Wi-Fi 7 in 2026: who it actually helps, and who should skip it
Router boxes now advertise five-digit numbers. Almost none of that figure is available to any single device in your house — but one of Wi-Fi 7's three headline features is genuinely worth paying for, and it is not the one on the front of the box.
A "BE19000" router is not a 19 Gbps router. That number is the theoretical maximum of every radio in the box added together, under laboratory conditions, with channel widths that may not be legal where you live. No client device sees it, ever. It exists so that the box with the bigger number sells more units than the box next to it.
Underneath the marketing, though, Wi-Fi 7 is a real standard with three specific changes. Two of them are conditional. One of them fixes something people complain about every day without knowing there was a fix.
On this page
The short answer
- Upgrade if your home has a lot of simultaneous devices, a mesh system without wired backhaul, or an internet plan above roughly 1 Gbps.
- Skip it if your plan is 500 Mbps or less and your devices are mostly Wi-Fi 5 and Wi-Fi 6. A better-placed Wi-Fi 6 router will beat a badly placed Wi-Fi 7 one every time.
- Multi-Link Operation is the real feature. It targets stalls and latency spikes, not peak throughput.
- 320 MHz channels need 6 GHz spectrum your country may not have allocated in full.
- 4K QAM works close to the router and quietly disappears as you walk away.
What Wi-Fi 7 actually adds
The Wi-Fi Alliance's own description of Wi-Fi CERTIFIED 7 lists the changes plainly. Channels can be 320 MHz wide in the 6 GHz band, which the Alliance describes as twice the throughput of Wi-Fi 6. Multi-Link Operation lets a device use more than one band at the same time. And 4K QAM — 4096-point quadrature amplitude modulation — packs more bits into each transmission, for roughly 20 % higher transmission rates than Wi-Fi 6's 1024-QAM.
There are less visible additions too: 512 Compressed Block Ack reduces acknowledgement overhead, and multiple resource units can now be assigned to a single device, which helps when parts of a channel are busy. Those are efficiency gains you will never see in a speed test but that show up in a crowded flat.
| Feature | What it needs | What you notice |
|---|---|---|
| 320 MHz channels | 6 GHz spectrum, short range, a 6 GHz client | Very high peak speeds in the same room |
| Multi-Link Operation | Wi-Fi 7 on both router and client | Fewer stalls, steadier latency, better mesh |
| 4K QAM | An excellent signal-to-noise ratio | About 20 % more throughput, close in |
| 512 Compressed Block Ack | Nothing on your part | Slightly more efficient busy networks |
The spectrum rule that changes the answer by country
Here is the part router marketing never mentions. A 320 MHz channel only fits in the 6 GHz band, and how much 6 GHz band exists depends on your regulator.
The United States opened the entire 5925–7125 MHz range — 1200 MHz — for licence-exempt use, and Canada, South Korea and several others did the same. The European Union and the United Kingdom opened only the lower portion, 5925–6425 MHz, which is 500 MHz. Ofcom has said it wants to authorise low-power indoor Wi-Fi across the upper 6 GHz band as quickly as possible, and the EU is still deciding whether that upper stretch is shared with mobile networks — but today, the difference is stark.
The practical translation: in the US, 320 MHz is a feature you can plan around. In Europe, it is a feature that works beautifully in a detached house and degrades in a dense building, because there is only one such channel to go around and everyone's router wants it. Many European routers sensibly default to 160 MHz for that reason.
Multi-Link Operation: the feature that earns its keep
Every previous generation of Wi-Fi made your device pick a band and stay there. If you were on 5 GHz and the channel got congested, or you walked behind a wall and the signal weakened, you either suffered or your device made a slow, visible decision to move to 2.4 GHz — the pause everyone knows as "the Wi-Fi dropped for a second".
Multi-Link Operation removes the choice. A Wi-Fi 7 client can associate with two bands at once and use them together, either splitting traffic across both for throughput or duplicating time-sensitive packets across both for reliability. When one link gets noisy, the other carries on. There is no re-association, no gap.
This is also the reason Wi-Fi 7 mesh systems are a bigger jump than Wi-Fi 7 single routers. A mesh node without an Ethernet cable back to the main router has to spend radio time talking to the router and to your devices. Being able to hold two links at once makes that backhaul far less fragile.
Keep reading
Advertisement
4K QAM and the fine print on distance
Modulation is how many distinct states a radio squeezes into each symbol it sends. Wi-Fi 6 tops out at 1024-QAM; Wi-Fi 7 adds 4096-QAM, worth about 20 % more throughput according to the Wi-Fi Alliance.
The catch is inherent to the technique. Packing 4096 states into the same signal space means the states sit closer together, so the receiver needs a much cleaner signal to tell them apart. In practice that means a strong signal at short range. Move to the next room and the link quietly steps down to a lower modulation — as it should, because a lower rate that works beats a higher rate that fails.
How to read a review
When a Wi-Fi 7 review reports a spectacular number, check the test distance. Figures taken at two metres with a 320 MHz channel and 4K QAM describe the best case. The interesting figure is the one from the far bedroom.
Your bottleneck is probably not the router
Before spending anything, work out which link in your chain is actually slowest. For most households it is the internet plan, and no router improves that. For a surprising number it is a single Wi-Fi 5 device — a smart TV, an older tablet — dragging down the airtime available to everything else.
- Run a speed test from a wired computer. If it matches your plan, the router's WAN side is fine and the problem is wireless. If it does not, call the ISP.
- Compare a same-room wireless test with a far-room one. A big gap is a coverage problem — solved by placement or a mesh node, not by a faster standard.
- Check what your devices actually support. A Wi-Fi 7 router talking to a Wi-Fi 5 laptop delivers Wi-Fi 5.
- Count the simultaneous streams. Wi-Fi 7's efficiency gains scale with device count; in a two-person flat with six devices there is little congestion to relieve.
Who should upgrade
| Situation | Verdict |
|---|---|
| Internet plan under 500 Mbps, few devices | Skip. Spend the money on router placement or one Ethernet run. |
| Multi-gigabit fibre, wired-capable devices | Worth it — mostly for the 2.5 GbE and 10 GbE ports that come with these routers. |
| Mesh without wired backhaul | Strongest case. MLO makes the backhaul noticeably steadier. |
| Flat in a dense building, Europe | Mixed. MLO helps; 320 MHz channels mostly will not. |
| Home working with constant video calls | Worth it for latency consistency, not for peak speed. |
| Router is five or more years old | Upgrade — but Wi-Fi 6 hardware at half the price may serve you equally well. |
Should you wait for Wi-Fi 8?
No — not on the current timeline, and not for the reasons people expect. IEEE 802.11bn, marketed as Wi-Fi 8, is explicitly not a speed generation. Its working name is Ultra High Reliability, and its stated targets are about 25 % higher throughput in poor signal conditions, about 25 % lower latency at the 95th percentile, and about 25 % fewer dropped packets while roaming between access points.
The schedule puts Draft 1.0 in mid-2025, certification launching around the end of 2027, and formal publication in March 2028. Consumer hardware follows certification, not the draft. Waiting means waiting years for a standard whose benefits are mostly invisible on a good network.
Which brings the whole thing back to a single sentence. Wi-Fi 7 is a genuine improvement in consistency dressed up as an improvement in speed. Buy it if consistency is your problem. If your problem is a 200 Mbps plan and a router in a cupboard, no standard on any roadmap will fix that.
