Is WiFi 7 wasted on your network?

Wireless throughput is a relay race, and the slowest runner sets the time. The chain runs: client radio → does the client support 6 GHz and MLO → AP radio → AP uplink port → switch port → cabling → gateway and ISP → PoE budget. A WiFi 7 AP only helps to the extent every other link can keep up.

The most common and most expensive mistake follows directly: hanging a WiFi 7 access point off a 1 GbE uplink. The radio can move multiple gigabits; the wire caps it at one (and around 940 Mbps usable after overhead). The shortfall between the capability you bought and the throughput you get isn’t a signal problem — it’s a hardware mismatch, and it’s the single most common WiFi 7 deployment error of 2025 and 2026.¹¹⁵ So the first question before choosing an AP is not “how fast is the radio?” but “what can my uplink, switch, cabling, ISP and clients actually carry?”

Here is the part the marketing quietly omits: MLO, 320 MHz and 4K-QAM all require WiFi 7 on the client device. A WiFi 6, 6E or 5 device connects to a WiFi 7 access point and works perfectly — at its own maximum standard, receiving none of WiFi 7’s exclusive features.¹⁰¹⁴ If your laptop’s adapter tops out at 160 MHz, the AP talks to it at 160 MHz and the 320 MHz capability simply sits idle.¹¹

And as of 2026, WiFi 7 clients are common but not yet the majority of what’s in the field — flagship phones and laptops from 2024 and 2025 increasingly ship WiFi 7 chipsets, while the broad installed base lags, and client devices generally trail the access points in stream count and capability.¹²¹⁴ On the infrastructure side, by contrast, the shift is already decisive: WiFi 7 reached about 44% of enterprise wireless access-point revenue in the first quarter of 2026, with that revenue up roughly 350% year over year — the gear has arrived ahead of the clients.¹⁷ The practical consequence: count your WiFi 7 clients before you buy for WiFi 7 features. The reliability and latency wins of MLO land only on MLO-capable clients; everyone else gets a (genuinely excellent) WiFi 6E-class experience from your shiny new AP — which is fine, as long as that’s what you meant to pay for.

UniFi’s WiFi 7 range is easiest to understand if you ignore the marketing tiers and sort by uplink port, because that’s where the real decision lives.

• U7 Pro (~$189): 6-stream tri-band, 2.5 GbE uplink, PoE+. The practical WiFi 7 AP for small offices and busy homes on gigabit/2.5 GbE switching.¹⁹
• U7 Pro Max (~$279): 8-stream, dedicated spectral-scanning radio, 500+ clients, ~1,750 sq ft, 2.5 GbE uplink, PoE+ (802.3at, 25 W).²⁴
• U7 Pro XG (~$199): 6-stream (2×2 all bands), 10 GbE uplink (1/2.5/5/10 G), PoE+ (22 W); Ubiquiti rates it over 3 Gbps aggregate, 300+ clients, ~1,500 sq ft.⁵⁷
• U7 Pro XGS (~$299): 8-stream (4×4 on 5 GHz), dedicated spectral scanning, Zero-Wait DFS, 10 GbE uplink, PoE++ (802.3bt, 29 W), 500+ clients, ~1,750 sq ft.⁴⁷⁸
• E7 (enterprise): 10-stream, 10 GbE uplink, AFC — the full-featured indoor flagship.⁹⁸
• Specialist models: U7 In-Wall (omits 6 GHz, integrated 2.5 GbE PoE switch) and U7 Mesh (dual-band, weatherproof outdoor-capable, 2.5 GbE injector included, gigabit-class real-world rates).⁹⁸

The crux is the Pro Max versus the XGS. On the radio they are nearly the same access point — same 8 streams, same spectral scanning, same coverage and client ratings; the differences that matter are the uplink (2.5 GbE versus 10 GbE), the power class (PoE+ versus PoE++), and Zero-Wait DFS on the XGS.⁴ You are not buying a faster radio at $299; you are buying a faster wire. And the corollary at the other end: pairing a 10 GbE access point like the U7 Pro XG with a 1 GbE switch port wastes most of what you paid for — the multi-gig port negotiates down and connects fine, but it caps a radio capable of several gigabits at a single gigabit, and on a 1 GbE-switched site the much cheaper U7 Pro delivers similar real-world performance — so the 10 GbE models only earn their price where there’s multi-gig (ideally 10 GbE) switching to feed them.⁵

This is where WiFi 7 budgets quietly balloon: the access point is the cheap part. To get value from a multi-gigabit radio, the wired backbone has to match it.

• Switching. A 2.5 GbE AP needs a 2.5 GbE PoE port; a 10 GbE AP (XG, XGS, E7) needs a 10 GbE PoE port — which is why Ubiquiti expanded its XG PoE switch line alongside these APs.⁷ Aggregate-throughput claims also assume the switch and its uplinks can carry the sum.
• PoE budget. Plan PoE+ (802.3at, ~25 W) for the Pro/Pro Max/XG and PoE++ (802.3bt, ~29 W) for the XGS, and sum the budget across every AP before you choose the switch — the XGS in particular will not run to full capability on a PoE+-only switch.⁴⁷
• Cabling. 10 GbE needs Cat6a to reach the full 100 m (Cat6 manages only ~55 m); 2.5 GbE is far more forgiving and usually runs fine on existing Cat5e.⁶ In older buildings the cabling — not the AP — is frequently the real project.

If a quote for “WiFi 7” doesn’t mention the switch, the PoE budget, and the cable, it’s quoting the cheap part and hiding the spend.

The strongest practical case for WiFi 7 in a city like New York isn’t the 320 MHz speed headline; it’s the 6 GHz band and MLO. The 6 GHz band is clean spectrum that hasn’t yet filled with neighbours, which is exactly the relief a congested environment needs — apartments, multi-tenant offices, dense building stacks where 2.4 and 5 GHz are saturated — and field trials confirm MLO doubles throughput under interference and roughly halves latency in real homes.¹⁰¹⁵

But two qualifications matter, and both cut against the spec sheet. First, 320 MHz channels live only on 6 GHz and the band holds only a small number of them, so 320 MHz is expected to be rarely usable in dense, enterprise-style deployments and is more of a home-router luxury — in a crowded building you may never get a clean one.¹³¹⁶ Second, 6 GHz has shorter range and loses more signal through walls than 5 GHz, so in New York’s plaster-and-brick construction, full 6 GHz coverage often means more access points, with MLO leaning on 5 GHz to fill the gaps.¹⁰¹³ (Outdoor and standard-power 6 GHz add AFC coordination, as on the E7 and U7 Pro Outdoor.⁸) The honest NYC takeaway: the congestion-relief case for 6 GHz and MLO is strong; the 320 MHz speed case is weak in dense spectrum. Design for coverage and a deliberate band policy, not for the headline channel width.

• Worth it when your clients are WiFi 7 (2023-and- later flagships) and you have the problem WiFi 7 solves — high density (think 50+ clients per AP), RF congestion, or latency-sensitive and 4K/8K work — and the wired side can carry it (2.5 or 10 GbE uplink, matching switch and PoE, Cat6/Cat6a).¹¹¹⁴⁶
• Wait or stay put when your internet plan is below ~500 Mbps, your devices are mostly more than four years old with no refresh planned, and you run fewer than ten devices with no 4K or real-time gaming — a WiFi 6 or 6E system already meets that need.¹¹
• Don’t spend on a 10 GbE AP feeding a 1 GbE switch, on WiFi 7 features your current clients can’t use, or on 320 MHz ambitions in a spectrum-dense building.
• The pragmatic path for new installs and refreshes: buy WiFi 7 on forward-compatibility grounds — skip WiFi 6E, since WiFi 7 now costs about the same and is strictly broader — but size the uplink, switch, PoE and cabling to the radio, and phase WiFi 7 into your highest-traffic, highest-density areas first while WiFi 6 covers the rest, which UniFi mixes cleanly in one network.¹⁰¹¹⁶

Match the access point to the actual bottleneck: a 1 GbE or 2.5 GbE-switched site is a U7 Pro or U7 Pro Max; a genuine 10 GbE backbone with real density is where the XGS or E7 finally earns the difference.

• Every throughput number here is third-party and representative. The ~609 / 1,133 Mbps U7 Pro figures, the 3.2 Gbps 6 GHz CNET result, the 2–4 and 6–15 Gbps enterprise ranges — all come from independent tests in their environments, not ours and not yours.⁹¹⁰¹⁴ The only number that describes your site is one measured at your site.
• “Wirelessly near-identical” is a most-deployments statement. The Pro Max and XGS do differ at the edges (Zero-Wait DFS, the 4×4 specifics), and in a true 10 GbE, very-high-density site the XGS’s headroom is real — but the radios are close enough that, for most, the uplink and PoE are the honest deciding factors.⁴
• The lineup and prices move. Models, pricing and availability change (the Pro Max has been intermittently sold out); confirm current specs on Ubiquiti’s tech-specs and store before buying.⁴³
• Client adoption is a moving target. WiFi 7 client share is climbing quickly, so the “your clients can’t use it” caveat weakens every year — buying for the next few years of devices is a legitimate reason even if today’s fleet is older.¹²¹¹
• 6 GHz rules vary by region and power class and use AFC for standard-power outdoor use; indoor low-power is simpler. Plan band policy deliberately rather than assuming the band is free to use everywhere.¹³⁸
• This is engineering judgement, not a benchmark report. We’ve cited what credible testers measured and reasoned from the systems chain; the unique answer for your building is a site survey and a throughput test — which is the work we’d actually do, and the only thing that turns “WiFi 7 is fast” into “WiFi 7 is faster here.”

The headline holds up to scrutiny. WiFi 7 is worth buying for most new installs on forward-compatibility alone — but whether it makes your network faster depends almost entirely on the parts that aren’t the access point. Size the uplink, the switch, the PoE and the cabling to the radio; count the clients that can actually use it; respect what 6 GHz can and can’t do in a dense building; and when in doubt, measure before you spend.