Ethernet vs WiFi: When Wired Actually Matters

By Paul Snyman · Published August 9, 2026 · 5 min read

Modern WiFi is good enough for most things now - that's genuinely true, and worth saying plainly rather than pretending ethernet is always the answer. But "most things" isn't "everything," and knowing exactly where the gap still matters saves you from either running cables you didn't need or skipping one you did.

Where the Gap Has Mostly Closed

For raw download throughput, current WiFi standards can match or exceed what most home internet plans deliver, so for streaming, browsing, and general use, a well-placed modern router genuinely performs close to wired for most people. If your WiFi feels slow, the far more common cause is placement, congestion, or your internet plan itself - see our mesh WiFi dead zones guide - not an inherent WiFi-vs-ethernet ceiling.

Where Ethernet Still Wins Clearly

The common thread isn't raw speed - it's latency and consistency. Even with strong signal and high throughput, WiFi involves more processing overhead and occasional retransmission from interference that a direct physical ethernet connection simply doesn't have. The gap is often small in absolute terms, but it's there consistently, which is exactly what latency-sensitive uses are sensitive to. See our cloud gaming latency guide for more on why this specifically matters there.

Testing Your Own Setup

Run Speed Test and Ping Test over WiFi, then run the same tests with the device connected via ethernet if that's an option. A meaningful gap - especially in latency or consistency across repeated tests, not just one reading - tells you whether wiring that specific device is actually worth the cable run.

The Gap That Has Not Closed: Consistency

Throughput comparisons between Ethernet and Wi-Fi have become genuinely close, and for most of what people do at home the headline numbers no longer decide anything. The difference that has not closed, and structurally cannot, is consistency.

Ethernet gives every device its own dedicated, full-duplex path to the switch. Your laptop can send and receive at the same time, and nothing else on the network affects when it gets to do so.

Wi-Fi is a shared radio medium^[source]. Only one device on a channel transmits at a time, so every device takes turns, and the more devices there are the longer the turns take to come round. On top of that, radio is lossy: a frame that is not acknowledged is sent again, and a retransmission is not a lost packet, it is a late one.

The result is that Wi-Fi latency is not a number, it is a distribution. A link might average 8 ms and spend a fraction of its time at 60. Ethernet on the same network sits at well under a millisecond and stays there. Averages hide this completely, which is why a Wi-Fi connection can benchmark beautifully and still feel worse for calls and games than a cable that tests slower.

Where Wi-Fi Genuinely Wins

This is not a one-sided argument, and pretending otherwise would be silly.

Wi-Fi is available everywhere in the house without drilling anything, which is not a small advantage. It costs nothing per device. Modern standards handle a crowded household far better than their predecessors did, largely because they let an access point talk to several devices in one transmission rather than strictly one at a time. And for anything that is not sensitive to a few tens of milliseconds, which is most browsing, streaming and downloading, the difference is genuinely invisible.

A 4K stream needs perhaps 25 Mbps and buffers several seconds ahead. It does not care about jitter at all. Running a cable to a television to improve Netflix is solving a problem that does not exist.

Choosing Per Device, Not Per House

The useful framing is not which is better but which each device should use.

Cable, where it is easy: a desktop, a games console used competitively, a NAS, a work machine that lives on video calls, and any mesh node you can reach. A mesh node with a wired path back to the router is the single highest-value cable in most homes, because it removes the wireless hop that every device behind it otherwise inherits.

Wi-Fi, without hesitation: phones, tablets, laptops that move, and every smart home device, all of which either cannot take a cable or gain nothing from one.

If the run is genuinely impossible, the in-between options are worth knowing. Powerline adapters carry Ethernet over the mains and are heavily dependent on your wiring, so results range from excellent to unusable and the only way to know is to try. MoCA carries it over existing coaxial television cabling and is far more consistent where that cabling exists.

Testing Which One You Are Actually On

Before changing anything, measure the thing that matters. Run a continuous ping on the device over Wi-Fi for a minute and note the average and, more importantly, the spread between best and worst. Then plug it in and do the same.

If both are flat, the cable will not improve anything you can perceive and you can stop. If the Wi-Fi run shows spikes and the wired one does not, you have just measured the entire argument, and you know exactly which devices deserve a cable.

One thing to check while you are there: a cable is only as good as its slowest part. A gigabit port at both ends with an old cable or a cheap unmanaged switch in the middle can silently negotiate 100 Mbps, and it will still say connected. If a wired device is inexplicably slower than the Wi-Fi it replaced, that is the first thing to look at.

Related Articles

• How to Reduce Latency for Cloud Gaming
• Mesh WiFi Dead Zones: How to Find and Fix Them
• Mesh WiFi vs. WiFi Extender
• Speed Test App for iPhone
• How PingKit compares to other network tools
• PS5 Keeps Disconnecting From WiFi: How to Prove Where It Drops
• Hue Bridge Not Found on iPhone: Is It on Your Network?