Your Wi-Fi is slow because radio bandwidth is being shared, weakened, or wasted—not created. Three things set the ceiling on a Wi-Fi link: the channel width (how much spectrum you occupy), the signal-to-noise ratio at your device (how many bits per hertz that spectrum can carry), and the airtime you win against everyone else on the same channel. When your speed drops, one of those three has changed. Most slow-Wi-Fi complaints trace to a shared channel, a congested band, or an old standard, not to your internet provider.
That explanation fits on one screen. The rest of this article explains each ceiling, why the numbers on the box never match the numbers you see, and which fixes actually move the needle.
The bandwidth ceiling
Wi-Fi speed is bounded by the Shannon–Hartley theorem: capacity equals bandwidth times the base-2 logarithm of one plus the signal-to-noise ratio. That formula has two practical consequences. Padding the equation with a fixed channel gives you the first one: more spectrum means more capacity up to a point. Your data rate ceiling is the channel width multiplied by the bits-per-hertz that the signal-to-noise ratio and encoding scheme allow.
In practice, a 20 MHz channel in the 2.4 GHz band carries only three non-overlapping channels (1, 6, and 11), so neighboring networks in an apartment building often share one. A 5 GHz channel is usually 20, 40, 80, or 160 MHz wide, which is why the 5 GHz band is far less crowded and why high-end routers advertise 160 MHz support. Wider channels carry more data per second but occupy more spectrum, so interference from nearby networks affects them more. The router chooses channel width at the access point; older routers fix it at 20 MHz to reduce interference, which caps throughput well below the advertised figure.
Signal-to-noise ratio, not signal strength alone, sets your speed. A strong signal drowned in noise from a neighboring access point performs worse than a weaker clean signal. Noise includes co-channel interference, Bluetooth traffic, microwave oven leakage in the 2.4 GHz band, and even your own devices.
What the numbers on the box mean
Manufacturers advertise the sum of all spatial streams across all channels and bands, which no single client ever reaches. A "Wi-Fi 6 AX3000" label sums the 2.4 GHz and 5 GHz theoretical data rates, often 600 Mbps and 2400 Mbps. A typical phone has two spatial streams and cannot use both bands simultaneously for one download, so its practical share is a fraction of 2400 Mbps. The real-world ceiling for a single 80 MHz Wi-Fi 6 connection at good signal is roughly 600–900 Mbps of the link rate, and actual file transfers lose additional overhead to management frames and retransmission.
Wi-Fi standards compound the gap. On a 20 MHz channel, 802.11n tops out near 72–87 Mbps per spatial stream, 802.11ac roughly twice that, and Wi-Fi 6 adds roughly 30 percent through denser encoding and better scheduling. A 2014 laptop with an 802.11n adapter on a modern router is capped by its own radio, not the router's.
The device decides the standard too: the access point and the client negotiate the fastest mode both support. A new router cannot push a legacy phone past its own limit.
Airtime sharing and congestion
Every Wi-Fi device on a channel competes for airtime, and 802.11 is a listen-before-talk protocol. A device that hears traffic backs off, so the more devices transmit, the more time is spent waiting rather than sending. Adding a dozen smart-home plugs that each send small bursts monthly is not the problem; adding two neighbors on the same 2.4 GHz channel is. On a busy apartment channel, throughput can fall well below a quarter of what a quiet channel delivers, even with a perfect signal to your router.
The hidden culprit is your own home. An old 2.4 GHz-only baby monitor, a cordless phone, or a legacy printer running 802.11b forces one of its own repeated low-rate frames onto a channel, and everyone sharing that channel pays the airtime tax. A single misconfigured device can drag a channel's aggregate speed down substantially.
Modern routers mitigate this with orthogonal frequency-division multiple access (OFDMA) and multi-user MIMO (MU-MIMO), which split airtime by subcarriers or spatial paths. These features help only when both access point and client support them, and on a crowded channel only against legitimate Wi-Fi; interference from non-Wi-Fi sources evades them entirely.
| Setting | 2.4 GHz | 5 GHz | 6 GHz (Wi-Fi 6E) |
|---|---|---|---|
| Non-overlapping 20 MHz channels | 3 | 24–25 | 59 |
| Typical channel widths | 20 MHz | 20/40/80/160 MHz | 20/40/80/160 MHz |
| Best for | Range, walls | Throughput, density | Throughput, clean spectrum |
| Congestion risk | High | Medium | Low |
| Penetration through walls | Best | Moderate | Worst |
| Realistic single-device rate | ~50–150 Mbps | ~400–900 Mbps | ~600–1200 Mbps |
The comparison assumes current Wi-Fi 6/6E gear and a single nearby client. Older clients and wider channels shift the numbers lower.
Regional rules and a hidden threshold
Regulatory rules set what your router can legally transmit. In the United States, the FCC allows 2.4 GHz channels 1 through 11 and limits transmit power to 1 watt (30 dBm) of effective isotropic radiated power on that band. The 5 GHz band permits higher power on some channels and lower power on others; the 6 GHz band comes with lower power unless the access point uses a special mode for outdoor use. These rules are why some routers sold globally disable certain channels when the region is set to a low-power profile.
The counterintuitive professional caveat: the strongest possible signal is not the goal. On a mesh or multi-access-point network, dumping maximum power on every unit creates roaming problems, because devices cling to a distant strong access point instead of switching to a nearby one. The reference access points use power budgets that keep cells roughly equal in received strength at the handoff boundary. A single router is often better off at moderate power with the device closer, not at maximum power from far away.
Red flags and quick triage
Speeds collapse only on one device: check its Wi-Fi adapter standard and driver version before blaming the router. Slow in the evening but fast at noon: congestion or ISP shaping, not your radio. Speed good at the router and bad in one room: signal attenuation and walls, not channel width. Wi-Fi fast but downloads slow: the internet path, DNS, or the server, not the radio. Only the 2.4 GHz band is slow: interference from neighbors or a legacy device on the channel. Random eight-second stalls: background scans by phones looking for a better network.
Practical actions that move the needle
Move downstream tests to the 5 GHz band first. A phone on 2.4 GHz near the router is often limited to roughly 100–150 Mbps by that band's capacity and channel width, while the same phone on 5 GHz reaches several hundred. If the 5 GHz band works, the problem is likely band selection or congestion.
Site-survey before you buy hardware. Use a free Wi-Fi analyzer app on a laptop or phone to list every neighboring network, its channel, and its signal level. If three or more networks sit within about 10 dBm of yours on the same channel, switch channels or move to 5 GHz. On 2.4 GHz, choose 1, 6, or 11 and keep the channel width at 20 MHz to avoid bleeding into neighbors.
Restrict legacy devices. If a tablet or printer only supports 802.11n, it limits the channel less than a broadband single-antenna client does, but any 802.11b-era device should be isolated. Where the router permits, disable legacy data rates below 6 Mbps, which shortens the airtime cost of every slow client's transmission.
Set a realistic target. Measure over Ethernet with a known server, then compare the Wi-Fi result on the same server. A single 5 GHz Wi-Fi 6 client at 3–5 meters through one drywall partition typically achieves 60–80 percent of the wired rate. Repeated File Transfer Protocol tests to a local server remove internet variables.
Check for driver and firmware updates from the adapter vendor and the router manufacturer's support pages, not the operating system's driver store. If the 2.4 and 5 GHz networks share a single name, disable band steering temporarily to see which band the device picks.
Mesh systems need wired backhaul or a dedicated backhaul radio. A wireless mesh hop can halve throughput before you ever see interference.
Limits and what the numbers cannot promise
Not every slow connection is fixable from your side. Licensed Wi-Fi channels in dense apartment buildings are a shared resource, and no setting makes more spectrum appear. If your provider delivers 200 Mbps to the modem but every device on 2.4 GHz tops out near 80 Mbps, that is radio physics, not a defect.
Signal strength reported by devices uses their own scale and is often rounded; compare the signal-to-noise ratio rather than the signal bars. Speeds measured through a browser or a speed test app are also bounded by that server, the operator's network, and the CPU of the device running the test.
Old devices, old drivers, and hidden interference can defeat every router setting. A clean 5 GHz channel with reasonable signal is the most reliable fix on the list, and if file transfers between two local devices are slow with wired speed confirmed, the radio is the bottleneck you can measure directly.
None of this diagnoses a specific house or recommends a purchase. It describes the mechanisms well enough that you can test the ceilings yourself.