Wi-Fi 6E and Wi-Fi 7 are frequently sold as a performance upgrade, and they can be. But the interesting question is almost never "is the new standard faster" — it is "what is the actual bottleneck on this site, and does the new standard address it?" On a large proportion of underperforming networks it does not, because the bottleneck is coverage design, configuration, or the age of the devices in people's hands.
What follows is what the newer standards genuinely change, what they cost, and how to decide whether a refresh is justified.
What 6 GHz actually gives you
Wi-Fi 6E is Wi-Fi 6 — the same 802.11ax radio technology — extended to operate in the 6 GHz band. Wi-Fi 7 (802.11be) is a further generation that can operate across 2.4, 5 and 6 GHz. The headline benefit of both, in practice, is the spectrum.
UK 6 GHz availability has moved in stages, and it is worth being current about it. Ofcom opened the lower 6 GHz band, 5925 to 6425 MHz, for licence-exempt wireless local area network use some years ago — 500 MHz of spectrum, for low-power indoor operation and for very low power devices, with very low power outdoor use subsequently permitted. More recently Ofcom set out a sharing arrangement for the upper part of the band: a further 160 MHz, 6425 to 6585 MHz, on a Wi-Fi priority basis alongside the existing licence-exempt terms, with the remainder up to 7125 MHz prioritised for mobile and available to Wi-Fi only under automated frequency coordination — a database system that checks a device's location against incumbent users before granting higher-power or outdoor operation.
Set against 5 GHz, where usable non-overlapping wide channels are scarce and part of the band is subject to dynamic frequency selection — meaning a radar detection can move your access points off a channel without notice — a clean 500 MHz-plus of contiguous spectrum with no legacy occupants is a genuine step change. It supports several non-overlapping 160 MHz channels where 5 GHz struggles to offer one or two, and it makes Wi-Fi 7's 320 MHz channel width physically possible at all — 320 MHz does not fit anywhere else.
Two further properties of 6 GHz matter more than they get credit for. There are no legacy 802.11b, a, g or n devices in the band, so none of the protection mechanisms and backward-compatibility overhead that burden 2.4 GHz apply. And security is not optional: WPA3 is required, with protected management frames, and open or WPA2-only association is not permitted. That is a real improvement, and also a real constraint — a device that only speaks WPA2 cannot use 6 GHz at all.
What 6 GHz costs you
The spectrum is not free. It is paid for in range, and in three separate ways that compound.
- Path loss rises with frequency. The free-space component is the smallest of the three effects — moving from the upper 5 GHz band to the middle of 6 GHz costs on the order of a decibel or two at the same distance — but it is in the same direction as everything else.
- Materials attenuate more. This is the significant one. Attenuation through walls, floors, glazing, racking and stock increases with frequency, and diffraction around obstacles becomes less effective as the wavelength shortens. A partition that costs you a few decibels at 5 GHz costs you more at 6, and a building that was marginal at 5 GHz will not be better at 6.
- Regulatory power limits are lower. Low-power indoor operation in 6 GHz is constrained in both total radiated power and power spectral density, and in many configurations an access point cannot transmit at the levels it may lawfully use in parts of the 5 GHz band. Higher power indoors and outdoor operation depend on coordination arrangements rather than being available by default.
There is a fourth effect that is not about the band at all but is always present when people move to it, because the whole point of 6 GHz is wide channels:
And upstream, the client sets the limit. Access points can be generous with power and antennas. A phone or a handheld terminal cannot. A design built around what the access point can reach, rather than what the client can return, fails in the same way at 6 GHz as it does at 5 — only sooner.
Why a 6E design usually needs more access points
Put those effects together and the conclusion is unavoidable, and it is the opposite of what most people expect from a newer standard: a 6 GHz design typically requires more access points to cover the same building than the 5 GHz design it replaces, not fewer.
Higher attenuation, lower permitted power and the signal-to-noise penalty of wide channels all shrink the usable cell. If the coverage target is "6 GHz at a level that supports the wide channels we bought it for", the cells get smaller and the count goes up.
There is a discovery consideration too. Passive scanning across a large number of 20 MHz channels would be slow, so 6 GHz uses a set of preferred scanning channels spaced across the band that clients check first, along with reduced neighbour report elements carried in 2.4 and 5 GHz beacons that advertise the presence of a co-located 6 GHz network. In practice this means the 6 GHz network is usually discovered via the lower bands rather than found on its own, which shapes how SSIDs and bands are configured.
None of this is an argument against 6 GHz. It is an argument against treating it as a like-for-like swap. Replacing existing access points one for one with 6E units and expecting the 6 GHz radio to cover what the 5 GHz radio covered is the most common and most expensive misunderstanding in this area — and it produces a network that is measurably worse in the places that mattered, on brand new equipment.
What Wi-Fi 7 adds, and what actually helps
Wi-Fi 7 brings several features. They are not equally useful in a commercial environment.
| Feature | What it does | Where it helps in practice |
|---|---|---|
| 320 MHz channels | Doubles the maximum channel width, in 6 GHz only. | Impressive on a specification sheet and useful for a small number of point-to-point or very high throughput applications. In a multi-access-point deployment, wide channels reduce the number of non-overlapping channels available and increase co-channel contention. With the spectrum currently available in the UK, one 320 MHz channel fits in the lower band; a channel plan built on 320 MHz across a whole building is generally a poor trade. |
| Multi-link operation (MLO) | A client and access point maintain links on more than one band simultaneously, aggregating them or using one as a backup. | The genuinely valuable enterprise feature. It reduces latency and jitter and improves reliability, because a link that hits interference on one band can carry on over another. This matters far more to voice, video and real-time control traffic than raw peak throughput does — and it is the feature most likely to justify Wi-Fi 7 on an operational site. |
| 4096-QAM (4K-QAM) | A denser modulation, carrying more bits per symbol. | Requires a very high signal-to-noise ratio, so it applies close to the access point in clean conditions. Adds peak rate in the best part of the cell and nothing at all at the edge, which is where the problems usually are. |
| Multi-RU and preamble puncturing | Allows a wide channel to be used with part of it blocked out, rather than abandoning the width entirely because of interference in one section. | Quietly useful. It makes wide channels more practical in real spectrum that is not clean, which is most spectrum. |
The real constraint is the client fleet
An access point's capabilities are a ceiling, not a floor. What a network delivers is determined by the devices connecting to it, and in most organisations that fleet is mixed and old.
Before specifying anything, the useful exercise is an audit of what is actually associating today:
- How many devices support 6 GHz at all? Only 6E and Wi-Fi 7 clients do. A large share of laptops, and almost all handheld terminals, industrial devices, IoT sensors, printers, building-services controllers and payment terminals in service today do not — and many will not be replaced for years.
- How many support WPA3? Devices that cannot do WPA3 with protected management frames cannot join a 6 GHz network, regardless of what else they support.
- What are those devices doing? A warehouse full of scanners running short transactional exchanges is not throughput-limited. Doubling the available channel width changes nothing for it. A design office moving large files, or a lecture theatre full of video, is a different case.
- Which devices are your problem devices? If the complaints come from voice handsets and scanners, they are almost certainly 2.4 or 5 GHz devices, and a 6 GHz radio will not touch the fault.
The pattern we see repeatedly is a site buying a generation of access points to solve a problem the previous generation was fully capable of solving — because the real causes were coverage gaps, transmit power left on automatic, overlapping channels, legacy rates still enabled and roaming thresholds untouched since installation. New hardware inherits every one of those faults.
When a refresh is justified — and when it isn't
Reasonable grounds for moving to 6E or Wi-Fi 7
- The existing hardware is genuinely out of headroom — too few radios, too few spatial streams, or a generation that cannot handle the client density you now have.
- The device fleet has already moved. If a substantial and growing share of clients support 6 GHz, leaving that spectrum unused is leaving capacity on the table — particularly in dense environments where 2.4 and 5 GHz are congested.
- You are in contested spectrum. Multi-tenant buildings, dense high streets, blocks of flats, campuses. 6 GHz is quiet in a way 2.4 GHz has not been for twenty years, and that quiet is worth real money in the right building.
- Latency-sensitive traffic on a busy site. This is the multi-link operation argument, and it is a legitimate one.
- End of support. Hardware out of vendor support, or a controller or licensing model being retired, is a perfectly good reason on its own.
- You are replacing it anyway. If the estate is due a refresh on age, specifying current-generation equipment is sensible. The design still has to be redone; the hardware choice is the easy part.
Poor grounds
- The Wi-Fi is bad and nobody knows why. Find out first. A refresh is an expensive diagnostic.
- Because the standard is newer. Not a reason.
- To fix coverage. New access points in the same positions cover the same area, and the 6 GHz radio covers less of it.
- To fix roaming. Roaming is a design and configuration problem. It follows the hardware across.
- Because throughput figures on the datasheet are higher. Those are theoretical aggregate rates under ideal conditions with an unrealistic client. No site achieves them, and no site needs to.
Let the survey decide it
The consistent conclusion is that the standard is rarely the bottleneck. Coverage design, channel and power planning, roaming configuration, airtime policy and the age of the client fleet account for most of the difference between a network that works and one that does not — and none of those are printed on a datasheet.
What settles the question is measurement: what the spectrum in your building actually looks like, what is on the network today and what it can support, where the coverage genuinely fails and at what signal level, what the neighbours are doing, and what the applications actually require. From that, the answer is usually one of three things — reconfigure what you have, extend it, or replace it — and only the third is a refresh.
Where a refresh is right, the same measurement produces the design: a 6 GHz-capable layout with the access point count the physics requires rather than the count the old plan happened to have, a band and channel-width strategy suited to the fleet, and a migration that keeps the older devices working throughout. That is a survey and a design question, and it is worth answering before the purchase order rather than after it. If the current network is simply underperforming and nobody can say why, start with troubleshooting instead — it is a much cheaper way to find out whether you have a hardware problem at all.
Let's find out what your Wi-Fi is actually doing.
Book a survey and you get measured data, a design you can build to, and a number you can budget against — not an opinion.