What do 2x2 and 4x4 router antennas mean? The phone's speed is stuck at 867Mbps, so I have calculated the spatial flow issue clearly

When

buy the router, the details say "4x4 MIMO" and "4 spatial streams." At first, I thought it meant the number of antennas—more antennas mean more threads pass through the wall. Until last month, I plugged in a 39-yuan USB wireless card for the desktop, and no matter how fast I tested, it was only about 70 Mbps. When I checked the negotiated speed of 144Mbps, I followed that the whole set of 2x2, 4x4, 867, and 1201 had nothing to do with wall penetration; it only managed how many lanes it could drive at once.

to clarify first: 2x2 is not two antennas, but two spatial streams

the term 'spatial flow' sounds mysterious, but it's easier to understand if you treat it as a driveway. WiFi data transmission: one stream is a parallel lane, and with two lanes running simultaneously, the speed doubles. The two streams on the router end are sent out, and the two streams on the phone end are retracted, making it a 2x2 stream.

there are two pitfalls here that I've stepped on before. First, antenna number does not equal the spatial flow . My Xiaomi AX3000T has four antennas standing outside. The 5G band is actually 2x2, and those four antennas are used separately for two bands, not four streams. The antenna number actually corresponds to gain and band division, which I calculated in the antenna gain. The second pitfall is even worse: negotiation to exploit weaknesses . The router has 4 streams, the phone has 2 streams, but in reality, only 2 can run, with the remaining two idle. My ASUS RT-AX86U is 4x4, connected to my Xiaomi 13 (2x2), and the negotiated speed is still capped at 2402Mbps, not a cent more.

so don't get too excited just by looking at the router's specs; first check how many streams are on the other side of the phone. Nowadays, whether it's 2,000 or 7,000 yuan phones, almost all on the market are 2x2; The commonly used AX200 and AX210 network cards on laptops are also 2x2. The real 1x1 models are cheap items like TV boxes, generic USB network cards, and old cameras.

How are the numbers

867, 1201, and 2402 calculated

these numbers aren't made on a whim; the formula is simple: total rate = number of spatial streams × the base price of a single stream . The base price is determined by WiFi algebra (modulation method) and bandwidth (20/40/80/160MHz).

In the

WiFi5 era, the 5G band had 80MHz bandwidth, one stream at 433Mbps, two lines at 867—the famous 866Mbps on old phones came from this way. By WiFi 6, modulation was upgraded to 1024QAM, one stream at 80MHz increased to 600Mbps, and two streams totaled 1201Mbps; Bandwidth doubles to 160MHz, with two streams at 2402Mbps. The AX3000 is the '3000' made up of the 2.4G 574 and 5G 2402. I calculated the generational differences between WiFi 6 and WiFi 7 before upgrade. Here's the comparison table, so you can look up the number of monsters directly in the future:

one detail worth remembering: many phones are default locked to the 80MHz bandwidth. Even if both the router and phone support 160MHz, you might only see 1201. The reason is simple: 160MHz occupies two consecutive 80MHz bands, with a DFS radar channel in between, so when radar signals are detected, it must actively yield. To save power and maintain stability, phone manufacturers lock many models at 80MHz. Within gigabit broadband, negotiation for 1201 is actually sufficient, so there's no need to overthink it.

add a note about the relationship between negotiation speed and actual internet speed, so you don't think you're being scammed just because the numbers don't match: negotiation is the theoretical limit, and in reality, it can reach 60-70% of the rate. For phones negotiated by 866, 5G real-world tests generally support 500 to 600 Mbps; The 2402 is negotiated, and gigabit broadband can reach around 940. Screenshots from the air interface, encryption costs, and the neighbor's WiFi all take a cut from this, which is a normal loss.

my test: 4x4 against a single 2x2 phone, basically a wasted

to verify this, I placed the RT-AX86U (5G 4x4) and AX3000T (5G 2x2) in the same position, with the same Xiaomi 13 connected to 5G at intervals of three meters, running five rounds of speed tests to average the average.

results were far from exciting: both devices negotiated speeds at 2402Mbps, with a speed of 941 versus 934, and the 7Mbps difference was purely a fluctuation. The few hundred yuan more expensive than the 2x2 4x4 hardly shows up in the "single phone" scenario. Its real effect is at the edge of weak signals: 4x4 beamforming is stronger, and the rate attenuation slows down between two walls. In my bathroom, 4x4 is stable around 400 Mbps, while the 2x2 unit drops to around 300 Mbps—there's a difference, but it's far from doubling. Also, the bottleneck now is mostly in signal power, not in the number of streams. The solution the article on signal strength tuning more targeted.

the real negative example is my 39-piece USB network card: 1x1, which can only connect to 2.4G, negotiated 144Mbps, and tested 72 to 80MB. The real culprit for the desktop gaming lag is not broadband, but the network card having only one lane. In this situation, changing routers is completely useless.

When will 4x4 really come into use

you can't tell the difference with a single device, multi-device concurrency is the real domain of 4x4. Four streams can be split and feed two 2x2 phones each with two streams at the same time. That's what MU-MIMO does. I've written a plain version of the on the principle. At home, when binge-watching, gaming, and casting at home, the 4x4's high-density scheduling is truly effortless.

second scenario is Mesh wireless backhaul. Sub-routers and main routers transmit data wirelessly. If the backhaul is only 2x2 and you have to free up your own equipment to serve, the bandwidth is split in half. Sub-routers are provided by 5G to 4x4, with two backhaul and two service devices, which can support the slogan "wireless backhaul without speed drops." Of course, if you have a network cable, prioritize wired backhaul. I covered all the connection methods and pitfalls in this .

on the other hand, there's no need to feel sorry for those 1x1 devices at home. Cameras and smart sockets only get several KB of data per second, more than enough for a single lane. What really causes them to lag isn't low traffic numbers, but too crowded 2.4G channels. I've written about this troubleshooting approach in my article on channel interference.

How to check your device's status: a single command

Windows computer, you don't need administrator privileges; just type '

netsh wlan show interfaces

output, look for 'Number of receiving antennas' and 'Number of transmission antennas.' My laptop shows 2 and 2, so it's 2x2.

router side also has something to look at: Xiaomi enters the device list in the backend, and clicking on the specific device shows negotiated speeds; In the ASUS backend client list under "Wireless Network," each device's connected frequency band and negotiated speed are listed in a row, making it clear who is sneaking away.

phone doesn't show this, so I use negotiated speed to deduce: after connecting to 5G, check the connection speed. 866/867 is WiFi 5 2x2, 1201 is WiFi 6 2x2 80MHz, 2402 is WiFi 6 2x2 160MHz, 287/300 is mostly 2.4G, and seeing 144 is basically 1x1.

I followed this method to run through my home: all three phones were 2x2, the laptop was 2x2, the TV box was 1x1 (I found out the reason it was old buffering), and the two cameras were 1x1. FamilyMart only has a 4x4 router, which means four lanes are half the sky. By the way, if the cable port speed is off, that's a different story. Check network port status check .

Conclusion: How to spend this money

order is clear: first count the number of devices at home, then check the bandwidth bandwidth, and finally it's time to pick the router. With broadband within gigabit and all devices being 2x2, a 2x2 machine at the AX3000 level can run the full time. The price difference saved can be spent on a cable for wired backhaul, which is much more practical than stacking 4x4. If you really have twenty or more devices running concurrently, or Mesh can only backlog wirelessly, then consider a mid-to-high-end 4x4 machine.

answered the two frequently asked questions casually. First: "Will my old phone get faster after switching to a 4x4 router?" No, old phones are 2x2 or 2x2, and negotiating whether to use the 867 or 867, at most the edge signal is a bit more stable. Second: "Should WiFi 7 phones be equipped with 4x4?" Nowadays, WiFi 7 phones are basically 2x2, and their speed boost relies on MLO dual-band aggregation methods, which is different from stacking data numbers. Don't let the details page lead the way.

remember one sentence: speed lag, it depends on how many lanes there are on the other end of the phone, not how many antennas the router has.