mobile network.
Deployment also changes the measurement. After private 5G goes in, the metric people optimise shifts from throughput to predictability, and after Wi-Fi 7 it shifts from predictability to unit price. That shift is why two teams can look at the same production data and reach opposite conclusions about which option won.
Part Three: Where the simple model breaks
Handover is the specific tax. A device crossing between Wi-Fi access points initiates its own scan and reassociation, and the resulting gap is variable. A cellular handover is orchestrated by the network with make-before-break behaviour, so the gap is engineered rather than discovered. For a person streaming video both are invisible; for an autonomous cart whose controller treats a 120 millisecond gap as a lost heartbeat, one triggers a safety stop and the other does not.
A 6 ms median is irrelevant when a 120 ms reassociation gap at the 99.9th percentile triggers a safety stop on every aisle crossing.
Part Four: The cost that arrives later
There is a deployment asymmetry worth naming. Wi-Fi access points are cheap and can be added by whoever owns the building; cellular coverage is planned once and changing it is a project. That makes Wi-Fi better for a floor that gets reconfigured every quarter, and private 5G better for a fixed process line that runs for a decade. The flexibility difference is often more decisive than the radio specification.
Adoption will not be a single flip. Early movers tend to be the cases with the most extreme value of the deciding variable, then the middle of the market follows once tooling and reference designs exist, and the long tail stays with Wi-Fi 7 until the cost of switching falls below the cost of an exception. A reasonable reading of the next few years is coexistence with a slowly moving boundary rather than a replacement event.
Part Five: What this changes outside the engineering team
Most real sites end up with both: cellular for the moving, safety-relevant control plane and Wi-Fi for bulk data, cameras and human devices. The interesting engineering is the policy layer that decides which traffic goes where, and it is a software problem rather than a spectrum problem.
The everyday analogue is a video call that freezes as you walk from the kitchen to the balcony. You tolerate it because nothing depends on those missing frames. A warehouse robot cannot, and the difference between a nuisance and a stopped production line is a tail percentile, not an average speed.
A market sizing sanity check
Industrial wireless infrastructure is a 12 to 18 billion dollar annual market by the end of the decade, with private cellular perhaps a fifth of new large-site spend. Add the policy, SIM management and edge-core software that a dual-network site needs and the addressable layer is 2 to 4 billion dollars a year by 2030, mostly recurring rather than one-off hardware.
All figures here are an author's synthesis of published ranges and a stated model, not a measurement of a shipped system. The point of the arithmetic is to locate the crossover and expose which assumption drives it; a vendor quote will differ, and the direction of that difference is more informative than the final number.
The comparison that matters
Wi-Fi 7 optimises for cheap bandwidth, easy reconfiguration and a consumer-grade supply chain. Private 5G optimises for bounded tail latency, orchestrated handover and admission control. Decide on what a 99.9th-percentile gap costs your process, not on a speed test at the network centre.
A change in regulation or in the supply of a single component could invalidate the ranking overnight. The ranking is a snapshot of prices and constraints in industrial wireless, not a property of the technologies, and anyone reusing it in a different year should re-run the threshold rather than reuse the conclusion.
Licensed spectrum does not buy you more speed; it buys you the right to decide who transmits and when.
SOL: HCqPmPJM3NVtS6YrmVaTZYnuP8XZ2PFgQZKT94VK1LkS
Creating this takes real effort — thank you for the support.
