2×2, 4×4, Whatever It Takes

How many antennas does an apartment actually need?

In the classic 1983 movie Mr. Mom, Michael Keaton is asked if he’s wiring his house for 220 volts, famously bluffing back: “220, 221, whatever it takes.” This has become vernacular for the universal urge to sound decisive while quietly assuming bigger numbers simply mean better results.  This mindset mirrors the current state of multifamily Wi-Fi design. Faced with the choice of gateway hardware, planners often look at antenna configurations with that exact same nonchalance: “So, you putting in Wi-Fi?  2×2, 4×4, whatever it takes.” But in a dense multi-dwelling unit (MDU), blindly maxing out antenna counts can create real collateral damage. 

The loudest arguments tend to be about controller platforms, cloud versus on-prem, which vendor’s RRM is smartest. But one decision sits upstream of all of that and is independent of whichever controller you run: how many antennas each radio should have. Wi-Fi 7 gateways ship in 2×2 and 4×4 flavors, and the instinct is to grab the bigger number. Too few antennas and you can’t blanket an apartment with a strong enough signal. Too many and you flood the building with signal you don’t need, making radio resource management harder for every unit around you. The right answer is “whatever it takes”, and for the average apartment, that turns out to be less than you’d think.

The average apartment

The average U.S. apartment is about 910 square feet. One-bedrooms are the largest share of new construction at roughly 48%, two-bedrooms at roughly 39% and averaging about 1,095 square feet.

Interior partitions are light: studs and half-inch gypsum board, each wall costing only a few dB at 5 GHz. The longest path inside the unit, bedroom-corner AP to far kitchen corner, is roughly 35 to 45 feet through one or two of those walls. What separates you from your neighbors is heavier, often doubled and sometimes concrete, but the lower bands cross even those with ease.

What the test data shows

We surveyed two Wi-Fi 7 devices across an entire floor of a real multifamily building, with the AP inside one apartment. One gateway runs 4×4 on all three bands; the other gateway runs 2×2 on 2.4 and 5 GHz and 4×4 on 6 GHz. Because the survey covers the whole floor, each map shows two things at once: coverage inside the host unit. Our bar for “adequate” is an RSSI better than –70 dBm, enough for a modern client to hold a high-rate link.

2.4 GHz: everyone gets covered, including the neighbors. Both radios paint not just the host apartment but virtually the entire floor red, with readings in the –40s and –50s dBm two and three units away from the AP. A 2×2 already over-covers the host unit; the 4×4 simply shouts louder into homes it was never meant to serve. The second pair of chains here is hardware you buy, tune, certify and ship so the radio can reach units you don’t sell to.

5 GHz: enough at home, still loud next door. Inside the host apartment, both configurations hold better than –72 dBm in every room. The 4×4’s extra streams buy margin mainly at the floor’s far corners, other people’s living rooms, where readings drop into the –70s. Judged by what the paying subscriber experiences, 2×2 and 4×4 are indistinguishable. Judged by what the neighbors’ radios must plan around, the 4×4 is a bigger interference source. That is a premium per unit for an outcome nobody asked for.

6 GHz: physics does the cell planning. Both devices run 4×4 here, and their maps are nearly identical: red and orange in and around the host unit, green mid-floor, blue everywhere beyond. Even a full 4×4 array cannot push 6 GHz across the building, and in an MDU that is a feature, not a bug. The band’s indoor rules also cap radiated power, so extra chains are the only lever left for link margin. This is the one place a 4×4 clearly earns its cost: multi-gigabit rates where the subscriber lives, minimal outside the residence, giving every apartment its own contained cell and letting channels be reused a few doors down.

The heatmap, in one picture

The simulated heatmaps below scale the in-unit behavior onto an average-size apartment footprint with the Wi-Fi gateway placed in a corner of the residence.  Read down each column: 2.4 and 5 GHz are saturated for both configurations, and 6 GHz is nearly identical because both arrays are 4×4 there. The change from 2×2 from 4×4 on the 2.4G and 5G bands only changes the far corner of the apartment by ~4dB.  There is still plenty of margin to maintain max throughput.  Spending the extra $ for the larger antenna array does not improve the performance in the small residence.

Simulated RSSI over an average ~900 sq ft apartment, AP in the bedroom corner.

But doesn’t 4×4 mean faster Wi-Fi?

Only sometimes. Spatial streams are capped by the weaker end of the link, and nearly every phone and laptop is 2×2, so a 4×4 AP still sends at most two streams to a single client. What the bigger array adds is roughly 3 to 5 dB of beamforming gain and 3 dB of receive combining, which indoors buys only 15% to 25% more range.

That range converts to speed only where the client sits below about –65 dBm. Above that line the client is already at top modulation, and extra signal buys literally zero throughput. In a 900 sq ft unit, nearly the entire floor plan is above that line on 2.4 and 5 GHz. The genuine 4×4 win is MU-MIMO, serving two 2×2 clients in one transmit opportunity, worth 20% to 50% aggregate capacity in device-dense homes. That is a capacity argument, not a coverage one, and worth paying for only where device density demands it.

Why more isn’t better in an MDU

The Wi-Fi Heatmaps make the abstract argument concrete. Every dB your gateway radiates past its own walls becomes part of the neighbor’s noise floor, and the survey shows the lower bands arriving next door strong enough to force co-channel avoidance. Multiply that by every unit on the floor and RRM’s job becomes triage: fewer clean channels, more deferrals, more airtime lost for the resident.

The operator ends up paying a hardware premium on every gateway to radiate signal your subscriber cannot use, causing airtime challenges into nearby apartments you also serve and support. At MDU scale, that is spending money to make your own network harder to run. A building of right-sized 2×2 lower-band radios is far easier for any controller to coordinate than a building of 4x4s shouting over one another.

What else moves coverage

Antenna count is rarely the decisive input, and the things that beat it are mostly free.

Placement matters more. A centered AP beats a corner one by 6 dB or more to the worst room, at no cost at all.

Construction matters more still. One metal shaft on the surveyed floor carves a deep blue hole in an otherwise saturated 2.4 GHz map, while drywall barely registers. Concrete, tile and foil-backed insulation do the same at scale, and no antenna count fixes them.

Ambient interference sets a ceiling no transmit power can punch through: the neighbors’ APs above, below and beside. In a dense MDU the environment, not raw power, decides the experience, which is why extra chains rarely reduce the support calls and truck rolls that dominate the real cost of the service.

So how many antennas does an apartment need?

For the average North American apartment, a 2×2 radio already blankets 2.4 and 5 GHz; the survey shows both bands reaching well beyond the unit at either antenna count. On 6 GHz, where propagation is hardest and transmit power capped, a 4×4 array is worth the cost.

Every chain you add is a power amplifier, a filter, an antenna and a share of the thermal, test and certification budget, repeated on every unit you ship. The question on each band is whether the space needs it: on 2.4 and 5 GHz in an average apartment it does not, and on 6 GHz it does.

The smart design isn’t “max everything.” It’s matching the array to the band and the space: take advantage where propagation is easy, use the extra antennas on 6 GHz where it isn’t. Minimize cost with a 2×2 / 2×2 / 4×4 split Wi-Fi gateway makes sense for in unit MDU deployments. 2×2, 4×4, whatever it takes, with the emphasis on what the space actually needs.