Why Your Wi-Fi Is Slow: A Technical Deep Dive

"Reboot the router" is the universal fix because the router is rarely the problem and the reboot feels like progress. The actual reasons home Wi-Fi underperforms are physical, knowable, and mostly fixable — if you know which layer to blame. This is a walk through the radio physics, the protocol mechanics, and the one measurement that tells you which one you're fighting.

The spectrum: three neighborhoods, three personalities

Your Wi-Fi lives in three bands with very different physics:

• 2.4 GHz — long range, terrible bandwidth. It penetrates walls like a rumor, but the band is only ~80 MHz wide and shared with Bluetooth, microwaves, and every neighbor's router. Real-world throughput tops out around 50–100 Mbps even on "300 Mbps" gear.
• 5 GHz — the workhorse. Roughly 500 MHz of usable spectrum, less penetration, far more capacity. This is where modern Wi-Fi 5/6 lives.
• 6 GHz — Wi-Fi 6E/7's gift: a whole clean band with no legacy devices and no DFS radar restrictions. Shorter range still, but if you're within 5 meters of an AP, it's the fastest thing on the block.

The first diagnosis is always: which band are you actually connected to? Half of "my Wi-Fi is slow" turns out to be a device parked on 2.4 GHz because it has one bar of 5 GHz signal, and the router's band-steering gave up.

Channel width: the bandwidth you're not using

A Wi-Fi "channel" is a slice of spectrum, and its width determines raw speed: a 20 MHz channel carries a fraction of a 160 MHz channel. Modern routers default to wide channels (80 or 160 MHz) for speed — but wide channels are also more likely to overlap neighbors, and the protocol is polite: if anyone else is using part of your channel, both of you slow down. The classic fix that actually works: check your channel width with a Wi-Fi analyzer, and if you're in an apartment building, drop from 160 to 80 or 40 MHz. You'll lose headline numbers and gain real-world stability.

SNR: the number that explains everything

Ignore megabytes for a second. The single most predictive metric in Wi-Fi is signal-to-noise ratio — how much louder your router is than the noise floor. Below ~25 dB SNR, the radio falls back to slower modulation schemes (more redundancy per bit), and throughput collapses long before the signal "drops."

1
# Linux: see SNR and rate live for the current link
2
iw dev wlan0 link
3
#    connected to 84:xx:xx:xx:xx:xx (on wlan0)
4
#        signal: -52 dBm
5
#        tx bitrate: 866.7 MBit/s

The mental model: -50 dBm with 866 Mbps is a great link; -75 dBm with 144 Mbps is the same router being polite about a bad situation. When throughput is bad, check the link rate first. If the link rate is high and throughput is low, you have a different problem (below).

The retry problem: why high signal can still mean slow

Here's the counterintuitive case: full bars, high link rate, terrible throughput. The culprit is usually retries. Wi-Fi is half-duplex and acknowledged: the sender transmits, waits for an ACK, and retransmits on silence. In a congested or hidden-node environment, the channel is full of collisions and retries — every packet effectively sent two or three times. The radio is working; the medium is full.

Symptoms: slow speed that's consistent regardless of distance, and a router that's warm. Fixes, in order: move to a less-congested channel (analyzer again), reduce channel width, and if you're in a dense apartment — consider that Wi-Fi was never designed for 40 routers in one building and stop chasing it with a bigger router.

MIMO and spatial streams: the antenna math

Your router has multiple antennas and so does your phone; MIMO lets them communicate over parallel spatial streams. Consumer gear is typically 2×2 (two streams, ~866 Mbps on 5 GHz at 80 MHz) or 3×3. The catch: the link uses the minimum of both sides' stream counts. A 4×4 router talking to a 2×2 phone runs at 2 streams — the router's extra antennas are idle. This is why "AX6000" marketing numbers are fiction for a single device: those numbers assume 4 streams and 160 MHz and a perfect room.

Mesh vs. access points: the backhaul lie

Mesh systems are the most oversold home-networking product in existence. A mesh node connected wirelessly to the main router has to receive and retransmit everything on the same radio — your "fast" mesh node is typically operating at half its speed, because half the airtime goes to talking to the main node. Wired backhaul (Ethernet between nodes, or tri-band mesh with a dedicated backhaul radio) fixes this; wireless single-band backhaul doesn't.

The rule: if you can run one Ethernet cable, do. A $20 cable beats $300 of mesh. Access points on Ethernet are strictly better than wireless mesh at every price point, and the difference grows with traffic.

The ISP-shaped ceiling

Before you buy anything: test with a cable. speedtest over Ethernet tells you your internet speed; over Wi-Fi it tells you the weaker of (internet, Wi-Fi). If wired is 500 Mbps and wireless is 200, your Wi-Fi is the bottleneck and you can fix it. If wired is 80 Mbps, your ISP is the bottleneck — no router purchase will help, and "the Wi-Fi is slow" was never Wi-Fi at all.

A practical diagnostic order

1. Cable test — establish the ceiling. Wired slow → call the ISP.
2. Check the band — are you on 5/6 GHz? Split SSIDs temporarily if band-steering is misbehaving.
3. Check the link rate — low rate → signal problem: move the AP, reduce distance, add an AP.
4. Check channel width + congestion — apartment building → 40/80 MHz, pick a clean channel.
5. Count retries — high retries → congestion, not signal: change channel, reduce width, or accept physics.
6. Backhaul — wireless mesh? Run a cable. It's the last 10%, and it's the most expensive 10% to fix.

Wi-Fi is physics with a protocol on top. Once you know which layer is lying to you — signal, congestion, or the ISP — the fix is usually small, specific, and cheap. The router reboot was never the answer; the answer was knowing what to measure.