The 6 GHz band is the biggest wireless spectrum expansion in twenty years. Wi-Fi 7 builds on top of it with 320 MHz channels and Multi-Link Operation that lets a single device run across three bands at once. Wi-Fi 8 is in late draft, targeting reliability instead of raw speed. None of that is the hard part. The hard part — the part that quietly turns a wireless RFP into a building project — is everything underneath: the cable, the switch, and the UPS that have to change to support it. This is the conversation we have with most customers who call us about a Wi-Fi upgrade.
Where the standards actually are right now
For a lot of buildings we walk, the Wi-Fi conversation collapses three different decisions into one — partly because the marketing collapses them too. So here's the brief, current state of each:
- Wi-Fi 6E is the version most enterprise rollouts are still landing on. The 6 GHz band has been usable in the U.S. since 2020, and 6E gear is mature, broadly supported, and priced for mainstream procurement.
- Wi-Fi 7 (IEEE 802.11be) is here. The Wi-Fi Alliance began certifying Wi-Fi 7 devices in January 2024, and IEEE Std 802.11be-2024 was officially published on July 22, 2025. [1] A second wave of certified features (Release 2) opened in December 2025. Enterprise-grade Wi-Fi 7 access points are widely available in 2026 from every major vendor.
- Wi-Fi 8 (IEEE 802.11bn, "Ultra High Reliability") is in draft. Wi-Fi Alliance certification is targeted for January 2028 and IEEE final approval for March 2028. [2] Notably, Wi-Fi 8 is not a speed play — its theoretical peak rate is the same as Wi-Fi 7. The goal is 25% better throughput in challenging signal conditions, 25% lower 95th-percentile latency, and 25% fewer dropped packets during roaming. [3]
The practical takeaway: most buildings should be planning a Wi-Fi 6E or Wi-Fi 7 deployment in the next 12 to 24 months, and they should plan it as a 10-year cable plant decision regardless of which generation they buy first. Wi-Fi 8 won't change the cable, the switch, or the UPS conversation — it'll just make the radios smarter at staying connected.
And in case the timing sounds aggressive — it isn't, the rest of the industry is already moving. Cisco's State of Wireless 2026, a survey of 6,098 wireless professionals across 30 markets, found that 19% of organizations are currently deploying Wi-Fi 6E or 7, with 59% planning to deploy in the next year. [12] Cisco's own deployment telemetry across its installed base recorded a 23% increase in 6E and 7 access point shipments in the second half of 2025, and a 60% jump in 6 GHz clients going live across the year. [12] The refresh wave isn't speculative. It's already moving — and what it does to the closet is the rest of this article.
What 6 GHz and 320 MHz actually give you
For the kinds of rooms our clients build — classrooms, conference rooms, hospital wards, manufacturing floors, warehouse cross-docks — the practical benefit of 6E and 7 isn't "faster Wi-Fi", it's predictable Wi-Fi under density. You stop having to explain why the room worked on Tuesday and didn't on Wednesday.
Here's where the numbers come from. The 6 GHz band gives you up to seven non-overlapping 160 MHz channels, versus the two you can realistically use in 5 GHz once DFS and radar avoidance chew up the rest. [4] Crucially, no legacy clients exist on 6 GHz — if a device is on the band, it's Wi-Fi 6E or newer. That alone removes a huge source of the "why is this slow" calls.
Wi-Fi 7 then adds three things on top of that foundation: 320 MHz channels (doubling the per-channel width), 4096-QAM modulation (denser data per symbol), and Multi-Link Operation (MLO) — the ability for a client and AP to use 2.4, 5, and 6 GHz simultaneously rather than picking one. [4] MLO is mandatory for Wi-Fi 7 certification and is what lets Wi-Fi 7 hit the headline 46 Gbps theoretical peak rate. [5]
Now the catch: a Wi-Fi 7 access point running all three radios at multi-gigabit rates is doing real data work, and that data has to go somewhere over the wire. [5] Which brings us to the part of the project most RFPs leave out.
The cabling problem nobody puts in the quote
A Wi-Fi 6E/7 access point is a tri-band radio doing concurrent multi-gigabit work. That means your AP backhaul has to actually move data — 2.5 Gbps is the floor for a reasonable enterprise deployment, 5 Gbps is where the busier APs want to live, and 10 Gbps shows up on flagship and high-density configurations.
Cat 5e won't do it. Cat 6 is marginal past 2.5G over typical run lengths. Cat 6A is the floor for any 6E or 7 project that's going to stay in service for the lifetime of the cable plant — and a decade is the right planning horizon, because nobody re-pulls horizontal runs that are passing traffic. The IEEE 802.3bt standard itself recommends Cat 6A for new installations; lower categories are technically permitted at lower power levels, but every modern deployment we walk lands on Cat 6A. [6]
Copper vs. copper-clad aluminum — the one we always have to talk about
This is where the cable conversation gets serious, and where we sometimes have to deliver news the customer doesn't want to hear. Genuine solid-copper Cat 6A is expensive, stiffer than its predecessors, and harder to pull. The market has responded with copper-clad aluminum (CCA) cable — aluminum conductors with a thin copper plating on the outside. It's cheaper, lighter, easier to install, and on some spec sheets it appears to pass the performance numbers.
It does not pass the codes. CCA is explicitly excluded from ANSI/TIA-568.2-D, which requires conductors in Category-rated cable to be solid or stranded copper. [7] CCA is prohibited under the National Electrical Code (NFPA 70) for twisted-pair Category cable communications because of fire-hazard concerns. [8] CCA cables do not have a valid UL safety listing per the NEC, and per Fluke Networks "can't be legally installed into areas of buildings that require CM, CMG, CMX, CMR or CMP rated cables", which covers virtually every commercial installation. [9]
The performance is also a problem. CCA has roughly 55% greater electrical resistance than copper for the same diameter, and it routinely fails the DC Resistance Unbalance test that PoE applications rely on. [9] What that translates to in the field: more voltage drop across the run, more heat in the cable bundle, and under-voltage at the powered device — which means your access points reset, your cameras flicker, your phones go offline, and your help desk gets the calls. The cable looks fine; the symptoms are everywhere else.
We've seen CCA sold as "Cat 6A compatible" through three different distribution channels in the last 18 months. If you're inheriting an existing cable plant that's about to support a 6E or 7 rollout, getting a sample tested before you commit to the AP count is one of the cheapest insurance policies on the project.
The power problem
A modern Wi-Fi 6E or Wi-Fi 7 access point in three-radio operation wants real power at the port. The first generation of 6E APs sat comfortably on PoE+ (802.3at, 30W). Wi-Fi 7 flagship APs — with multiple radios, sensor radios, IoT support, and dual multi-gigabit Ethernet ports — push past that ceiling. [10]
Cisco's published Wi-Fi 7 design guidance is unambiguous: "since the introduction of Wi-Fi 6E, it is best to plan for 802.3bt power or 60W ports." [10] The Cisco Meraki Wi-Fi 7 technical guide is even more direct: "802.3bt (UPOE) is the recommended power input for full operation of Wi-Fi 7 Access Points." [5]
So the power conversation, simplified: plan for 802.3bt at the switch port. PoE+ may technically light up the AP, but you'll be running on reduced functionality — disabled radios, disabled USB, disabled IoT — and you'll discover the limits at the worst possible time.
Figures from IEEE 802.3bt-2018 (incorporated into IEEE 802.3-2022). [11][6]
What 802.3bt does to the rest of the closet
Here's the load math most planning conversations skip. A 48-port PoE+ switch with an average of 15W per port draws about 720W of PoE output, plus 100–150W for the switch itself. Call it ~900W at the receptacle. A 48-port 802.3bt switch — populated with Wi-Fi 6E or 7 access points pulling 40–60W each — can sit at 2.5 to 3 kW at full density. You've just multiplied the switch closet load by a factor of three.
The UPS that was sized for the old load doesn't know it's on a different switch plan. A 1.5 kVA or 2.2 kVA rack UPS comfortably supporting the legacy switch is suddenly at 150% of its rated capacity, running on overload protection until it shuts down under fault. This is the most common version of "the AP upgrade broke the closet" we walk into.
The fix, in order of preference: size up the UPS to the new load with the runtime requirement explicit; add intelligent load shedding that drops PoE to non-critical ports under battery; or — for closets carrying real density — move from VRLA to a lithium UPS that handles the higher discharge rate and tighter footprint more comfortably. (We've written about that conversation in detail here.) Any of these decisions has to happen before the new switch hits the dock, not after.
What Wi-Fi 7 specifically asks of your network
Beyond the power and cabling baseline that 6E established, Wi-Fi 7 adds a few specific requirements to the wired side that are worth knowing about:
- Multi-gigabit switching is no longer optional. The combined throughput of all three radios on a Wi-Fi 7 AP exceeds 1 Gbps in routine operation. A 1G switch port creates a backhaul bottleneck that defeats the upgrade. [5] Plan for 2.5G access ports at minimum.
- WPA3 becomes mandatory. The 802.11be data rate and Multi-Link Operation features require WPA3 / Enhanced Open. WPA2 was permitted on Wi-Fi 6 and below; on Wi-Fi 7 it isn't. [5] Your authentication infrastructure (RADIUS, certificates, identity provider integration) needs to be ready for it.
- RF coverage on 6 GHz is shorter than 5 GHz at equivalent power levels, because higher-frequency signals attenuate faster through walls and people. Expect to deploy more APs for the same coverage area than you would have for an equivalent 802.11ac plan — which compounds every cable, switch, and UPS conversation above.
What Wi-Fi 8 changes — and what it doesn't
The short answer: Wi-Fi 8 doesn't change the infrastructure conversation in any meaningful way. Same spectrum (2.4, 5, 6 GHz), same maximum channel width (320 MHz), same 4096-QAM, same eight-spatial-stream ceiling. [3] What changes is what the APs do with that foundation: Multi-AP coordination (APs sharing channel and beamforming decisions across the whole RF cell), distributed Multi-Link Operation, Enhanced Long Range for edge clients, and improved roaming behavior. The goal across all of it is reliability, not headline throughput. [2]
For planning purposes that means: if you're sized correctly for Wi-Fi 7, you're sized correctly for Wi-Fi 8. Cat 6A still floors you. 802.3bt still feeds you. WPA3 still authenticates you. The new behaviors will arrive as firmware and silicon updates on top of the same physical infrastructure. So the 10-year cable plant decision you make today doesn't need to wait for 2028 — it needs to be right for the platform that ships now.
The planning order that actually works
If we were spec'ing a 6E or 7 rollout in your building today, here's the order of decisions we'd make:
- AP count and placement first. Site survey based on 6 GHz propagation. Expect more APs than you had under 802.11ac for equivalent coverage.
- Cable plant second. Cat 6A solid copper, F/UTP minimum, jacket rating verified against pathway capacity. If existing cable is in place, sample-test before you commit to the AP count — particularly if the cable was installed in the last few years and the price seemed too good.
- Switch and PoE strategy third. 802.3bt-capable access switching, multi-gigabit ports, PoE budget calculated against the new AP count.
- UPS and cooling fourth. Size the UPS to the new load with runtime requirements made explicit. Consider whether a lithium UPS makes sense for the new density profile.
- Authentication and management last. WPA3 readiness, controller / cloud platform updates for Wi-Fi 7 visibility (MLO analytics, 320 MHz channel planning), and security policy review.
Doing the work in this order means every downstream decision is grounded in the one above it. Doing it in any other order — which is how most projects actually go, because the APs were ordered first — means you pay for half the infrastructure twice.
- IEEE 802.11 Working Group. Wireless LAN standards timeline (IEEE Std 802.11be-2024 publication date and certification milestones). ieee802.org/11
- IEEE 802.11bn (Wi-Fi 8). Ultra High Reliability — standards timeline, technical scope, performance targets. en.wikipedia.org/wiki/IEEE_802.11bn
- Tom's Hardware (2025). Next-gen Wi-Fi 8 focuses on reliability instead of speed. tomshardware.com
- IEEE 802.11be-2024 (Wi-Fi 7). Multi-Link Operation, 320 MHz channels, 4096-QAM, certification timeline. en.wikipedia.org/wiki/IEEE_802.11be
- Cisco Meraki (2025). Wi-Fi 7 (802.11be) Technical Guide — power, cabling, security, MLO requirements. documentation.meraki.com
- Eaton / Tripp Lite. Overview of Power over Ethernet (PoE) Technology — IEEE 802.3bt specifications and Cat 6A recommendation. tripplite.eaton.com
- trueCABLE. The ANSI/TIA 568 Series of Specifications: What is Most Important to Know for Copper. truecable.com (TIA-568 reference)
- trueCABLE (2026). Copper Clad Aluminum vs Copper: The Ultimate Test on a Fluke Versiv DSX Cable Analyzer. truecable.com (CCA test results)
- Fluke Networks. Copper Clad Aluminum (CCA) Cables — application note on standards non-compliance and DC Resistance Unbalance failures. flukenetworks.com
- Cisco Systems. Wi-Fi 7 and the Growing Future of Wireless Design Guide — explicit 802.3bt power planning recommendation. cisco.com
- Power over Ethernet (Wikipedia). IEEE 802.3bt-2018 Type 3 / Type 4 specifications and 4-pair PoE technical detail. en.wikipedia.org/wiki/Power_over_Ethernet
- Cisco Systems (April 2026). State of Wireless 2026: Unlocking the Multiplier Effect — global survey of 6,098 wireless professionals across 30 markets, plus Cisco platform deployment telemetry. cisco.com (State of Wireless 2026)