Wi-Fi 7 range explained: 6 GHz vs 5 GHz through walls and floors

The short version

• Physics costs 6 GHz less than you'd think. In free space, a signal at 6.5 GHz loses 1.5 dB more than one at 5.5 GHz over the same distance. At equal power, that's about 85% of the distance, not half.
• Walls add a little more. In NIST's laboratory measurements, drywall, glass, plywood, lumber and brick each lose at most about half a decibel more at 6.5 GHz than at 5.5 GHz. A 203 mm (8 in) hollow concrete block wall loses about 1 dB more, and 102 mm (4 in) of solid concrete about 2 dB more. The wall itself costs far more than the change of band.
• The bigger gap can be the FCC power rules. Most home routers run 6 GHz as "low-power indoor" access points, limited to 5 dBm per megahertz. On 5 GHz, the rules allow 6 to 12 dB more power per megahertz on an 80 MHz channel. Where a 5 GHz router uses that allowance, the gap is several times the physics.
• Wider 6 GHz channels don't reach farther. The limit is set per megahertz, so a 320 MHz channel gets more total power (30 dBm against 18 dBm at 20 MHz) but the same signal-to-noise ratio. It carries more data where the signal is strong. It doesn't push the edge out.
• Your phone is the weaker end. A phone on a low-power indoor 6 GHz network may transmit at most −1 dBm per megahertz, 6 dB below the router.
• Worked example: 10 m (33 ft) away through two drywall partitions, 6 GHz at 160 MHz still reaches its top Wi-Fi 7 rate on paper, with twice the link rate of 5 GHz at 80 MHz. Through one concrete block wall at 12 m, the low-power indoor 6 GHz link falls about 6 dB short of even the Wi-Fi 6-level top rate, while 5 GHz still reaches it.
• Standard power with AFC lifts the 6 GHz limit to 36 dBm, but only on routers and devices that support it. We found documentation for two ASUS routers, plus an ASUS model whose product page describes it, a TP-Link outdoor unit that says support will come in an update, and a Google mesh system its maker calls "AFC ready".

Why higher frequencies lose more, and how much more

Free-space path loss is the loss between two antennas with nothing in the way. The ITU's recommendation on free-space attenuation (ITU-R P.525) gives it for two isotropic antennas as:

The frequency term is the part that matters here. Between two frequencies at the same distance, the difference is 20 log₁₀(f₂ / f₁). For 6.5 GHz, in the middle of the 6 GHz band, against 5.5 GHz, near the middle of 5 GHz Wi-Fi:

From the bottom of 5 GHz Wi-Fi (5.15 GHz) to the top of 6 GHz (7.125 GHz), the same formula gives 2.8 dB. That fits Cisco's Wi-Fi 6E white paper, which models the difference between the two bands as "actually minimal", puts the total from the bottom of 5 GHz to the top of 6 GHz under 3 dB, and adds: "While it's true that higher frequencies don't go as far, it's not at all a deal breaker."

Why frequency matters at all: the receiving antenna shrinks

A radio wave spreads out at the same rate at every frequency. The extra loss at higher frequencies comes from the receiving antenna, not from the air. An antenna's effective capture area for a given gain is proportional to the wavelength squared (for a 0 dBi antenna it is λ²/4π, the effective aperture ITU-R P.525 uses for an isotropic receiving antenna). The 4πd/λ in the ITU formula combines the spreading with that capture area. At 5.5 GHz the wavelength is 5.45 cm, and a 0 dBi antenna captures about 2.4 cm². At 6.5 GHz the wavelength is 4.61 cm, and the same antenna captures about 1.7 cm² (RouterHaus's arithmetic). Less area intercepts less of the passing wave, and that ratio is exactly the 1.45 dB.

In free space, 1.45 dB means the 6.5 GHz signal reaches the same strength at 5.5/6.5 = 85% of the distance. If that were the whole story, 6 GHz would cover nearly as much of your home as 5 GHz. It doesn't, so look at the power.

The bigger gap: FCC power limits

The FCC opened 1,200 MHz at 5.925 to 7.125 GHz in April 2020. Its announcement "authorizes indoor low-power operations over the full 1,200 megahertz and standard-power devices in 850 megahertz". The limits are in 47 CFR 15.407. They differ from the rules for the main 5 GHz Wi-Fi bands in two ways. They are stated as EIRP, the power actually radiated including antenna gain, where those 5 GHz rules limit conducted power. And for low-power indoor access points and their clients, the class most home routers and phones use, the binding limit is a power spectral density (PSD), a limit per megahertz, rather than a total. (Standard-power and geofenced access points reach their total cap at 20 MHz and wider.)

EIRP by channel width: per megahertz versus a total cap

A per-megahertz limit means the total grows with the channel: limit + 10 log₁₀(width in MHz), up to the total cap. On 5 GHz it works the other way. The rules for 5.15–5.25 GHz (U-NII-1) and 5.725–5.85 GHz (U-NII-3) set a total of 1 W (30 dBm) conducted, the rules for the radar-shared bands between them (U-NII-2) set 250 mW (24 dBm), and each band also has a power-density limit (per megahertz, or per 500 kHz in U-NII-3). Both totals allow an antenna gain of up to 6 dBi without a reduction, which makes 36 dBm and 30 dBm EIRP. So on 5 GHz the total stays put and a wider channel spreads it thinner.

Cisco's white paper and ASUS's AFC support page tabulate the same 18, 21, 24 and 27 dBm figures, and ASUS lists 30 dBm at 320 MHz. At 80 MHz, the regulatory ceiling on 5 GHz is 12 dB higher per megahertz than 6 GHz in U-NII-1 and U-NII-3, and 6 dB higher in U-NII-2. Real routers may not reach the 5 GHz ceilings; few home router makers publish EIRP. Cisco's white paper, for one, models a 5 GHz access point at 23 dBm against a 6 GHz low-power indoor one at 24 dBm, both on 80 MHz channels, and finds the two cells "pretty much the same" size. The power gap only appears when a 5 GHz radio uses more of its allowance than that. One business access point that does publish its figures, Grandstream's GWN7672, lists a maximum transmit power of 25 dBm on 5 GHz and 24 dBm on 6 GHz, with 5.5 dBi antennas on both bands. The datasheet doesn't say whether that's conducted power, and it adds "Maximum power varies by country, frequency band and MCS rate". If it is conducted, the 6 GHz radio with its 5.5 dBi antennas could radiate up to 29.5 dBm, more than the 24 to 27 dBm the indoor rule allows at 80 or 160 MHz, so on 6 GHz the regulation, not the hardware, would set the limit.

What that means for range at a given data rate

Doubling a channel's width doubles the noise it collects, so the signal-to-noise ratio (SNR) falls by 3 dB. Under a per-megahertz limit the power doubles too, so the SNR stays the same. Cisco puts it this way: "The Signal-to-Noise Ratio (SNR) decreases by 3 dB for each doubling of the channel width", and "The power limits expressed in PSD remove the SNR penalties long associated with bonded channels."

The GWN7672 datasheet shows the noise side. To hold its top Wi-Fi 6 rate (MCS 11) on 6 GHz, its receiver needs −57 dBm on an 80 MHz channel, −53 dBm at 160 MHz and −51 dBm at 320 MHz. For the top Wi-Fi 7 rate (MCS 13), it needs −52, −50 and −47 dBm. Each doubling needs 2 to 4 dB more signal, and a 6 GHz low-power router gets 3 dB more power for each doubling. In practice:

• On 6 GHz low-power indoor, every channel width reaches roughly the same distance at the same modulation. The wider channel carries proportionally more data inside that range. A 320 MHz channel doesn't extend coverage, and narrowing to 80 MHz doesn't either.
• On 5 GHz it's different. The total is fixed, so narrowing a channel concentrates power. That's why our dead-zone guide suggests capping 5 GHz at 80 MHz for a shaky far room. The same trick does nothing for 6 GHz under low-power indoor rules.
• Any 20 MHz transmission on 6 GHz is capped at 18 dBm under low-power indoor rules, against up to 36 dBm on 5 GHz U-NII-1 or U-NII-3. Signal-strength readings at the same spot can differ by more than the 1.5 dB of physics suggests. How an iPhone uses those readings to choose a band is in why an iPhone stays on 5 GHz.

320 MHz also needs a client that supports it. Apple's iPhones, iPads and Macs stop at 160 MHz; our generation guide lists the devices that go wider.

The uplink: your phone is the weaker end

A Wi-Fi link has to work both ways. On a low-power indoor 6 GHz network, the client limit is −1 dBm per megahertz, 6 dB below the router's 5 dBm. That's 12 dBm at 20 MHz and 21 dBm at 160 MHz. Cisco: client power "must always remain 6 dB below the maximum EIRP of the AP the client is associated to". If the phone and the router had equally good receivers and antennas, the phone-to-router direction would have 6 dB less margin than the router-to-phone direction. At regulatory maximums, the uplink runs out first.

On 5 GHz the regulatory ceiling for clients is higher: 250 mW (24 dBm) conducted and 11 dBm per megahertz in U-NII-1, and the same totals as access points in the other 5 GHz bands. What phones actually transmit on either band isn't published by the phone makers we checked, so we can't put a number on the 5 GHz uplink. Turning the router's power up doesn't fix the uplink; our dead-zone guide quotes Cisco's warning that it "only increases coverage on the downlink".

Walls and floors: measured losses at both bands

NIST measured building materials between 3 and 8 GHz in 1997 (NIST IR 6055). Horn antennas faced each other 2 m apart, with each specimen in a window in a metal shield that blocked other paths. Each specimen was measured at ten positions and the results averaged. The report publishes curve-fit equations for most specimens, including all of those below. The table evaluates those equations at 5.5 and 6.5 GHz (RouterHaus's arithmetic; the fits match the report's plotted curves).

Three things to take from it. First, NIST's conclusion holds: loss generally rises with frequency and, for non-metallic materials, with density. Second, the step from 5 GHz to 6 GHz adds 0 to 2 dB per obstacle, most for dense concrete. Third, the material matters far more than the band. One 4-inch concrete wall costs more than 20 dB at either band, and nothing about Wi-Fi 7 changes that.

Read the numbers as a floor, not a forecast. They are bare sheets in a lab with the signal hitting them straight on. A real interior wall is two layers of drywall over studs with wiring, insulation and sometimes metal, and a signal crossing it at an angle travels through more material. The drywall "gain" is thin-panel resonance: NIST notes that its thinner drywall and glass specimens sometimes measured at or slightly above 100% transmission at certain frequencies. Its 6 mm and 16 mm drywall curves also stay within 1 dB of no loss from 3 to 8 GHz. The glass was what NIST calls ordinary architectural window glass (soda-lime float glass); NIST didn't test metal-coated low-E glass. Floors aren't in the table. The ITU's indoor planning recommendation (ITU-R P.1238) gives 7 dB for a house floor at 5.2 GHz and doesn't list a 6 GHz figure; our walls-and-floors section has the rest of that table.

A worked link budget: 5 GHz at 80 MHz against 6 GHz at 160 and 320 MHz

A link budget adds up everything between the router's transmitter and your device's receiver. If the signal that arrives beats the receiver's sensitivity for a given data rate, that rate is possible, and the difference is the margin. Every input below is either sourced or labeled as an assumption. Replace the assumptions with your own figures where you have them.

• 5 GHz router EIRP: 30 dBm on an 80 MHz channel. Assumption. It's below the 36 dBm FCC ceiling in U-NII-1 and U-NII-3, equal to the U-NII-2 ceiling, and within what the GWN7672's 25 dBm and 5.5 dBi antennas would allow if that's conducted power. Home router makers rarely publish EIRP. At the 23 dBm Cisco's model uses, every 5 GHz margin below would be 7 dB smaller.
• 6 GHz router EIRP: the FCC low-power indoor maximum for each width (27 dBm at 160 MHz, 30 dBm at 320 MHz), and 36 dBm for standard power. Standard power is the ceiling; the AFC system sets the actual figure at each location.
• Receiver sensitivity: the GWN7672 datasheet's figures for MCS 13 (top Wi-Fi 7 rate) and MCS 11 (top Wi-Fi 6 rate). Assumption: we use an access point's receiver as a stand-in for your phone's, because we found no published receive sensitivity for phones.
• Phone antenna gain: 0 dBi. Assumption. Phone makers don't publish it.
• Losses: free-space loss from ITU-R P.525 at 5.5 and 6.5 GHz; walls from NIST IR 6055 (two drywall partitions counted as no loss, because NIST measured under 0.2 dB per 13 mm sheet at both frequencies); a floor at 7 dB from ITU-R P.1238 at 5.2 GHz, assumed the same at 6.5 GHz because the ITU gives no 6 GHz figure.

Link rates are the IEEE 802.11be figures for a two-stream device that the RouterHaus methodology uses. Link rate isn't throughput, and a margin of a few decibels isn't a guarantee: the model leaves out furniture, the angle through the floor, a hand around the phone and other networks. It also uses free-space loss for distance, while P.1238 adds its floor factors to an indoor distance model whose coefficients already allow for walls and obstacles, so the absolute margins are optimistic on both bands. How to read it:

• Two rooms away (A), every option clears its top rate. 6 GHz at 160 MHz has half the margin of 5 GHz (8 dB against 17 dB) but twice the link rate. This is where 6 GHz pays off.
• One floor up (B), 6 GHz low-power indoor clears its top rate by about 3 dB, which leaves almost nothing for what the model leaves out. 5 GHz has 12 dB to spare. This is the "6 GHz disappears upstairs" case.
• Through masonry (C), low-power indoor 6 GHz misses even the Wi-Fi 6-level top rate by about 6 dB at either width, and drops to a lower rate (the datasheet doesn't list sensitivities for lower rates, so we can't say which). 5 GHz still reaches its Wi-Fi 6-level top rate.
• The gap between the bands is about 8.5 dB of margin at the top Wi-Fi 7 rate, 6 GHz at 160 MHz against 5 GHz at 80 MHz, and 9.5 dB through the block wall. Of the 8.5 dB, 3 dB is the assumed EIRP difference, 1.5 dB is free-space physics and 4 dB is the sensitivity figures: 2 dB for twice the channel width and 2 dB because the datasheet's 6 GHz receiver is less sensitive than its 5 GHz one at the same width. The block wall adds its extra 1.1 dB at 6.5 GHz. In free space, 8.5 dB is a factor of about 2.7 in distance; through walls and floors, it buys less distance than that.
• Standard power closes the gap. At 160 MHz it adds 9 dB over low-power indoor, which puts 6 GHz level with this 5 GHz example, with twice the link rate.

The uplink is worse for 6 GHz low-power indoor: the phone's limit is 6 dB below the router's, so every 6 GHz margin above shrinks by up to 6 dB in the phone-to-router direction if the phone transmits at its ceiling.

Standard power and AFC: what it is and which routers document it

Standard power lets an access point use up to 23 dBm per megahertz and 36 dBm total on 6 GHz, indoors or out, but only in U-NII-5 and U-NII-7 and only under an automated frequency coordination (AFC) system. Under 15.407, the access point registers its coordinates, antenna height, FCC ID and serial number, gets back a list of frequencies and the maximum power allowed on each at that spot, and must check in at least once a day. AFC protects licensed microwave links, so the power you get depends on where you live, and some channels may not be available at full power.

AFC began commercially in 2024. On February 23, 2024, the FCC's Office of Engineering and Technology approved seven AFC systems for commercial operation: Qualcomm, Federated Wireless, Sony, Comsearch, Wi-Fi Alliance Services Corporation, the Wireless Broadband Alliance and Broadcom.

The router has to support standard power, and so does the phone or laptop. These are the routers and mesh systems in our catalog whose makers mention AFC, from a search of each 6 GHz model's product page and ASUS's AFC support page on November 7, 2026:

• ASUS ROG Strix GS-BE18000 and ROG Rapture GT-BE19000AI. ASUS's AFC support page (updated December 2, 2025) applies to these two. It says AFC works only in the United States, "Currently only support AFC on Router operation mode", and "AiMesh node will support AFC via later firmware upgrade". It also says "Both the access point (AP) and client device must support AFC", and lists the devices it calls officially announced: the iPhone 15 Pro, 15 Pro Max and iPhone 16 models with iOS 18.4 or later, several M2 and M4 iPads with iPadOS 18.4 or later, the Samsung Galaxy S24 Ultra and S25 Ultra, Intel AX210 and BE200 adapters with driver 23.130 or later, and MediaTek's MT7925. That is ASUS's list, not the device makers' own. The GS-BE18000 product page also says "AFC service included for the first 4 years; additional charges apply thereafter". ASUS doesn't publish the later charge.
• ASUS ROG Strix GS-BE12000. Its product page describes AFC standard-power mode and carries the same four-year footnote, but the next footnote on that page, about where and when AFC is available, names the GS-BE18000, and ASUS's AFC support page doesn't list the GS-BE12000. Confirm with ASUS before buying it for AFC.
• TP-Link Deco BE65-Outdoor. An outdoor mesh unit. Its product page promotes a dedicated 6 GHz band "now available outdoors with AFC", but its footnote says the AFC function "will be supported through subsequent software updates, and the support time is subject to actual conditions".
• Google Nest Wifi Pro. A Wi-Fi 6E mesh system. Google's spec page says it is "Designed to support higher-power 6 GHz channels, when available (AFC ready)". We found no Google documentation saying AFC has been turned on, so treat it as a possibility, not a feature.

ASUS's own claim for the benefit: "With the commonly used 160MHz bandwidth at a mid-to-long distance of 52.8 feet, AFC can provide around a 1.6x higher performance." That's the maker's figure under its test conditions. Our link budget gives 9 dB more power at 160 MHz, which is consistent with a gain at the edge of coverage and none close to the router, where low-power indoor already reaches the top rate.

Two things are still moving. In February 2026 the FCC proposed letting AFC systems take building entry loss into account for access points that work in both low-power indoor and standard-power modes. The FCC's notice says this "will provide stronger signals and better coverage inside homes and buildings". That is a proposal, with comments due in March and April 2026, not a rule. And this list covers the makers' documentation we found; check a router's spec sheet for "AFC" or "standard power" before assuming either way.

What to do with this

6 GHz can be worth having even if it doesn't reach every room. It's less crowded and has no radar-shared channels, as our Wi-Fi 6, 6E or 7 guide explains. A Wi-Fi 7 device can also hold a 5 GHz link alongside 6 GHz through multi-link operation, so when 6 GHz fades it has another link already in place; our MLO explainer lists which devices do that. The iPhone side, including why a phone may stay on 5 GHz after it comes back in range, is in why your iPhone won't connect to 6 GHz and our iPhone Wi-Fi 7 guide.